48
University of Wollongong Research Online Faculty of Informatics - Papers (Archive) Faculty of Engineering and Information Sciences 2006 A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses S. J. Kim KyungHee University, Korea H. D. Kim University of Wollongong Y. Lee Taeduk Radio Astronomy Observatory, Korea Y. C. Minh Taeduk Radio Astronomy Observatory, Korea R. Balasubramanyam Raman Research Institute, Bangalore, India See next page for additional authors Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: [email protected] Publication Details is article was originally published as: Kim, SJ, Kim, HD, Minh, YC, et al, A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses, e Astrophysical Journal Supplement Series, 2006, 162, 161-206. Copyright 2006 University of Chicago Press. e journal can be found here.

A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

  • Upload
    others

  • View
    1

  • Download
    0

Embed Size (px)

Citation preview

Page 1: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

University of WollongongResearch Online

Faculty of Informatics - Papers (Archive) Faculty of Engineering and Information Sciences

2006

A Molecular Line Survey of W3(OH) and W3 IRS5 from 847 to 1156 GHz Observational Data andAnalysesS J KimKyungHee University Korea

H D KimUniversity of Wollongong

Y LeeTaeduk Radio Astronomy Observatory Korea

Y C MinhTaeduk Radio Astronomy Observatory Korea

R BalasubramanyamRaman Research Institute Bangalore India

See next page for additional authors

Research Online is the open access institutional repository for the University of Wollongong For further information contact the UOW Libraryresearch-pubsuoweduau

Publication DetailsThis article was originally published as Kim SJ Kim HD Minh YC et al A Molecular Line Survey of W3(OH) and W3 IRS 5 from847 to 1156 GHz Observational Data and Analyses The Astrophysical Journal Supplement Series 2006 162 161-206 Copyright2006 University of Chicago Press The journal can be found here

A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156GHz Observational Data and Analyses

AbstractWe have carried out observations toward the W3 complex and G343+015 using the TRAO 14 m radiotelescope to examine in detail the chemical variations occurring while molecular clouds evolve from theprestellar to the H II region phase Observations include spectral surveys of these objects between 847 and1156 GHz mapping observations toward W3(OH) with the emissions of CS (2ndash1) HCN (1ndash0) HNC(1ndash0) and HCO+ (1ndash0) and mapping of CS (2ndash1) emission toward W3 IRS 5 Chemical model calculationsare used to estimate the age of W3(OH) by comparing with the fractional abundances of detected moleculesWe found that G343+015 and W3(OH) are at a similar evolutionary stage although large differences in thefractional abundances are found in CH3CN and HC3N Overall the properties of the detected species andabundances in three regions support the view that chemistry varies as molecular clouds evolve from a coldcollapsing phase to a high-temperature phase such as the hot core and H II phase Chemical modelcalculations for W3(OH) indicate that the evolutionary age of the cloud is 104ndash105 yr with temperature inthe range 10ndash60 K

DisciplinesPhysical Sciences and Mathematics

Publication DetailsThis article was originally published as Kim SJ Kim HD Minh YC et al A Molecular Line Survey ofW3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses The AstrophysicalJournal Supplement Series 2006 162 161-206 Copyright 2006 University of Chicago Press The journal canbe found here

AuthorsS J Kim H D Kim Y Lee Y C Minh R Balasubramanyam M G Burton T J Millar and D W Lee

This journal article is available at Research Online httprouoweduauinfopapers409

A MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 FROM 847 TO 1156 GHzOBSERVATIONAL DATA AND ANALYSES

Sang-Joon Kim1Hun-Dae Kim

2Youngung Lee

3Young Chol Minh

3Ramesh Balasubramanyam

4

Michael G Burton5Tom J Millar

6and Dong-Wook Lee

1

Received 2004 August 29 accepted 2005 May 31

ABSTRACT

Wehave carried out observations toward theW3 complex andG343+015 using the TRAO14m radio telescope toexamine in detail the chemical variations occurring while molecular clouds evolve from the prestellar to the H ii

region phase Observations include spectral surveys of these objects between 847 and 1156 GHz mapping obser-vations towardW3(OH) with the emissions of CS (2ndash1) HCN (1ndash0) HNC (1ndash0) and HCO+ (1ndash0) and mapping ofCS (2ndash1) emission toward W3 IRS 5 Chemical model calculations are used to estimate the age of W3(OH) bycomparing with the fractional abundances of detected molecules We found that G343+015 and W3(OH) are at asimilar evolutionary stage although large differences in the fractional abundances are found in CH3CN and HC3NOverall the properties of the detected species and abundances in three regions support the view that chemistry variesas molecular clouds evolve from a cold collapsing phase to a high-temperature phase such as the hot core and H ii

phase Chemical model calculations for W3(OH) indicate that the evolutionary age of the cloud is 104ndash105 yr withtemperature in the range 10ndash60 K

Subject headinggs H ii regions mdash ISM abundances mdash ISM individual (W3(OH) W3 IRS 5 G343+015)

1 INTRODUCTION

It is known that chemical properties of molecular clouds varyduring their evolution The onset of central protostars comparedwith the cold dark collapsing phase plays a critical role in driv-ing the chemistry in the surrounding dense region by providinga high temperature and density Molecular line surveys amongmany observational techniques are one of the most powerfulmethods for the investigation of the chemistry The identifiedmolecules which are sensitive to various chemical reactions inthemolecular clouds can be used to investigate the formation andevolution of molecular clouds Representative line surveys per-formed over the last few decades are listed in Table 1 Valuableinformation on interstellar chemistry and the evolution of mo-lecular clouds has been obtained from the surveys However asshown in Table 1 they have been limited to a few objects such asSgr B2 (Cummins et al1986 Turner1989 Sutton et al1991) andthe nearby Orion-KL region (Johansson et al 1984 Jewell et al1989 Turner 1989 Blake et al 1986 1996 ZiurysampMcGonagle1993 Schilke et al 1997)

Recently other sources have been chosen for molecular linesurveys the ultracompact H ii region G343+015 (MacDonaldet al 1996 Kim et al 2000) the low-mass star-forming regionIRAS 162932422 (van Dishoeck et al 1995) the protostellarregion IRAS 174702853 (Kim et al 2002) and the late-typestar IRC +10216 (Kawaguchi et al1995 Cernicharo et al 2000)It is important to study the sources for which significant chemicalvariations are expected as predicted by models Detailed ex-

amination has been undertaken for individual species present inOrion-KL and Sgr B2 (Turner 1991) in order to test the modelsAlthough the results toward these two sources have provided uswith interesting information on the nature of dense cores the com-plex physical structure of Orion-KL and Sgr B2 has limited ourunderstanding of the chemical processes The molecular abun-dances derived from these objects contain large uncertainties dueto the blending of adjacent transitions compared to low-mass andisolated objects Therefore Orion-KL and Sgr B2 do not providea favorable environment for the study of relative chemical varia-tions and ultimately for the determination of the evolutionarystage although observations toward the objects have enabled usto understand the close relationship between the chemistry and theevolution of molecular clouds

Observations toward the hot core (MacDonald et al1996) andthe cold halo (Thompson amp MacDonald 1999) of the sourceG343+015 have revealed significant chemical variations be-tween these two distinct physical components within this objectA distinct chemistry was also noticed from observations of SO2CH3OH and H2CS (Helmich et al1994) and submillimeter mo-lecular line surveys (Helmich amp van Dishoeck 1997) of the W3region

Models of gas-phase ion-molecule chemistry can reasonablyreproduce the observed abundance of carbon-chain species in thecold dark clouds (Herbst et al 1984 Millar amp Freeman 1984)However such ion-molecule chemical models alone cannot sat-isfactorily account for the observed high abundance of mole-cules containing H or S atoms in star-forming molecular cloudsCoupling of gas-phase and grain-surface reactions (eg vanDishoeck amp Blake 1998) has been emphasized in recent chem-ical models in order to cope with the discrepancy Representativemodels have been developed for hot cores byMillar et al (1991)and by Charnley et al (1995) According to these models hydro-gen sulfur ionized and neutral molecules are readily depositedonto dust surfaces as the density increases during the collapsingphase of interstellar clouds The formation of complex heavy spe-cies containing hydrogen becomes inhibited leading to slow re-actions of those species in the earlier collapsing phase (t lt 104 yr)

1 Department of Astronomy and Space Science KyungHee UniversityYoungin Kyunggi 449-701 Korea sjkim1khuackr

2 School of Information Technology and Computer Science University ofWollongong NSW 2500 Australia

3 TaedukRadioAstronomyObservatory San36-1Whaam-DongYouseong-GuTaejeon 305-348 Korea

4 Department of Astronomy and Astrophysics Raman Research InstituteSadashivanagar Bangalore 560080 India

5 School of Physics University of New South Wales Sydney NSW 2052Australia

6 Department of Physics UMIST PO Box 88 Manchester M60 1QD UK

161

The Astrophysical Journal Supplement Series 162161ndash206 2006 January

2006 The American Astronomical Society All rights reserved Printed in USA

Radiation from the protostar then heats up the surrounding gas andgrains releasing volatile species from the dust surface back to gasphase This stage is called the hot core phase and lasts 104ndash105 yrafter the onset of star formation (van Dishoeck amp Blake 1998)

We surveyedW3 region to investigate the chemical evolution aspredicted by recent chemical models We also observed G343+015 in order to identify several missing lines from a previous sur-vey (Kim et al 2000) In addition high-spectral-resolution ob-servations for molecules of interest (CS HCO+ HCN HNCCH3CN and CH3CCH) were carried out in G343+015 TheW3region contains several objects at different evolutionary stagesandG343+015 is awell-known hot core associatedwith an ultra-compact H ii region Our TRAO observations toward W3 andG343+015 provide us with valuable information on interstellarchemistry because the chemical properties of these sources dif-fering in mass core size and evolutionary stage are useful for in-vestigating the relationships between the chemical variations andthese parameters

The contents of this paper are as follows x 1 presents back-ground information on line surveys evolutionary scenarios andmodels presented in the literature In x 2 general aspects of theW3 complex focusing on W3(OH) and W3 IRS 5 are given Inxx 3 and 4 we present the observational procedures line iden-tifications and spectral displays In x 5 we discuss individualspecies of interest detected in the survey In x 6 we discuss thechemical variations as well as the isotopic ratios of C and Satoms found from G343+015 W3(OH) and W3 IRS 5 In x 7the cloud morphology found from CS HCO+ HCN and HNCmappings is presented In x 8 chemical results from model cal-culations are given We summarize the results in x 9

2 W3 COMPLEX

The W3 complex consists of several infrared sources (Wynn-Williams amp Becklin 1974 Jaffe et al 1983 Colley 1980) at dif-ferent evolutionary stages and core masses The various objectsinclude well-developed shell structure in the H ii regions (W3BW3C and near IRS 3 and IRS 4) and several ultracompact H ii

regions (W3FW3M and near IRS 7 and IRS 5) Strong OH andH2O masers viz W3(OH) and W3(H2O) separated by 6

00 fromeach other are also observed

At least two hot cores have been revealed in W3 from obser-vations of NH3 lines (Mauersberger et al 1988) Its distance is

estimated to be 183 kpc residing in the Perseus arm of the gal-axy (Imai et al 2000) Helmich amp van Dishoeck (1997) per-formed spectral line surveys through the 345GHzwindow towardW3 IRS 4W3 IRS 5 andW3(OH) Significant differences in thephysical and chemical conditions were found in IRS 4 IRS 5 andW3(H2O) and the gas temperature seems highest (T 220 K)toward W3(H2O) compared with those in IRS 4 and IRS 5(Helmich et al 1994) High-velocity (V frac14 52 km s1) flowsnear W3 IRS 5 and an intermediate-velocity (V frac14 26 km s1)wing near W3(OH) are found in 12CO emission (Bally amp Lada1983)

21 W3(OH)

The ultracompact H ii regionW3(OH) is an extensively stud-ied site for the investigation of star formation and its influence onthe surrounding regions (Wyrowski et al 1999) Three peaks aredistinguished with subarcsecond resolution observations in thethermal dust continuum emission originating from the hot coreregion of W3(H2O Wyrowski et al 1999) Continuum obser-vations reveal thatW3(OH) has a shell structure in the H ii region(Dreher ampWelch 1981) H66 H76 and H110 radio recom-bination lines having full width at half-maximum (FWHM) of36 41 and 69 km s1 respectively are observed (Sams et al1996) H41 CH3CN [5(2)ndash4(2)] HCN (1ndash0) HCO+ (1ndash0)and 13CO (1ndash0) are detected with amplitudes of 04 01 59 55and 76 K respectively from observations with the Nobeyama45m telescope (Forster et al 1990) Keto et al (1992)mademid-infrared and molecular line observations of W3(OH) and com-pared them with G343+015

22 W3 IRS 5

W3 IRS 5 is a bright infrared source with a total luminosity of3 105 L (Campbell et al 1995) and is separated by7500 fromW3 IRS 4 which is brighter than W3 IRS 5 in the submillimeterwavelength range (Jaffe et al 1983) W3 IRS 5 is younger thanW3(OH) and appears to be at an earlier stage evolving toward ahot corelike object (Helmich et al 1994) Very long baseline in-terferometry measurement (Imai et al 2000) of radial velocitiesand propermotions ofwatermasers in theW3 IRS 5 region revealstwo groups of maser components which are associated with atleast two different outflows

TABLE 1

Reported Molecular Line Surveys

Sources Telescope FWHM

Frequency

(GHz)

Resolution

(MHz)

Line Density

(TransitionGHz) Observations

Orion A OSO 20 m 4700 72ndash91 1 85 Johansson et al (1984)

IRC +10216 23

Orion A OVRO 104 m 0A5 247ndash263 103 152 Blake et al (1986)

Sgr B2 BTL 7 m 2A9 70ndash150 1 914 Cummins et al (1986)

Orion A NRAO 12 m 2400 330ndash360 2 54 Jewell et al (1989)

Orion A OVRO 104 m 0A5 215ndash247 162 170 Sutton et al (1985)

Sgr B2(M) CSO 104 m 2000 330ndash355 1 67 Sutton et al (1991)

Sgr B2Orion-KL NRAO 11 m 8300 70ndash115 1 143171 Turner (1989)

G343+015 JCMT 15 m 1300 330ndash360 0756 06 Thompson amp MacDonald (1999)

G343+015 TRAO 14 m 5800 86ndash1653 1 20 Kim et al (2000 2001)

G343+015 JCMT 15 m 1400 330ndash360 033 114 MacDonald et al (1996)

G589039 JCMT 15 m 1300 330ndash360 0756 47 Thompson amp MacDonald (1999)

IRC +10216 IRAM 30 m 1900 129ndash173 1 86 Cernicharo et al (2000)

IRC +10216 Nobeyama 45 m 3500 28ndash50 025 86 Kawaguchi et al (1995)

Orion-KL FCRAO 14 m 4000 150ndash160 1 180 Ziurys amp MacGonagle (1993)

IRAS 174702853 Mopra 22 m 3000 90ndash96 05 35 Kim et al (2002)

KIM ET AL162 Vol 162

Strong 12CO self-absorption and 13COandHCNemissionpeaksare seen toward IRS 5 the maximum 12CO emission is observedtoward the compact H ii region and a CS peak is found 800 west ofthe OH1720 MHz maser (Dickel et al 1980) The possible pres-ence of disklike structure around a cluster of embedded stars in themid-infrared images of IRS 5 is inferred (Persi et al 1996) Thetotal H2 mass of the molecular cores associated with IRS 5 is esti-mated to be 1100 M derived from 13CO (1ndash0) and C18O (1ndash0)observations (Roberts et al 1997) A luminosity of 60 105 Land a cloud mass of 1600 M have been derived using 230 GHzcontinuum observations (Gordon 1987) for the entire complex ofW3

3 OBSERVATIONS

We carried out molecular line surveys toward W3(OH)and W3 IRS 5 using the Taeduk Radio Astronomy Observa-tory (TRAO) 14 m telescope between 2001 November and 2002February The receiver system was a low-noise SIS receiver oper-ating in single-sideband mode For the observations of W3(OH)the back ends were two filter banks in serial mode each of whichhas a 1MHz resolution and 256MHz bandwidth providing a totalbandwidth of 512 MHz at each local oscillator (LO) tuning Sen-sitive 250 kHz measurements toward W3(OH) and W3 IRS 5were also performed using two 256 channel filter banks in parallelmode after completing the 1MHz observations ofW3(OH) IRAS184500200 and IRAS 183350711

During the observations the typical system temperature was300 K between 85 and 110 GHz and was 600 K at 115 GHzPosition switching mode observations were carried out taking areference position to be 300 away in right ascension which wasconfirmed to be free of emission Calibration was performedusing the standard chopper-wheel method which corrects for at-mospheric attenuation and scattering of forward spillover andthe intensity was derived in terms of T

A scale (Ulich amp Hass1976) The beam sizes of the TRAO 14 m telescope are 6400 and5500 at 85 and 115 GHz respectively

The total on-source integration time of each LO tuning for the1MHz observation was 300 s achieving channel-to-channel rmsnoise of 003ndash005 K which is consistent with that in G343+015 (Kim et al 2000) Therefore these observations providedan opportunity to compare the chemical properties of theW3 andG343+015 regions with data taken with the same telescope sys-tem and sensitivity Pointing and focusing were checked using Cygni

Almost all of the molecular transitions previously detectedfrom the 1 MHz surveys toward G343+015 (Kim et al 2000)W3(OH) IRAS183350711 and IRAS184500200 (thiswork)were also considered for the 250 kHz surveys towardW3(OH) andW3 IRS 5 We also observed selected molecules (HCO+ CSHCN HNC and their isotopic variants) with 250 kHz resolutiontoward the ultracompact H ii region G343+015 to examine iso-topic ratios and the HCN hyperfine ratio in detail We measuredline emissions from SO2 HNCO CH3OH [0(0)ndash1(1) E ] andCH3OH [2(1)ndash1(1) A] transitions toward G343+015

The detectedmolecules and their transitions show close agree-ment betweenG343+015 and IRAS 184500200 and betweenW3(OH) and IRAS 183350711 suggesting that physicalchemical condition and evolutionary stage are similar To estimatephysical conditions as well as to determine the initial parametersfor chemical model calculations mapping observations were car-ried out towardW3(OH) for the transitions of CS (2ndash1) and HCN(1ndash0) over 100 100 HNC (1ndash0) over 80 80 and HCO+ (1ndash0)over 60 70 We carried out only CS (2ndash1) mapping toward W3

IRS 5 because the line strengths of other species are too weak tomap with the relatively large-beam size of TRAO Mapping ob-servations of CS (2ndash1) and HCO+ in IRAS 184500200 andIRAS183350711were performed covering almost the same areaas W3(OH) CS (2ndash1) mapping of IRAS 184500200 shows ir-regular and elongated distribution in theN-S directionwhereas forIRAS 183350711 spherical distribution is shown The CS (2ndash1)distribution is consistent with that of 86 GHz radio continuum(Walsh et al 1998) There is no report on continuummap for IRAS183350711Wewill present images and detailed analyses ofmo-lecular lines detected in IRAS 184500200 and IRAS 183350711 in future work

The source positions of W3(OH) and W3 IRS 5 and thecentral position of G343+015 are as follows

W3(OH )mdash frac14 02h23m16s9 frac14 6138056B4 (19500

Helmich et al 1994)W3 IRS 5mdash frac14 02h21m53s1 frac14 6152020B0 (19500

Helmich et al 1994)G343+015mdash frac14 18h50m46s3 frac14 0111013B0 (19500

Kim et al 2000)

4 DATA REDUCTION AND DISPLAY

41 Spectra and Images

Spectra taken with the 1 MHz resolution of the 512 channelsspectrometer in serial mode have been averaged and then base-line subtracted using the SPA data reduction package developedat the FCRAO observatory Some bad channels especially at theedge of the spectra were significant at the upper frequency regionwhen LSB mode observations were performed The number ofbad channels in the worst spectra is about 15ndash20 equivalentto 15ndash20 MHz The total bandwidth for every observation is512MHz but observations were carried out by stepping the cen-tral frequency by 500 MHz resulting in an overlap of approxi-mately 6 MHz between spectra

For the 250 kHz resolution observations the spectrometerswere configured in a parallel mode and an average of signalswas taken The typical noise in an averaged spectrum is 003ndash005 K for 1 MHz data and 006ndash009 K for 250 kHz data Thereduced spectra presented in Figures 1 and 2 were taken towardW3(OH) with the 1MHz and 250 kHz bandwidths respectivelyThe spectra in Figure 1 were obtained by combining two neigh-boring spectra giving a bandwidth of1 GHz In Figures 3 and4 250 kHz spectra taken from W3 IRS 5 and G343+015 re-spectively are presented

Data taken with mapping observations were made into imagesusing the Contour and Trigrid functions in IDL and then theimages were compared with those processed using CLASS Thetwo images which were drawn in terms of antenna temperatureshowed a good agreement with each other The contour maps inFigure 5 were made by taking signals stronger than 5 whichcorresponds to 04 K

42 Line Identification

Line identifications were made exclusively using the catalogof Lovas (1992) The catalog lists molecular transitions detectedfrom Orion-KL Sgr B2 and from various interstellar molecularclouds The number of detected lines in the catalog is sufficientfor the use in the line identifications of our 3 mm observationsThe identified line frequencies were determined by Gaussian fit-ting after correcting by the corresponding Doppler shifts of46and 39 km s1 for W3(OH) and W3 IRS 5 respectively andthen compared with the rest frequencies from the Lovas catalog

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 163No 1 2006

Fig 1mdashSpectra of W3(OH) from 848 to 1154 GHz obtained during 2001 NovemberndashDecember with the TRAO 14 m telescope Each panel combines twoneighboring frequency measurements to have about 1 GHz bandwidth

164

Fig 1mdashContinued

165

Fig 1mdashContinued

166

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 2: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156GHz Observational Data and Analyses

AbstractWe have carried out observations toward the W3 complex and G343+015 using the TRAO 14 m radiotelescope to examine in detail the chemical variations occurring while molecular clouds evolve from theprestellar to the H II region phase Observations include spectral surveys of these objects between 847 and1156 GHz mapping observations toward W3(OH) with the emissions of CS (2ndash1) HCN (1ndash0) HNC(1ndash0) and HCO+ (1ndash0) and mapping of CS (2ndash1) emission toward W3 IRS 5 Chemical model calculationsare used to estimate the age of W3(OH) by comparing with the fractional abundances of detected moleculesWe found that G343+015 and W3(OH) are at a similar evolutionary stage although large differences in thefractional abundances are found in CH3CN and HC3N Overall the properties of the detected species andabundances in three regions support the view that chemistry varies as molecular clouds evolve from a coldcollapsing phase to a high-temperature phase such as the hot core and H II phase Chemical modelcalculations for W3(OH) indicate that the evolutionary age of the cloud is 104ndash105 yr with temperature inthe range 10ndash60 K

DisciplinesPhysical Sciences and Mathematics

Publication DetailsThis article was originally published as Kim SJ Kim HD Minh YC et al A Molecular Line Survey ofW3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses The AstrophysicalJournal Supplement Series 2006 162 161-206 Copyright 2006 University of Chicago Press The journal canbe found here

AuthorsS J Kim H D Kim Y Lee Y C Minh R Balasubramanyam M G Burton T J Millar and D W Lee

This journal article is available at Research Online httprouoweduauinfopapers409

A MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 FROM 847 TO 1156 GHzOBSERVATIONAL DATA AND ANALYSES

Sang-Joon Kim1Hun-Dae Kim

2Youngung Lee

3Young Chol Minh

3Ramesh Balasubramanyam

4

Michael G Burton5Tom J Millar

6and Dong-Wook Lee

1

Received 2004 August 29 accepted 2005 May 31

ABSTRACT

Wehave carried out observations toward theW3 complex andG343+015 using the TRAO14m radio telescope toexamine in detail the chemical variations occurring while molecular clouds evolve from the prestellar to the H ii

region phase Observations include spectral surveys of these objects between 847 and 1156 GHz mapping obser-vations towardW3(OH) with the emissions of CS (2ndash1) HCN (1ndash0) HNC (1ndash0) and HCO+ (1ndash0) and mapping ofCS (2ndash1) emission toward W3 IRS 5 Chemical model calculations are used to estimate the age of W3(OH) bycomparing with the fractional abundances of detected molecules We found that G343+015 and W3(OH) are at asimilar evolutionary stage although large differences in the fractional abundances are found in CH3CN and HC3NOverall the properties of the detected species and abundances in three regions support the view that chemistry variesas molecular clouds evolve from a cold collapsing phase to a high-temperature phase such as the hot core and H ii

phase Chemical model calculations for W3(OH) indicate that the evolutionary age of the cloud is 104ndash105 yr withtemperature in the range 10ndash60 K

Subject headinggs H ii regions mdash ISM abundances mdash ISM individual (W3(OH) W3 IRS 5 G343+015)

1 INTRODUCTION

It is known that chemical properties of molecular clouds varyduring their evolution The onset of central protostars comparedwith the cold dark collapsing phase plays a critical role in driv-ing the chemistry in the surrounding dense region by providinga high temperature and density Molecular line surveys amongmany observational techniques are one of the most powerfulmethods for the investigation of the chemistry The identifiedmolecules which are sensitive to various chemical reactions inthemolecular clouds can be used to investigate the formation andevolution of molecular clouds Representative line surveys per-formed over the last few decades are listed in Table 1 Valuableinformation on interstellar chemistry and the evolution of mo-lecular clouds has been obtained from the surveys However asshown in Table 1 they have been limited to a few objects such asSgr B2 (Cummins et al1986 Turner1989 Sutton et al1991) andthe nearby Orion-KL region (Johansson et al 1984 Jewell et al1989 Turner 1989 Blake et al 1986 1996 ZiurysampMcGonagle1993 Schilke et al 1997)

Recently other sources have been chosen for molecular linesurveys the ultracompact H ii region G343+015 (MacDonaldet al 1996 Kim et al 2000) the low-mass star-forming regionIRAS 162932422 (van Dishoeck et al 1995) the protostellarregion IRAS 174702853 (Kim et al 2002) and the late-typestar IRC +10216 (Kawaguchi et al1995 Cernicharo et al 2000)It is important to study the sources for which significant chemicalvariations are expected as predicted by models Detailed ex-

amination has been undertaken for individual species present inOrion-KL and Sgr B2 (Turner 1991) in order to test the modelsAlthough the results toward these two sources have provided uswith interesting information on the nature of dense cores the com-plex physical structure of Orion-KL and Sgr B2 has limited ourunderstanding of the chemical processes The molecular abun-dances derived from these objects contain large uncertainties dueto the blending of adjacent transitions compared to low-mass andisolated objects Therefore Orion-KL and Sgr B2 do not providea favorable environment for the study of relative chemical varia-tions and ultimately for the determination of the evolutionarystage although observations toward the objects have enabled usto understand the close relationship between the chemistry and theevolution of molecular clouds

Observations toward the hot core (MacDonald et al1996) andthe cold halo (Thompson amp MacDonald 1999) of the sourceG343+015 have revealed significant chemical variations be-tween these two distinct physical components within this objectA distinct chemistry was also noticed from observations of SO2CH3OH and H2CS (Helmich et al1994) and submillimeter mo-lecular line surveys (Helmich amp van Dishoeck 1997) of the W3region

Models of gas-phase ion-molecule chemistry can reasonablyreproduce the observed abundance of carbon-chain species in thecold dark clouds (Herbst et al 1984 Millar amp Freeman 1984)However such ion-molecule chemical models alone cannot sat-isfactorily account for the observed high abundance of mole-cules containing H or S atoms in star-forming molecular cloudsCoupling of gas-phase and grain-surface reactions (eg vanDishoeck amp Blake 1998) has been emphasized in recent chem-ical models in order to cope with the discrepancy Representativemodels have been developed for hot cores byMillar et al (1991)and by Charnley et al (1995) According to these models hydro-gen sulfur ionized and neutral molecules are readily depositedonto dust surfaces as the density increases during the collapsingphase of interstellar clouds The formation of complex heavy spe-cies containing hydrogen becomes inhibited leading to slow re-actions of those species in the earlier collapsing phase (t lt 104 yr)

1 Department of Astronomy and Space Science KyungHee UniversityYoungin Kyunggi 449-701 Korea sjkim1khuackr

2 School of Information Technology and Computer Science University ofWollongong NSW 2500 Australia

3 TaedukRadioAstronomyObservatory San36-1Whaam-DongYouseong-GuTaejeon 305-348 Korea

4 Department of Astronomy and Astrophysics Raman Research InstituteSadashivanagar Bangalore 560080 India

5 School of Physics University of New South Wales Sydney NSW 2052Australia

6 Department of Physics UMIST PO Box 88 Manchester M60 1QD UK

161

The Astrophysical Journal Supplement Series 162161ndash206 2006 January

2006 The American Astronomical Society All rights reserved Printed in USA

Radiation from the protostar then heats up the surrounding gas andgrains releasing volatile species from the dust surface back to gasphase This stage is called the hot core phase and lasts 104ndash105 yrafter the onset of star formation (van Dishoeck amp Blake 1998)

We surveyedW3 region to investigate the chemical evolution aspredicted by recent chemical models We also observed G343+015 in order to identify several missing lines from a previous sur-vey (Kim et al 2000) In addition high-spectral-resolution ob-servations for molecules of interest (CS HCO+ HCN HNCCH3CN and CH3CCH) were carried out in G343+015 TheW3region contains several objects at different evolutionary stagesandG343+015 is awell-known hot core associatedwith an ultra-compact H ii region Our TRAO observations toward W3 andG343+015 provide us with valuable information on interstellarchemistry because the chemical properties of these sources dif-fering in mass core size and evolutionary stage are useful for in-vestigating the relationships between the chemical variations andthese parameters

The contents of this paper are as follows x 1 presents back-ground information on line surveys evolutionary scenarios andmodels presented in the literature In x 2 general aspects of theW3 complex focusing on W3(OH) and W3 IRS 5 are given Inxx 3 and 4 we present the observational procedures line iden-tifications and spectral displays In x 5 we discuss individualspecies of interest detected in the survey In x 6 we discuss thechemical variations as well as the isotopic ratios of C and Satoms found from G343+015 W3(OH) and W3 IRS 5 In x 7the cloud morphology found from CS HCO+ HCN and HNCmappings is presented In x 8 chemical results from model cal-culations are given We summarize the results in x 9

2 W3 COMPLEX

The W3 complex consists of several infrared sources (Wynn-Williams amp Becklin 1974 Jaffe et al 1983 Colley 1980) at dif-ferent evolutionary stages and core masses The various objectsinclude well-developed shell structure in the H ii regions (W3BW3C and near IRS 3 and IRS 4) and several ultracompact H ii

regions (W3FW3M and near IRS 7 and IRS 5) Strong OH andH2O masers viz W3(OH) and W3(H2O) separated by 6

00 fromeach other are also observed

At least two hot cores have been revealed in W3 from obser-vations of NH3 lines (Mauersberger et al 1988) Its distance is

estimated to be 183 kpc residing in the Perseus arm of the gal-axy (Imai et al 2000) Helmich amp van Dishoeck (1997) per-formed spectral line surveys through the 345GHzwindow towardW3 IRS 4W3 IRS 5 andW3(OH) Significant differences in thephysical and chemical conditions were found in IRS 4 IRS 5 andW3(H2O) and the gas temperature seems highest (T 220 K)toward W3(H2O) compared with those in IRS 4 and IRS 5(Helmich et al 1994) High-velocity (V frac14 52 km s1) flowsnear W3 IRS 5 and an intermediate-velocity (V frac14 26 km s1)wing near W3(OH) are found in 12CO emission (Bally amp Lada1983)

21 W3(OH)

The ultracompact H ii regionW3(OH) is an extensively stud-ied site for the investigation of star formation and its influence onthe surrounding regions (Wyrowski et al 1999) Three peaks aredistinguished with subarcsecond resolution observations in thethermal dust continuum emission originating from the hot coreregion of W3(H2O Wyrowski et al 1999) Continuum obser-vations reveal thatW3(OH) has a shell structure in the H ii region(Dreher ampWelch 1981) H66 H76 and H110 radio recom-bination lines having full width at half-maximum (FWHM) of36 41 and 69 km s1 respectively are observed (Sams et al1996) H41 CH3CN [5(2)ndash4(2)] HCN (1ndash0) HCO+ (1ndash0)and 13CO (1ndash0) are detected with amplitudes of 04 01 59 55and 76 K respectively from observations with the Nobeyama45m telescope (Forster et al 1990) Keto et al (1992)mademid-infrared and molecular line observations of W3(OH) and com-pared them with G343+015

22 W3 IRS 5

W3 IRS 5 is a bright infrared source with a total luminosity of3 105 L (Campbell et al 1995) and is separated by7500 fromW3 IRS 4 which is brighter than W3 IRS 5 in the submillimeterwavelength range (Jaffe et al 1983) W3 IRS 5 is younger thanW3(OH) and appears to be at an earlier stage evolving toward ahot corelike object (Helmich et al 1994) Very long baseline in-terferometry measurement (Imai et al 2000) of radial velocitiesand propermotions ofwatermasers in theW3 IRS 5 region revealstwo groups of maser components which are associated with atleast two different outflows

TABLE 1

Reported Molecular Line Surveys

Sources Telescope FWHM

Frequency

(GHz)

Resolution

(MHz)

Line Density

(TransitionGHz) Observations

Orion A OSO 20 m 4700 72ndash91 1 85 Johansson et al (1984)

IRC +10216 23

Orion A OVRO 104 m 0A5 247ndash263 103 152 Blake et al (1986)

Sgr B2 BTL 7 m 2A9 70ndash150 1 914 Cummins et al (1986)

Orion A NRAO 12 m 2400 330ndash360 2 54 Jewell et al (1989)

Orion A OVRO 104 m 0A5 215ndash247 162 170 Sutton et al (1985)

Sgr B2(M) CSO 104 m 2000 330ndash355 1 67 Sutton et al (1991)

Sgr B2Orion-KL NRAO 11 m 8300 70ndash115 1 143171 Turner (1989)

G343+015 JCMT 15 m 1300 330ndash360 0756 06 Thompson amp MacDonald (1999)

G343+015 TRAO 14 m 5800 86ndash1653 1 20 Kim et al (2000 2001)

G343+015 JCMT 15 m 1400 330ndash360 033 114 MacDonald et al (1996)

G589039 JCMT 15 m 1300 330ndash360 0756 47 Thompson amp MacDonald (1999)

IRC +10216 IRAM 30 m 1900 129ndash173 1 86 Cernicharo et al (2000)

IRC +10216 Nobeyama 45 m 3500 28ndash50 025 86 Kawaguchi et al (1995)

Orion-KL FCRAO 14 m 4000 150ndash160 1 180 Ziurys amp MacGonagle (1993)

IRAS 174702853 Mopra 22 m 3000 90ndash96 05 35 Kim et al (2002)

KIM ET AL162 Vol 162

Strong 12CO self-absorption and 13COandHCNemissionpeaksare seen toward IRS 5 the maximum 12CO emission is observedtoward the compact H ii region and a CS peak is found 800 west ofthe OH1720 MHz maser (Dickel et al 1980) The possible pres-ence of disklike structure around a cluster of embedded stars in themid-infrared images of IRS 5 is inferred (Persi et al 1996) Thetotal H2 mass of the molecular cores associated with IRS 5 is esti-mated to be 1100 M derived from 13CO (1ndash0) and C18O (1ndash0)observations (Roberts et al 1997) A luminosity of 60 105 Land a cloud mass of 1600 M have been derived using 230 GHzcontinuum observations (Gordon 1987) for the entire complex ofW3

3 OBSERVATIONS

We carried out molecular line surveys toward W3(OH)and W3 IRS 5 using the Taeduk Radio Astronomy Observa-tory (TRAO) 14 m telescope between 2001 November and 2002February The receiver system was a low-noise SIS receiver oper-ating in single-sideband mode For the observations of W3(OH)the back ends were two filter banks in serial mode each of whichhas a 1MHz resolution and 256MHz bandwidth providing a totalbandwidth of 512 MHz at each local oscillator (LO) tuning Sen-sitive 250 kHz measurements toward W3(OH) and W3 IRS 5were also performed using two 256 channel filter banks in parallelmode after completing the 1MHz observations ofW3(OH) IRAS184500200 and IRAS 183350711

During the observations the typical system temperature was300 K between 85 and 110 GHz and was 600 K at 115 GHzPosition switching mode observations were carried out taking areference position to be 300 away in right ascension which wasconfirmed to be free of emission Calibration was performedusing the standard chopper-wheel method which corrects for at-mospheric attenuation and scattering of forward spillover andthe intensity was derived in terms of T

A scale (Ulich amp Hass1976) The beam sizes of the TRAO 14 m telescope are 6400 and5500 at 85 and 115 GHz respectively

The total on-source integration time of each LO tuning for the1MHz observation was 300 s achieving channel-to-channel rmsnoise of 003ndash005 K which is consistent with that in G343+015 (Kim et al 2000) Therefore these observations providedan opportunity to compare the chemical properties of theW3 andG343+015 regions with data taken with the same telescope sys-tem and sensitivity Pointing and focusing were checked using Cygni

Almost all of the molecular transitions previously detectedfrom the 1 MHz surveys toward G343+015 (Kim et al 2000)W3(OH) IRAS183350711 and IRAS184500200 (thiswork)were also considered for the 250 kHz surveys towardW3(OH) andW3 IRS 5 We also observed selected molecules (HCO+ CSHCN HNC and their isotopic variants) with 250 kHz resolutiontoward the ultracompact H ii region G343+015 to examine iso-topic ratios and the HCN hyperfine ratio in detail We measuredline emissions from SO2 HNCO CH3OH [0(0)ndash1(1) E ] andCH3OH [2(1)ndash1(1) A] transitions toward G343+015

The detectedmolecules and their transitions show close agree-ment betweenG343+015 and IRAS 184500200 and betweenW3(OH) and IRAS 183350711 suggesting that physicalchemical condition and evolutionary stage are similar To estimatephysical conditions as well as to determine the initial parametersfor chemical model calculations mapping observations were car-ried out towardW3(OH) for the transitions of CS (2ndash1) and HCN(1ndash0) over 100 100 HNC (1ndash0) over 80 80 and HCO+ (1ndash0)over 60 70 We carried out only CS (2ndash1) mapping toward W3

IRS 5 because the line strengths of other species are too weak tomap with the relatively large-beam size of TRAO Mapping ob-servations of CS (2ndash1) and HCO+ in IRAS 184500200 andIRAS183350711were performed covering almost the same areaas W3(OH) CS (2ndash1) mapping of IRAS 184500200 shows ir-regular and elongated distribution in theN-S directionwhereas forIRAS 183350711 spherical distribution is shown The CS (2ndash1)distribution is consistent with that of 86 GHz radio continuum(Walsh et al 1998) There is no report on continuummap for IRAS183350711Wewill present images and detailed analyses ofmo-lecular lines detected in IRAS 184500200 and IRAS 183350711 in future work

The source positions of W3(OH) and W3 IRS 5 and thecentral position of G343+015 are as follows

W3(OH )mdash frac14 02h23m16s9 frac14 6138056B4 (19500

Helmich et al 1994)W3 IRS 5mdash frac14 02h21m53s1 frac14 6152020B0 (19500

Helmich et al 1994)G343+015mdash frac14 18h50m46s3 frac14 0111013B0 (19500

Kim et al 2000)

4 DATA REDUCTION AND DISPLAY

41 Spectra and Images

Spectra taken with the 1 MHz resolution of the 512 channelsspectrometer in serial mode have been averaged and then base-line subtracted using the SPA data reduction package developedat the FCRAO observatory Some bad channels especially at theedge of the spectra were significant at the upper frequency regionwhen LSB mode observations were performed The number ofbad channels in the worst spectra is about 15ndash20 equivalentto 15ndash20 MHz The total bandwidth for every observation is512MHz but observations were carried out by stepping the cen-tral frequency by 500 MHz resulting in an overlap of approxi-mately 6 MHz between spectra

For the 250 kHz resolution observations the spectrometerswere configured in a parallel mode and an average of signalswas taken The typical noise in an averaged spectrum is 003ndash005 K for 1 MHz data and 006ndash009 K for 250 kHz data Thereduced spectra presented in Figures 1 and 2 were taken towardW3(OH) with the 1MHz and 250 kHz bandwidths respectivelyThe spectra in Figure 1 were obtained by combining two neigh-boring spectra giving a bandwidth of1 GHz In Figures 3 and4 250 kHz spectra taken from W3 IRS 5 and G343+015 re-spectively are presented

Data taken with mapping observations were made into imagesusing the Contour and Trigrid functions in IDL and then theimages were compared with those processed using CLASS Thetwo images which were drawn in terms of antenna temperatureshowed a good agreement with each other The contour maps inFigure 5 were made by taking signals stronger than 5 whichcorresponds to 04 K

42 Line Identification

Line identifications were made exclusively using the catalogof Lovas (1992) The catalog lists molecular transitions detectedfrom Orion-KL Sgr B2 and from various interstellar molecularclouds The number of detected lines in the catalog is sufficientfor the use in the line identifications of our 3 mm observationsThe identified line frequencies were determined by Gaussian fit-ting after correcting by the corresponding Doppler shifts of46and 39 km s1 for W3(OH) and W3 IRS 5 respectively andthen compared with the rest frequencies from the Lovas catalog

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 163No 1 2006

Fig 1mdashSpectra of W3(OH) from 848 to 1154 GHz obtained during 2001 NovemberndashDecember with the TRAO 14 m telescope Each panel combines twoneighboring frequency measurements to have about 1 GHz bandwidth

164

Fig 1mdashContinued

165

Fig 1mdashContinued

166

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 3: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

A MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 FROM 847 TO 1156 GHzOBSERVATIONAL DATA AND ANALYSES

Sang-Joon Kim1Hun-Dae Kim

2Youngung Lee

3Young Chol Minh

3Ramesh Balasubramanyam

4

Michael G Burton5Tom J Millar

6and Dong-Wook Lee

1

Received 2004 August 29 accepted 2005 May 31

ABSTRACT

Wehave carried out observations toward theW3 complex andG343+015 using the TRAO14m radio telescope toexamine in detail the chemical variations occurring while molecular clouds evolve from the prestellar to the H ii

region phase Observations include spectral surveys of these objects between 847 and 1156 GHz mapping obser-vations towardW3(OH) with the emissions of CS (2ndash1) HCN (1ndash0) HNC (1ndash0) and HCO+ (1ndash0) and mapping ofCS (2ndash1) emission toward W3 IRS 5 Chemical model calculations are used to estimate the age of W3(OH) bycomparing with the fractional abundances of detected molecules We found that G343+015 and W3(OH) are at asimilar evolutionary stage although large differences in the fractional abundances are found in CH3CN and HC3NOverall the properties of the detected species and abundances in three regions support the view that chemistry variesas molecular clouds evolve from a cold collapsing phase to a high-temperature phase such as the hot core and H ii

phase Chemical model calculations for W3(OH) indicate that the evolutionary age of the cloud is 104ndash105 yr withtemperature in the range 10ndash60 K

Subject headinggs H ii regions mdash ISM abundances mdash ISM individual (W3(OH) W3 IRS 5 G343+015)

1 INTRODUCTION

It is known that chemical properties of molecular clouds varyduring their evolution The onset of central protostars comparedwith the cold dark collapsing phase plays a critical role in driv-ing the chemistry in the surrounding dense region by providinga high temperature and density Molecular line surveys amongmany observational techniques are one of the most powerfulmethods for the investigation of the chemistry The identifiedmolecules which are sensitive to various chemical reactions inthemolecular clouds can be used to investigate the formation andevolution of molecular clouds Representative line surveys per-formed over the last few decades are listed in Table 1 Valuableinformation on interstellar chemistry and the evolution of mo-lecular clouds has been obtained from the surveys However asshown in Table 1 they have been limited to a few objects such asSgr B2 (Cummins et al1986 Turner1989 Sutton et al1991) andthe nearby Orion-KL region (Johansson et al 1984 Jewell et al1989 Turner 1989 Blake et al 1986 1996 ZiurysampMcGonagle1993 Schilke et al 1997)

Recently other sources have been chosen for molecular linesurveys the ultracompact H ii region G343+015 (MacDonaldet al 1996 Kim et al 2000) the low-mass star-forming regionIRAS 162932422 (van Dishoeck et al 1995) the protostellarregion IRAS 174702853 (Kim et al 2002) and the late-typestar IRC +10216 (Kawaguchi et al1995 Cernicharo et al 2000)It is important to study the sources for which significant chemicalvariations are expected as predicted by models Detailed ex-

amination has been undertaken for individual species present inOrion-KL and Sgr B2 (Turner 1991) in order to test the modelsAlthough the results toward these two sources have provided uswith interesting information on the nature of dense cores the com-plex physical structure of Orion-KL and Sgr B2 has limited ourunderstanding of the chemical processes The molecular abun-dances derived from these objects contain large uncertainties dueto the blending of adjacent transitions compared to low-mass andisolated objects Therefore Orion-KL and Sgr B2 do not providea favorable environment for the study of relative chemical varia-tions and ultimately for the determination of the evolutionarystage although observations toward the objects have enabled usto understand the close relationship between the chemistry and theevolution of molecular clouds

Observations toward the hot core (MacDonald et al1996) andthe cold halo (Thompson amp MacDonald 1999) of the sourceG343+015 have revealed significant chemical variations be-tween these two distinct physical components within this objectA distinct chemistry was also noticed from observations of SO2CH3OH and H2CS (Helmich et al1994) and submillimeter mo-lecular line surveys (Helmich amp van Dishoeck 1997) of the W3region

Models of gas-phase ion-molecule chemistry can reasonablyreproduce the observed abundance of carbon-chain species in thecold dark clouds (Herbst et al 1984 Millar amp Freeman 1984)However such ion-molecule chemical models alone cannot sat-isfactorily account for the observed high abundance of mole-cules containing H or S atoms in star-forming molecular cloudsCoupling of gas-phase and grain-surface reactions (eg vanDishoeck amp Blake 1998) has been emphasized in recent chem-ical models in order to cope with the discrepancy Representativemodels have been developed for hot cores byMillar et al (1991)and by Charnley et al (1995) According to these models hydro-gen sulfur ionized and neutral molecules are readily depositedonto dust surfaces as the density increases during the collapsingphase of interstellar clouds The formation of complex heavy spe-cies containing hydrogen becomes inhibited leading to slow re-actions of those species in the earlier collapsing phase (t lt 104 yr)

1 Department of Astronomy and Space Science KyungHee UniversityYoungin Kyunggi 449-701 Korea sjkim1khuackr

2 School of Information Technology and Computer Science University ofWollongong NSW 2500 Australia

3 TaedukRadioAstronomyObservatory San36-1Whaam-DongYouseong-GuTaejeon 305-348 Korea

4 Department of Astronomy and Astrophysics Raman Research InstituteSadashivanagar Bangalore 560080 India

5 School of Physics University of New South Wales Sydney NSW 2052Australia

6 Department of Physics UMIST PO Box 88 Manchester M60 1QD UK

161

The Astrophysical Journal Supplement Series 162161ndash206 2006 January

2006 The American Astronomical Society All rights reserved Printed in USA

Radiation from the protostar then heats up the surrounding gas andgrains releasing volatile species from the dust surface back to gasphase This stage is called the hot core phase and lasts 104ndash105 yrafter the onset of star formation (van Dishoeck amp Blake 1998)

We surveyedW3 region to investigate the chemical evolution aspredicted by recent chemical models We also observed G343+015 in order to identify several missing lines from a previous sur-vey (Kim et al 2000) In addition high-spectral-resolution ob-servations for molecules of interest (CS HCO+ HCN HNCCH3CN and CH3CCH) were carried out in G343+015 TheW3region contains several objects at different evolutionary stagesandG343+015 is awell-known hot core associatedwith an ultra-compact H ii region Our TRAO observations toward W3 andG343+015 provide us with valuable information on interstellarchemistry because the chemical properties of these sources dif-fering in mass core size and evolutionary stage are useful for in-vestigating the relationships between the chemical variations andthese parameters

The contents of this paper are as follows x 1 presents back-ground information on line surveys evolutionary scenarios andmodels presented in the literature In x 2 general aspects of theW3 complex focusing on W3(OH) and W3 IRS 5 are given Inxx 3 and 4 we present the observational procedures line iden-tifications and spectral displays In x 5 we discuss individualspecies of interest detected in the survey In x 6 we discuss thechemical variations as well as the isotopic ratios of C and Satoms found from G343+015 W3(OH) and W3 IRS 5 In x 7the cloud morphology found from CS HCO+ HCN and HNCmappings is presented In x 8 chemical results from model cal-culations are given We summarize the results in x 9

2 W3 COMPLEX

The W3 complex consists of several infrared sources (Wynn-Williams amp Becklin 1974 Jaffe et al 1983 Colley 1980) at dif-ferent evolutionary stages and core masses The various objectsinclude well-developed shell structure in the H ii regions (W3BW3C and near IRS 3 and IRS 4) and several ultracompact H ii

regions (W3FW3M and near IRS 7 and IRS 5) Strong OH andH2O masers viz W3(OH) and W3(H2O) separated by 6

00 fromeach other are also observed

At least two hot cores have been revealed in W3 from obser-vations of NH3 lines (Mauersberger et al 1988) Its distance is

estimated to be 183 kpc residing in the Perseus arm of the gal-axy (Imai et al 2000) Helmich amp van Dishoeck (1997) per-formed spectral line surveys through the 345GHzwindow towardW3 IRS 4W3 IRS 5 andW3(OH) Significant differences in thephysical and chemical conditions were found in IRS 4 IRS 5 andW3(H2O) and the gas temperature seems highest (T 220 K)toward W3(H2O) compared with those in IRS 4 and IRS 5(Helmich et al 1994) High-velocity (V frac14 52 km s1) flowsnear W3 IRS 5 and an intermediate-velocity (V frac14 26 km s1)wing near W3(OH) are found in 12CO emission (Bally amp Lada1983)

21 W3(OH)

The ultracompact H ii regionW3(OH) is an extensively stud-ied site for the investigation of star formation and its influence onthe surrounding regions (Wyrowski et al 1999) Three peaks aredistinguished with subarcsecond resolution observations in thethermal dust continuum emission originating from the hot coreregion of W3(H2O Wyrowski et al 1999) Continuum obser-vations reveal thatW3(OH) has a shell structure in the H ii region(Dreher ampWelch 1981) H66 H76 and H110 radio recom-bination lines having full width at half-maximum (FWHM) of36 41 and 69 km s1 respectively are observed (Sams et al1996) H41 CH3CN [5(2)ndash4(2)] HCN (1ndash0) HCO+ (1ndash0)and 13CO (1ndash0) are detected with amplitudes of 04 01 59 55and 76 K respectively from observations with the Nobeyama45m telescope (Forster et al 1990) Keto et al (1992)mademid-infrared and molecular line observations of W3(OH) and com-pared them with G343+015

22 W3 IRS 5

W3 IRS 5 is a bright infrared source with a total luminosity of3 105 L (Campbell et al 1995) and is separated by7500 fromW3 IRS 4 which is brighter than W3 IRS 5 in the submillimeterwavelength range (Jaffe et al 1983) W3 IRS 5 is younger thanW3(OH) and appears to be at an earlier stage evolving toward ahot corelike object (Helmich et al 1994) Very long baseline in-terferometry measurement (Imai et al 2000) of radial velocitiesand propermotions ofwatermasers in theW3 IRS 5 region revealstwo groups of maser components which are associated with atleast two different outflows

TABLE 1

Reported Molecular Line Surveys

Sources Telescope FWHM

Frequency

(GHz)

Resolution

(MHz)

Line Density

(TransitionGHz) Observations

Orion A OSO 20 m 4700 72ndash91 1 85 Johansson et al (1984)

IRC +10216 23

Orion A OVRO 104 m 0A5 247ndash263 103 152 Blake et al (1986)

Sgr B2 BTL 7 m 2A9 70ndash150 1 914 Cummins et al (1986)

Orion A NRAO 12 m 2400 330ndash360 2 54 Jewell et al (1989)

Orion A OVRO 104 m 0A5 215ndash247 162 170 Sutton et al (1985)

Sgr B2(M) CSO 104 m 2000 330ndash355 1 67 Sutton et al (1991)

Sgr B2Orion-KL NRAO 11 m 8300 70ndash115 1 143171 Turner (1989)

G343+015 JCMT 15 m 1300 330ndash360 0756 06 Thompson amp MacDonald (1999)

G343+015 TRAO 14 m 5800 86ndash1653 1 20 Kim et al (2000 2001)

G343+015 JCMT 15 m 1400 330ndash360 033 114 MacDonald et al (1996)

G589039 JCMT 15 m 1300 330ndash360 0756 47 Thompson amp MacDonald (1999)

IRC +10216 IRAM 30 m 1900 129ndash173 1 86 Cernicharo et al (2000)

IRC +10216 Nobeyama 45 m 3500 28ndash50 025 86 Kawaguchi et al (1995)

Orion-KL FCRAO 14 m 4000 150ndash160 1 180 Ziurys amp MacGonagle (1993)

IRAS 174702853 Mopra 22 m 3000 90ndash96 05 35 Kim et al (2002)

KIM ET AL162 Vol 162

Strong 12CO self-absorption and 13COandHCNemissionpeaksare seen toward IRS 5 the maximum 12CO emission is observedtoward the compact H ii region and a CS peak is found 800 west ofthe OH1720 MHz maser (Dickel et al 1980) The possible pres-ence of disklike structure around a cluster of embedded stars in themid-infrared images of IRS 5 is inferred (Persi et al 1996) Thetotal H2 mass of the molecular cores associated with IRS 5 is esti-mated to be 1100 M derived from 13CO (1ndash0) and C18O (1ndash0)observations (Roberts et al 1997) A luminosity of 60 105 Land a cloud mass of 1600 M have been derived using 230 GHzcontinuum observations (Gordon 1987) for the entire complex ofW3

3 OBSERVATIONS

We carried out molecular line surveys toward W3(OH)and W3 IRS 5 using the Taeduk Radio Astronomy Observa-tory (TRAO) 14 m telescope between 2001 November and 2002February The receiver system was a low-noise SIS receiver oper-ating in single-sideband mode For the observations of W3(OH)the back ends were two filter banks in serial mode each of whichhas a 1MHz resolution and 256MHz bandwidth providing a totalbandwidth of 512 MHz at each local oscillator (LO) tuning Sen-sitive 250 kHz measurements toward W3(OH) and W3 IRS 5were also performed using two 256 channel filter banks in parallelmode after completing the 1MHz observations ofW3(OH) IRAS184500200 and IRAS 183350711

During the observations the typical system temperature was300 K between 85 and 110 GHz and was 600 K at 115 GHzPosition switching mode observations were carried out taking areference position to be 300 away in right ascension which wasconfirmed to be free of emission Calibration was performedusing the standard chopper-wheel method which corrects for at-mospheric attenuation and scattering of forward spillover andthe intensity was derived in terms of T

A scale (Ulich amp Hass1976) The beam sizes of the TRAO 14 m telescope are 6400 and5500 at 85 and 115 GHz respectively

The total on-source integration time of each LO tuning for the1MHz observation was 300 s achieving channel-to-channel rmsnoise of 003ndash005 K which is consistent with that in G343+015 (Kim et al 2000) Therefore these observations providedan opportunity to compare the chemical properties of theW3 andG343+015 regions with data taken with the same telescope sys-tem and sensitivity Pointing and focusing were checked using Cygni

Almost all of the molecular transitions previously detectedfrom the 1 MHz surveys toward G343+015 (Kim et al 2000)W3(OH) IRAS183350711 and IRAS184500200 (thiswork)were also considered for the 250 kHz surveys towardW3(OH) andW3 IRS 5 We also observed selected molecules (HCO+ CSHCN HNC and their isotopic variants) with 250 kHz resolutiontoward the ultracompact H ii region G343+015 to examine iso-topic ratios and the HCN hyperfine ratio in detail We measuredline emissions from SO2 HNCO CH3OH [0(0)ndash1(1) E ] andCH3OH [2(1)ndash1(1) A] transitions toward G343+015

The detectedmolecules and their transitions show close agree-ment betweenG343+015 and IRAS 184500200 and betweenW3(OH) and IRAS 183350711 suggesting that physicalchemical condition and evolutionary stage are similar To estimatephysical conditions as well as to determine the initial parametersfor chemical model calculations mapping observations were car-ried out towardW3(OH) for the transitions of CS (2ndash1) and HCN(1ndash0) over 100 100 HNC (1ndash0) over 80 80 and HCO+ (1ndash0)over 60 70 We carried out only CS (2ndash1) mapping toward W3

IRS 5 because the line strengths of other species are too weak tomap with the relatively large-beam size of TRAO Mapping ob-servations of CS (2ndash1) and HCO+ in IRAS 184500200 andIRAS183350711were performed covering almost the same areaas W3(OH) CS (2ndash1) mapping of IRAS 184500200 shows ir-regular and elongated distribution in theN-S directionwhereas forIRAS 183350711 spherical distribution is shown The CS (2ndash1)distribution is consistent with that of 86 GHz radio continuum(Walsh et al 1998) There is no report on continuummap for IRAS183350711Wewill present images and detailed analyses ofmo-lecular lines detected in IRAS 184500200 and IRAS 183350711 in future work

The source positions of W3(OH) and W3 IRS 5 and thecentral position of G343+015 are as follows

W3(OH )mdash frac14 02h23m16s9 frac14 6138056B4 (19500

Helmich et al 1994)W3 IRS 5mdash frac14 02h21m53s1 frac14 6152020B0 (19500

Helmich et al 1994)G343+015mdash frac14 18h50m46s3 frac14 0111013B0 (19500

Kim et al 2000)

4 DATA REDUCTION AND DISPLAY

41 Spectra and Images

Spectra taken with the 1 MHz resolution of the 512 channelsspectrometer in serial mode have been averaged and then base-line subtracted using the SPA data reduction package developedat the FCRAO observatory Some bad channels especially at theedge of the spectra were significant at the upper frequency regionwhen LSB mode observations were performed The number ofbad channels in the worst spectra is about 15ndash20 equivalentto 15ndash20 MHz The total bandwidth for every observation is512MHz but observations were carried out by stepping the cen-tral frequency by 500 MHz resulting in an overlap of approxi-mately 6 MHz between spectra

For the 250 kHz resolution observations the spectrometerswere configured in a parallel mode and an average of signalswas taken The typical noise in an averaged spectrum is 003ndash005 K for 1 MHz data and 006ndash009 K for 250 kHz data Thereduced spectra presented in Figures 1 and 2 were taken towardW3(OH) with the 1MHz and 250 kHz bandwidths respectivelyThe spectra in Figure 1 were obtained by combining two neigh-boring spectra giving a bandwidth of1 GHz In Figures 3 and4 250 kHz spectra taken from W3 IRS 5 and G343+015 re-spectively are presented

Data taken with mapping observations were made into imagesusing the Contour and Trigrid functions in IDL and then theimages were compared with those processed using CLASS Thetwo images which were drawn in terms of antenna temperatureshowed a good agreement with each other The contour maps inFigure 5 were made by taking signals stronger than 5 whichcorresponds to 04 K

42 Line Identification

Line identifications were made exclusively using the catalogof Lovas (1992) The catalog lists molecular transitions detectedfrom Orion-KL Sgr B2 and from various interstellar molecularclouds The number of detected lines in the catalog is sufficientfor the use in the line identifications of our 3 mm observationsThe identified line frequencies were determined by Gaussian fit-ting after correcting by the corresponding Doppler shifts of46and 39 km s1 for W3(OH) and W3 IRS 5 respectively andthen compared with the rest frequencies from the Lovas catalog

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 163No 1 2006

Fig 1mdashSpectra of W3(OH) from 848 to 1154 GHz obtained during 2001 NovemberndashDecember with the TRAO 14 m telescope Each panel combines twoneighboring frequency measurements to have about 1 GHz bandwidth

164

Fig 1mdashContinued

165

Fig 1mdashContinued

166

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 4: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Radiation from the protostar then heats up the surrounding gas andgrains releasing volatile species from the dust surface back to gasphase This stage is called the hot core phase and lasts 104ndash105 yrafter the onset of star formation (van Dishoeck amp Blake 1998)

We surveyedW3 region to investigate the chemical evolution aspredicted by recent chemical models We also observed G343+015 in order to identify several missing lines from a previous sur-vey (Kim et al 2000) In addition high-spectral-resolution ob-servations for molecules of interest (CS HCO+ HCN HNCCH3CN and CH3CCH) were carried out in G343+015 TheW3region contains several objects at different evolutionary stagesandG343+015 is awell-known hot core associatedwith an ultra-compact H ii region Our TRAO observations toward W3 andG343+015 provide us with valuable information on interstellarchemistry because the chemical properties of these sources dif-fering in mass core size and evolutionary stage are useful for in-vestigating the relationships between the chemical variations andthese parameters

The contents of this paper are as follows x 1 presents back-ground information on line surveys evolutionary scenarios andmodels presented in the literature In x 2 general aspects of theW3 complex focusing on W3(OH) and W3 IRS 5 are given Inxx 3 and 4 we present the observational procedures line iden-tifications and spectral displays In x 5 we discuss individualspecies of interest detected in the survey In x 6 we discuss thechemical variations as well as the isotopic ratios of C and Satoms found from G343+015 W3(OH) and W3 IRS 5 In x 7the cloud morphology found from CS HCO+ HCN and HNCmappings is presented In x 8 chemical results from model cal-culations are given We summarize the results in x 9

2 W3 COMPLEX

The W3 complex consists of several infrared sources (Wynn-Williams amp Becklin 1974 Jaffe et al 1983 Colley 1980) at dif-ferent evolutionary stages and core masses The various objectsinclude well-developed shell structure in the H ii regions (W3BW3C and near IRS 3 and IRS 4) and several ultracompact H ii

regions (W3FW3M and near IRS 7 and IRS 5) Strong OH andH2O masers viz W3(OH) and W3(H2O) separated by 6

00 fromeach other are also observed

At least two hot cores have been revealed in W3 from obser-vations of NH3 lines (Mauersberger et al 1988) Its distance is

estimated to be 183 kpc residing in the Perseus arm of the gal-axy (Imai et al 2000) Helmich amp van Dishoeck (1997) per-formed spectral line surveys through the 345GHzwindow towardW3 IRS 4W3 IRS 5 andW3(OH) Significant differences in thephysical and chemical conditions were found in IRS 4 IRS 5 andW3(H2O) and the gas temperature seems highest (T 220 K)toward W3(H2O) compared with those in IRS 4 and IRS 5(Helmich et al 1994) High-velocity (V frac14 52 km s1) flowsnear W3 IRS 5 and an intermediate-velocity (V frac14 26 km s1)wing near W3(OH) are found in 12CO emission (Bally amp Lada1983)

21 W3(OH)

The ultracompact H ii regionW3(OH) is an extensively stud-ied site for the investigation of star formation and its influence onthe surrounding regions (Wyrowski et al 1999) Three peaks aredistinguished with subarcsecond resolution observations in thethermal dust continuum emission originating from the hot coreregion of W3(H2O Wyrowski et al 1999) Continuum obser-vations reveal thatW3(OH) has a shell structure in the H ii region(Dreher ampWelch 1981) H66 H76 and H110 radio recom-bination lines having full width at half-maximum (FWHM) of36 41 and 69 km s1 respectively are observed (Sams et al1996) H41 CH3CN [5(2)ndash4(2)] HCN (1ndash0) HCO+ (1ndash0)and 13CO (1ndash0) are detected with amplitudes of 04 01 59 55and 76 K respectively from observations with the Nobeyama45m telescope (Forster et al 1990) Keto et al (1992)mademid-infrared and molecular line observations of W3(OH) and com-pared them with G343+015

22 W3 IRS 5

W3 IRS 5 is a bright infrared source with a total luminosity of3 105 L (Campbell et al 1995) and is separated by7500 fromW3 IRS 4 which is brighter than W3 IRS 5 in the submillimeterwavelength range (Jaffe et al 1983) W3 IRS 5 is younger thanW3(OH) and appears to be at an earlier stage evolving toward ahot corelike object (Helmich et al 1994) Very long baseline in-terferometry measurement (Imai et al 2000) of radial velocitiesand propermotions ofwatermasers in theW3 IRS 5 region revealstwo groups of maser components which are associated with atleast two different outflows

TABLE 1

Reported Molecular Line Surveys

Sources Telescope FWHM

Frequency

(GHz)

Resolution

(MHz)

Line Density

(TransitionGHz) Observations

Orion A OSO 20 m 4700 72ndash91 1 85 Johansson et al (1984)

IRC +10216 23

Orion A OVRO 104 m 0A5 247ndash263 103 152 Blake et al (1986)

Sgr B2 BTL 7 m 2A9 70ndash150 1 914 Cummins et al (1986)

Orion A NRAO 12 m 2400 330ndash360 2 54 Jewell et al (1989)

Orion A OVRO 104 m 0A5 215ndash247 162 170 Sutton et al (1985)

Sgr B2(M) CSO 104 m 2000 330ndash355 1 67 Sutton et al (1991)

Sgr B2Orion-KL NRAO 11 m 8300 70ndash115 1 143171 Turner (1989)

G343+015 JCMT 15 m 1300 330ndash360 0756 06 Thompson amp MacDonald (1999)

G343+015 TRAO 14 m 5800 86ndash1653 1 20 Kim et al (2000 2001)

G343+015 JCMT 15 m 1400 330ndash360 033 114 MacDonald et al (1996)

G589039 JCMT 15 m 1300 330ndash360 0756 47 Thompson amp MacDonald (1999)

IRC +10216 IRAM 30 m 1900 129ndash173 1 86 Cernicharo et al (2000)

IRC +10216 Nobeyama 45 m 3500 28ndash50 025 86 Kawaguchi et al (1995)

Orion-KL FCRAO 14 m 4000 150ndash160 1 180 Ziurys amp MacGonagle (1993)

IRAS 174702853 Mopra 22 m 3000 90ndash96 05 35 Kim et al (2002)

KIM ET AL162 Vol 162

Strong 12CO self-absorption and 13COandHCNemissionpeaksare seen toward IRS 5 the maximum 12CO emission is observedtoward the compact H ii region and a CS peak is found 800 west ofthe OH1720 MHz maser (Dickel et al 1980) The possible pres-ence of disklike structure around a cluster of embedded stars in themid-infrared images of IRS 5 is inferred (Persi et al 1996) Thetotal H2 mass of the molecular cores associated with IRS 5 is esti-mated to be 1100 M derived from 13CO (1ndash0) and C18O (1ndash0)observations (Roberts et al 1997) A luminosity of 60 105 Land a cloud mass of 1600 M have been derived using 230 GHzcontinuum observations (Gordon 1987) for the entire complex ofW3

3 OBSERVATIONS

We carried out molecular line surveys toward W3(OH)and W3 IRS 5 using the Taeduk Radio Astronomy Observa-tory (TRAO) 14 m telescope between 2001 November and 2002February The receiver system was a low-noise SIS receiver oper-ating in single-sideband mode For the observations of W3(OH)the back ends were two filter banks in serial mode each of whichhas a 1MHz resolution and 256MHz bandwidth providing a totalbandwidth of 512 MHz at each local oscillator (LO) tuning Sen-sitive 250 kHz measurements toward W3(OH) and W3 IRS 5were also performed using two 256 channel filter banks in parallelmode after completing the 1MHz observations ofW3(OH) IRAS184500200 and IRAS 183350711

During the observations the typical system temperature was300 K between 85 and 110 GHz and was 600 K at 115 GHzPosition switching mode observations were carried out taking areference position to be 300 away in right ascension which wasconfirmed to be free of emission Calibration was performedusing the standard chopper-wheel method which corrects for at-mospheric attenuation and scattering of forward spillover andthe intensity was derived in terms of T

A scale (Ulich amp Hass1976) The beam sizes of the TRAO 14 m telescope are 6400 and5500 at 85 and 115 GHz respectively

The total on-source integration time of each LO tuning for the1MHz observation was 300 s achieving channel-to-channel rmsnoise of 003ndash005 K which is consistent with that in G343+015 (Kim et al 2000) Therefore these observations providedan opportunity to compare the chemical properties of theW3 andG343+015 regions with data taken with the same telescope sys-tem and sensitivity Pointing and focusing were checked using Cygni

Almost all of the molecular transitions previously detectedfrom the 1 MHz surveys toward G343+015 (Kim et al 2000)W3(OH) IRAS183350711 and IRAS184500200 (thiswork)were also considered for the 250 kHz surveys towardW3(OH) andW3 IRS 5 We also observed selected molecules (HCO+ CSHCN HNC and their isotopic variants) with 250 kHz resolutiontoward the ultracompact H ii region G343+015 to examine iso-topic ratios and the HCN hyperfine ratio in detail We measuredline emissions from SO2 HNCO CH3OH [0(0)ndash1(1) E ] andCH3OH [2(1)ndash1(1) A] transitions toward G343+015

The detectedmolecules and their transitions show close agree-ment betweenG343+015 and IRAS 184500200 and betweenW3(OH) and IRAS 183350711 suggesting that physicalchemical condition and evolutionary stage are similar To estimatephysical conditions as well as to determine the initial parametersfor chemical model calculations mapping observations were car-ried out towardW3(OH) for the transitions of CS (2ndash1) and HCN(1ndash0) over 100 100 HNC (1ndash0) over 80 80 and HCO+ (1ndash0)over 60 70 We carried out only CS (2ndash1) mapping toward W3

IRS 5 because the line strengths of other species are too weak tomap with the relatively large-beam size of TRAO Mapping ob-servations of CS (2ndash1) and HCO+ in IRAS 184500200 andIRAS183350711were performed covering almost the same areaas W3(OH) CS (2ndash1) mapping of IRAS 184500200 shows ir-regular and elongated distribution in theN-S directionwhereas forIRAS 183350711 spherical distribution is shown The CS (2ndash1)distribution is consistent with that of 86 GHz radio continuum(Walsh et al 1998) There is no report on continuummap for IRAS183350711Wewill present images and detailed analyses ofmo-lecular lines detected in IRAS 184500200 and IRAS 183350711 in future work

The source positions of W3(OH) and W3 IRS 5 and thecentral position of G343+015 are as follows

W3(OH )mdash frac14 02h23m16s9 frac14 6138056B4 (19500

Helmich et al 1994)W3 IRS 5mdash frac14 02h21m53s1 frac14 6152020B0 (19500

Helmich et al 1994)G343+015mdash frac14 18h50m46s3 frac14 0111013B0 (19500

Kim et al 2000)

4 DATA REDUCTION AND DISPLAY

41 Spectra and Images

Spectra taken with the 1 MHz resolution of the 512 channelsspectrometer in serial mode have been averaged and then base-line subtracted using the SPA data reduction package developedat the FCRAO observatory Some bad channels especially at theedge of the spectra were significant at the upper frequency regionwhen LSB mode observations were performed The number ofbad channels in the worst spectra is about 15ndash20 equivalentto 15ndash20 MHz The total bandwidth for every observation is512MHz but observations were carried out by stepping the cen-tral frequency by 500 MHz resulting in an overlap of approxi-mately 6 MHz between spectra

For the 250 kHz resolution observations the spectrometerswere configured in a parallel mode and an average of signalswas taken The typical noise in an averaged spectrum is 003ndash005 K for 1 MHz data and 006ndash009 K for 250 kHz data Thereduced spectra presented in Figures 1 and 2 were taken towardW3(OH) with the 1MHz and 250 kHz bandwidths respectivelyThe spectra in Figure 1 were obtained by combining two neigh-boring spectra giving a bandwidth of1 GHz In Figures 3 and4 250 kHz spectra taken from W3 IRS 5 and G343+015 re-spectively are presented

Data taken with mapping observations were made into imagesusing the Contour and Trigrid functions in IDL and then theimages were compared with those processed using CLASS Thetwo images which were drawn in terms of antenna temperatureshowed a good agreement with each other The contour maps inFigure 5 were made by taking signals stronger than 5 whichcorresponds to 04 K

42 Line Identification

Line identifications were made exclusively using the catalogof Lovas (1992) The catalog lists molecular transitions detectedfrom Orion-KL Sgr B2 and from various interstellar molecularclouds The number of detected lines in the catalog is sufficientfor the use in the line identifications of our 3 mm observationsThe identified line frequencies were determined by Gaussian fit-ting after correcting by the corresponding Doppler shifts of46and 39 km s1 for W3(OH) and W3 IRS 5 respectively andthen compared with the rest frequencies from the Lovas catalog

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 163No 1 2006

Fig 1mdashSpectra of W3(OH) from 848 to 1154 GHz obtained during 2001 NovemberndashDecember with the TRAO 14 m telescope Each panel combines twoneighboring frequency measurements to have about 1 GHz bandwidth

164

Fig 1mdashContinued

165

Fig 1mdashContinued

166

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 5: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Strong 12CO self-absorption and 13COandHCNemissionpeaksare seen toward IRS 5 the maximum 12CO emission is observedtoward the compact H ii region and a CS peak is found 800 west ofthe OH1720 MHz maser (Dickel et al 1980) The possible pres-ence of disklike structure around a cluster of embedded stars in themid-infrared images of IRS 5 is inferred (Persi et al 1996) Thetotal H2 mass of the molecular cores associated with IRS 5 is esti-mated to be 1100 M derived from 13CO (1ndash0) and C18O (1ndash0)observations (Roberts et al 1997) A luminosity of 60 105 Land a cloud mass of 1600 M have been derived using 230 GHzcontinuum observations (Gordon 1987) for the entire complex ofW3

3 OBSERVATIONS

We carried out molecular line surveys toward W3(OH)and W3 IRS 5 using the Taeduk Radio Astronomy Observa-tory (TRAO) 14 m telescope between 2001 November and 2002February The receiver system was a low-noise SIS receiver oper-ating in single-sideband mode For the observations of W3(OH)the back ends were two filter banks in serial mode each of whichhas a 1MHz resolution and 256MHz bandwidth providing a totalbandwidth of 512 MHz at each local oscillator (LO) tuning Sen-sitive 250 kHz measurements toward W3(OH) and W3 IRS 5were also performed using two 256 channel filter banks in parallelmode after completing the 1MHz observations ofW3(OH) IRAS184500200 and IRAS 183350711

During the observations the typical system temperature was300 K between 85 and 110 GHz and was 600 K at 115 GHzPosition switching mode observations were carried out taking areference position to be 300 away in right ascension which wasconfirmed to be free of emission Calibration was performedusing the standard chopper-wheel method which corrects for at-mospheric attenuation and scattering of forward spillover andthe intensity was derived in terms of T

A scale (Ulich amp Hass1976) The beam sizes of the TRAO 14 m telescope are 6400 and5500 at 85 and 115 GHz respectively

The total on-source integration time of each LO tuning for the1MHz observation was 300 s achieving channel-to-channel rmsnoise of 003ndash005 K which is consistent with that in G343+015 (Kim et al 2000) Therefore these observations providedan opportunity to compare the chemical properties of theW3 andG343+015 regions with data taken with the same telescope sys-tem and sensitivity Pointing and focusing were checked using Cygni

Almost all of the molecular transitions previously detectedfrom the 1 MHz surveys toward G343+015 (Kim et al 2000)W3(OH) IRAS183350711 and IRAS184500200 (thiswork)were also considered for the 250 kHz surveys towardW3(OH) andW3 IRS 5 We also observed selected molecules (HCO+ CSHCN HNC and their isotopic variants) with 250 kHz resolutiontoward the ultracompact H ii region G343+015 to examine iso-topic ratios and the HCN hyperfine ratio in detail We measuredline emissions from SO2 HNCO CH3OH [0(0)ndash1(1) E ] andCH3OH [2(1)ndash1(1) A] transitions toward G343+015

The detectedmolecules and their transitions show close agree-ment betweenG343+015 and IRAS 184500200 and betweenW3(OH) and IRAS 183350711 suggesting that physicalchemical condition and evolutionary stage are similar To estimatephysical conditions as well as to determine the initial parametersfor chemical model calculations mapping observations were car-ried out towardW3(OH) for the transitions of CS (2ndash1) and HCN(1ndash0) over 100 100 HNC (1ndash0) over 80 80 and HCO+ (1ndash0)over 60 70 We carried out only CS (2ndash1) mapping toward W3

IRS 5 because the line strengths of other species are too weak tomap with the relatively large-beam size of TRAO Mapping ob-servations of CS (2ndash1) and HCO+ in IRAS 184500200 andIRAS183350711were performed covering almost the same areaas W3(OH) CS (2ndash1) mapping of IRAS 184500200 shows ir-regular and elongated distribution in theN-S directionwhereas forIRAS 183350711 spherical distribution is shown The CS (2ndash1)distribution is consistent with that of 86 GHz radio continuum(Walsh et al 1998) There is no report on continuummap for IRAS183350711Wewill present images and detailed analyses ofmo-lecular lines detected in IRAS 184500200 and IRAS 183350711 in future work

The source positions of W3(OH) and W3 IRS 5 and thecentral position of G343+015 are as follows

W3(OH )mdash frac14 02h23m16s9 frac14 6138056B4 (19500

Helmich et al 1994)W3 IRS 5mdash frac14 02h21m53s1 frac14 6152020B0 (19500

Helmich et al 1994)G343+015mdash frac14 18h50m46s3 frac14 0111013B0 (19500

Kim et al 2000)

4 DATA REDUCTION AND DISPLAY

41 Spectra and Images

Spectra taken with the 1 MHz resolution of the 512 channelsspectrometer in serial mode have been averaged and then base-line subtracted using the SPA data reduction package developedat the FCRAO observatory Some bad channels especially at theedge of the spectra were significant at the upper frequency regionwhen LSB mode observations were performed The number ofbad channels in the worst spectra is about 15ndash20 equivalentto 15ndash20 MHz The total bandwidth for every observation is512MHz but observations were carried out by stepping the cen-tral frequency by 500 MHz resulting in an overlap of approxi-mately 6 MHz between spectra

For the 250 kHz resolution observations the spectrometerswere configured in a parallel mode and an average of signalswas taken The typical noise in an averaged spectrum is 003ndash005 K for 1 MHz data and 006ndash009 K for 250 kHz data Thereduced spectra presented in Figures 1 and 2 were taken towardW3(OH) with the 1MHz and 250 kHz bandwidths respectivelyThe spectra in Figure 1 were obtained by combining two neigh-boring spectra giving a bandwidth of1 GHz In Figures 3 and4 250 kHz spectra taken from W3 IRS 5 and G343+015 re-spectively are presented

Data taken with mapping observations were made into imagesusing the Contour and Trigrid functions in IDL and then theimages were compared with those processed using CLASS Thetwo images which were drawn in terms of antenna temperatureshowed a good agreement with each other The contour maps inFigure 5 were made by taking signals stronger than 5 whichcorresponds to 04 K

42 Line Identification

Line identifications were made exclusively using the catalogof Lovas (1992) The catalog lists molecular transitions detectedfrom Orion-KL Sgr B2 and from various interstellar molecularclouds The number of detected lines in the catalog is sufficientfor the use in the line identifications of our 3 mm observationsThe identified line frequencies were determined by Gaussian fit-ting after correcting by the corresponding Doppler shifts of46and 39 km s1 for W3(OH) and W3 IRS 5 respectively andthen compared with the rest frequencies from the Lovas catalog

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 163No 1 2006

Fig 1mdashSpectra of W3(OH) from 848 to 1154 GHz obtained during 2001 NovemberndashDecember with the TRAO 14 m telescope Each panel combines twoneighboring frequency measurements to have about 1 GHz bandwidth

164

Fig 1mdashContinued

165

Fig 1mdashContinued

166

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 6: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashSpectra of W3(OH) from 848 to 1154 GHz obtained during 2001 NovemberndashDecember with the TRAO 14 m telescope Each panel combines twoneighboring frequency measurements to have about 1 GHz bandwidth

164

Fig 1mdashContinued

165

Fig 1mdashContinued

166

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 7: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashContinued

165

Fig 1mdashContinued

166

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 8: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashContinued

166

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 9: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashContinued

167

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 10: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashContinued

168

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 11: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashContinued

169

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 12: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashContinued

170

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 13: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashContinued

171

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 14: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 1mdashContinued

Fig 2mdashSpectra of individual lines in W3(OH) observed with 250 kHz resolution

172

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 15: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 2mdashContinued

173

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 16: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 2mdashContinued

174

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 17: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 2mdashContinued

175

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 18: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 2mdashContinued

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 19: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

A total of 43 transitions including blended lines from 18 spe-cies were detected in the 1 MHz observations toward W3(OH)whereas 75 lines from 28 species were detected in the 250 kHzresolution observations (Table 2) As shown in Table 2 most ofthe lines detected at 250 kHz resolution are weak (01ndash02 K)except for a few lines such as HC3N (12ndash11) and H13CO+ (1ndash0)The H13CO+ (1ndash0) transition was only observed in 250 kHz res-olution the nondetection of H13CO+ (1ndash0) in 1 MHz spectrumwas due to bad channels in the vicinity of the emission line Only20 lines were detected in W3 IRS 5 (Table 2) The detection ofH41 recombination line in W3 IRS 5 was interesting becauseit was not detected in W3(OH) but was observed in G343+015 (Kim et al 2000) HNCO [4(0 4)ndash3(0 3)] CH3OH [0(0)ndash1(1)E ] andCH3OH [2(1)ndash1(1)A] transitions were confirmedfrom high-resolution observations toward the ultracompact H ii

region G343+015 (Table 3)Though none of the SO2 transitions in the 3 mm wavelength

were detected from the 1 MHz spectral resolution surveys ofKim et al (2000) a number of SO2 transitionswere observed fromOrion-KL with strong intensities Therefore their detections wereexpected especially inG343+015 because in the context of chem-istry G343+015 is more like that in Orion-KL As a result weconfirmed the SO2 [3(1 3)ndash2(0 2)] transition in G343+015W3(OH) and W3 IRS 5 However SO2 [27(3 25)ndash26(4 22)]having an energy level difference of 370 K in temperature wasonly detected in G343+015 just above the noise level (Fig 4)SO2 [25(3 23)ndash24(4 20)] was not detected

In order to show the difference in sensitivity between the1 MHz and 250 kHz spectrometers we present relative sensitivitycurves for the two spectrometers in Figure 6 The sensitivity curves

were obtained from the observations of the SiO (v frac14 1 J frac14 2 1)line of a point source Cygni with a short exposure time

5 INDIVIDUAL MOLECULES

In this section we present derived abundances of moleculesSince our telescope beam size is large the derived abundancesare affected by the following beam-filling effects Assuming thatantenna response and source brightness distribution are bothGaussian the beam-filling factor (s) can be written as (egNummelin et al 1998)

s frac14 2(source)=frac122(telescope beam)thorn 2(source)

where (source) is the source size and (telescope beam) is thebeam size of the antenna Then the corrected brightness tem-perature is

Tbc frac14 Tmb=s

where Tmb is the main-beam brightness temperature If a sourceis smaller than the telescope beam the derived abundance will beunderestimated The source sizes are however often not knownTherefore we caution that our derived abundances are likely tobe underestimated for small sources Distribution of SiO mole-cules for instance is known to be compact whereas that of CO ismostly extended The derived column abundances of SiO listed inTable 4 therefore are underestimated On the other hand the COdistribution is expected to be larger than the telescope beam sizeIn this paper we consider the sidelobes of the beam by taking for-ward spillovers and scattering factors in the column densities

Fig 2mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 177

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 20: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 3mdashSpectra of lines in W3 IRS 5 observed with 250 kHz resolution

178

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 21: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 3mdashContinued

179

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 22: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 3mdashContinued

180

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 23: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 3mdashContinued

181

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 24: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 3mdashContinued

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 25: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Therefore the effects of beam filling would not be very signifi-cant for large sources Comparisons of our results with those ofother authors who used smaller beams should also be done withcaution

51 Ionized SpeciesmdashHCOthorn and N2Hthorn

The molecule HCO+ is important because it is formed at amore primitive level of chemistry than species like H2CO andNH3 and the role of ion-molecule chemistry can be tested di-rectly (Turner 1995) HCO+ is closely linked with CO in both thediffuse- and dense-cloud regimes hence it provides a probe ofion-molecule chemistry because the abundance of CO and itsisotopomers are particularly well understood as a result of themodels of van Dishoeck amp Blake (1986 1988)

In some sources containing low-mass young stellar objectsHCO+ appears to be influenced by stellar outflows and N2H

+ onthe other hand seems to trace preferentially the quiescent outerenvelope (Bachiller 1996a 1996b Mardones et al 1997 vanDishoeckampBlake 1998) N2H

+ is known to bemore abundant inthe quiescent medium than in shocked regions (Bachiller 1996a1996b) In contrast to HCO+ no line wings are seen in N2H

+

(Womack et al 1992) While N2H+ is known to be a good tracer

of cold clouds HCO+ has been widely used to investigate infallmotions of prestellar cores (Gregersen et al 1997) The high abun-dance ofHCO+ in such active regions can be accounted by desorp-tion process of ice mantle components H2O and CO by shockheating (Rawlings et al 2000)

In this survey we detected several ionized species such asHCO+ H13CO+ and N2H

+ in W3(OH) whereas N2H+ was not

detected in W3 IRS 5 Column densities for N2H+ have been

found to be 5 1012 cm2 and 1014 cm2 in cold and warmclouds respectively (Womack et al 1992) but it is apparentlyabsent from hot and shocked gas clouds (Womack et al 1992Bachiller 1996a 1996b) Therefore it is likely that either disper-sion by shocks generated by outflows from the embedded source(Womack et al 1992 Bachiller 1996a 1996b) or depletion of thespecies on the dust (Bergin et al 2002) is responsible for the non-detection of N2H

+ in our observations toward W3 IRS 5Self-absorption in the line of HCO+ was seen in G343+015

andW3 IRS 5 however no self-absorptionwas found inW3(OH)(Fig 7) The low column density of HCO+ in G343+015 there-fore might be influenced by absorption in the surrounding coldenvelope The spectral feature of HCO+ is found to be redshiftedby 3 km s1 from the central velocity (58 km s1) and an ab-sorption feature was seen at61 km s1 While the strong HCO+

peak in the G343+015 spectrum appears to be redshifted it is

blueshifted in W3 IRS 5 The HCO+ spectral feature of W3 IRS 5seems to be incompatible with the nominal spectral shapes of theinfallingcollapsing models (Zhou et al 1993) An explanationmay be because HCO+ in W3 IRS 5 might have been affected bythe outflows

Helmich amp van Dishoeck (1997) obtained an HCO+ columndensity of N frac14 2 1014 cm2 toward W3(H2O) [Note thatW3(OH) and W3(H2O) are separated by 600] The result is verysimilar to the value of N frac14 123 1014 cm2 found from ourobservations in the direction of W3(OH) which would includeW3(H2O) in the same beam

On the other hand the abundance ratio of HCO+ relative toN2H

+ was found to be22 inW3(OH) (Table 4) The excitationtemperature (5 K) adopted in our calculation is a reasonablevalue for heavy molecules and therefore the derived N2H

+

column density would not differ from the realistic density In ad-dition the line intensity (06 K) of N2H

+ observed inW3(OH) in-dicates that it is not optically thick A high ratio of HCO+N2H

+

is predicted by the reaction N2H++COHCO++N2 which in-

volves destruction of N2H+ and the production of HCO+ (Snyder

amp Watson 1977)

52 Sulfur-bearing SpeciesmdashOCS SO CS and SO2

The abundances of sulfur-bearing species in the star-formingregions have been found to be significantly different from thosein dark or quiescent regions (Charnley 1997) Sulfur-bearingspecies are known to form efficiently from reactions involvingparent H2S The widely accepted theory of H2S formation is thatsulfur freezes out onto grains during the collapsing phase and itremains on the grains in the form of H2S until it is evaporated byheating in a hot core (Charnley 1997 Hatchell et al 1998) CS isoften used as a tracer of dense gas region and it results from re-actions of S+ and S with CH and C2 (van Dishoeck amp Blake1998)

We observed several optically thin isotopic variants of CS inW3(OH) and G343+015 and we derived column densities andoptical depths using these optically thin lines (Table 4) CS col-umn density ratios between W3(OH) and G343+015 derivedfrom isotopic variants are similar to within a factor of 2 On theother hand CS column densities of W3(OH) derived from itsisotopic variants appear to be similar

We have carried out large velocity gradient (LVG) modelcalculations to construct models for comparison with observa-tions Comparisons were made for excitation temperatures op-tical depths and the observed intensities of CS (2ndash1) forW3(OH) and G343+015 In the models we examined H2

Fig 3mdashContinued

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 183

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 26: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 4mdashSpectra of lines in the ultracompact H ii region G343+015 observed with 250 kHz resolution

184

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 27: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 4mdashContinued

185

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 28: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 5mdashMap of CS (2ndash1) map (upper left panel ) HCO+ (1ndash0) (upper right panel ) HCN (1ndash0) (middle left panel ) and HNC (1ndash0) (middle right panel ) taken towardthe strong OH-emitting region W3(OH) and CS (2ndash1) map taken towardW3 IRS 5 (lower left panel ) Observations were carried out between 2001 November and 2002February Contours are drawn for the signals stronger than 04 K with 5 While the CS (2ndash1) map (lower right) of G343+015 drawn with the data taken from theindependent project is an integrated intensity map other map figures are contoured using the peak of signal taken from this project

186

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 29: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

TABLE 2

Observed Transitions in W3(OH) and W3 IRS 5 from a 1 MHz and 250 kHz Resolution Spectrometer

TA (K) (km s1)

RTA dv (K km s1)

(rest)(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

851391 OCS 7ndash6 015 702 111

853389 C3H2 2(1 2)ndash1(0 1) 016 021 688 829 117 191

015 939 150 1 MHz

854507 CH3CCH 5(2)ndash4(2) 017 281 050

854557 CH3CCH 5(1)ndash4(1) 014 252 038

854573 CH3CCH 5(0)ndash4(0) 010 779 022

008 1165 101 1 MHz (b)

860935 SO 2(2)ndash1(1) 024 019 354 620 089 122

863388 H13CN 1ndash0 F = 1ndash1 015 012 863 605 135 077

026 1546 355 1 MHz

863402 1ndash0 F = 2ndash1

863423 1ndash0 F = 0ndash1

867543 H13CO+ 1ndash0 036 010 451 860 174 091

868470 SiO 2ndash1 v = 0 009 512 050

870907 HN13C 1ndash0 F = 0ndash1 009 281 027

870909 HN13C 1ndash0 F = 1ndash1

873169 C2H 1ndash0 32ndash12 F = 2ndash1 052 019 429 835 234 166

040 413 163 1 MHz

873286 C2H 1ndash0 32ndash12 F = 1ndash0 028 015 365 521 108 082

033 312 109 1 MHz

874020 C2H 1ndash0 12ndash12 F = 1ndash1 026 026 317 310 086 085

025 628 171 1 MHz

874072 C2H 1ndash0 12ndash12 F = 0ndash1 019 420 084

879252 HNCO 4(0 4)ndash3(0 3) 009 204 017 019 769

886304 HCN 1ndash0 F = 1ndash1 112 033 324 301 384 105

886318 HCN 1ndash0 F = 2ndash1 182 064 396 372 765 252

886339 HCN 1ndash0 F = 0ndash1 066 055 437 402 305 235

160 1137 1882 1 MHz (b)

891885 HCO+ 1ndash0 258 061 535 938 1462 608

193 725 1485 1 MHz

906635 HCN 1ndash0 F = 0ndash1 120 024 395 785 500 200

906636 HCN 1ndash0 F = 2ndash1

906637 HCN 1ndash0 F = 1ndash1

100 455 609 1 MHz

909790 HC3N 10ndash9 055 017 365 513 214 090

035 487 186 1 MHz

919801 CH3CN 5(2)ndash4(2) 006 373 024

919853 CH3CN 5(1)ndash4(1) 009 269 026

919871 CH3CN 5(0)ndash4(0) 016 270 046

920344 H41 42ndash41 015 2116 343

924943 13CS 2ndash1 028 483 145

931716 N2H+ 1ndash0 F1 = 1ndash1 F = 0ndash1 060 703 446

931719 N2H+ 1ndash0 F1 = 1ndash1 F = 2ndash2

931721 N2H+ 1ndash0 F1 = 1ndash1 F = 1ndash0

931735 N2H+ 1ndash0 F1 = 2ndash1 F = 2ndash1

931738 N2H+ 1ndash0 F1 = 2ndash1 F = 3ndash2

931740 N2H+ 1ndash0 F1 = 2ndash1 F = 1ndash1

063 979 654 1 MHz

951694 CH3OH 8(0)ndash7(1) A+ 016 558 095

959143 CH3OH 2(1)ndash1(1) A+ 010 405 043

964130 C34S 2ndash1 032 361 120

045 599 284 1 MHz

967394 CH3OH 2(1)ndash1(1) E 043 377 173

967414 CH3OH 2(0)ndash1(0) A+ 049 486 251

967446 CH3OH 2(0)ndash1(0) E 017 401 072

034 1040 376 1 MHz (b)

967555 CH3OH 2(1)ndash1(1) E 013 396 055

013 484 068 1 MHz

971721 C33S 2ndash1 014 346 053

973012 OCS 8ndash7 008 262 023

979810 CS 2ndash1 243 087 450 747 1160 692

223 591 1399 1 MHz

992999 SO 3(2)ndash2(1) 092 048 584 670 572 339

187

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 30: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

densities up to 106 cm3 and used a kinetic temperature of 20 Kapproximately the value of T

R found from the CO (1ndash0) line ofW3(OH) Input parameters such as column densities opticaldepths and fractional abundances for the LVG calculations weretaken from Table 4 Good agreements in line intensities betweenmodels and the observational results were found at hydrogendensities of nH2

frac14 3 104 cm3 and nH2frac14 2 104 cm3 for

G343+015 andW3(OH) respectively As demonstrated in Fig-ure 8 the molecular hydrogen densities optical depths and ex-citation temperatures are determined at the points where observedtemperatures (Tex) and theoretical temperatures (T

R) meet Goodagreement in the excitation temperatures and optical depths be-tween observations and LVGmodels were also found at the cross-ing points where hydrogen densities are nH2

frac14 3 4eth THORN 104 cm3In Figure 8 the horizontal line is the observed T

R the solid line isthe variation of the excitation temperature (Tex) the dotted line isthe variation of T

R and the dashed line is the variation of theoptical depth

We detected SO [3(2)ndash2(1)] in W3(OH) W3 IRS 5 andG343+015 but the detection of SO [2(1)ndash1(0)] in W3 IRS 5wasmarginal Lower limits to SO column densities were estimatedfor the three sources From SO isotopic variants 34SO 33SO andS18ON (SO) frac14 13 1015 cm2 in W3(H2O) and N (SO)frac14 50 1015 cm2 inW3 IRS 5 were derived byHelmich amp van Dishoeck(1997) We obtained N (SO) frac14 15 1014 cm2 in W3(OH) and87 1013 cm2 in W3 IRS 5 from the SO [3(2)ndash2(1)] transi-tion (Table 4) Contrary to the observations of Helmich amp vanDishoeck which show a high abundance of SO inW3(OH) com-pared to that inW3 IRS 5 we found the opposite result Althoughthe lower limit columndensitieswere derived for the three sourcesthe abundances of SO in W3(OH) and G343+015 seem to behigh because even the lower limit columndensities approach to thecolumn density derived from the optically thick CS line On theother handN(SO) inW3 IRS5was found to be greater thanN(CS)The destruction of SOmay be contributing to the formation of

SO2 The SO2 line intensity of W3 IRS 5 was observed to be

TABLE 2mdashContinued

TA (K) (km s1)

RTA dv (K km s1)

(rest)

(GHz) Species Transition W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 W3(OH) W3 IRS 5 Notes

1000764 HC3N 11ndash10 052 320 175

033 534 188 1 MHz

1014778 H2CS 3(1 3)ndash2(1 2) 020 249 053

1025401 CH3CCH 6(2)ndash5(2) 016 420 053

1025460 CH3CCH 6(1)ndash5(1) 018 183 035

1025480 CH3CCH 6(0)ndash5(0) 021 301 066

1030404 H2CS 3(0 3)ndash2(0 2) 021 382 086

1040294 SO2 3(1 3)ndash2(0 2) 010 011 424 722 045 081

1046170 H2CS 3(1 2)ndash2(1 1) 026 436 118

1070139 CH3OH 3(1)ndash4(0) A+ 017 462 082 1 MHz

1088939 CH3OH 0(0)ndash1(1) E 019 329 065

1091736 HC3N 12ndash11 045 345 166

1097822 C18O 1ndash0 044 061 425 381 196 245

056 424 250 1 MHz

1102014 13CO 1ndash0 451 450 448 621 2139 2965

469 509 2530 1 MHz

1103650 CH3CN 6(3)ndash5(3) 010 475 052

1103761 CH3CN 6(2)ndash5(2) 011 197 024

1103814 CH3CN 6(1)ndash5(1) 010 617 062 (b)

1103835 CH3CN 6(0)ndash5(0)

1123588 C17O 1ndash0 F = 32ndash52 028 013 433 534 129 072

1123590 C17O 1ndash0 F = 72ndash52

1123600 C17O 1ndash0 F = 52ndash52

1131442 CN 1ndash0 J = 12ndash12 F = 12ndash32 028 019 205 540 062 107

034 341 125 1 MHz

1131705 1ndash0 J = 12ndash12 F = 32ndash12 023 466 111 1 MHz

1131913 1ndash0 J = 12ndash12 F = 32ndash32 035 427 159 1 MHz

1134881 1ndash0 J = 32ndash12 F = 32ndash12 037 021 471 816 184 178

031 282 132 1 MHz

1134910 1ndash0 J = 32ndash12 F = 52ndash32 055 534 311

060 502 319 1 MHz

1134996 1ndash0 J = 32ndash12 F = 12ndash12 021 449 098

010 246 050 1 MHz

1135089 1ndash0 J = 32ndash12 F = 32ndash32 025 019 291 480 076 094

025 415 112 1 MHz (b)

1152712 CO 1ndash0 1049 991 751 973 8348 10216

1085 801 9216 1 MHz

NotesmdashResults of detected molecules and their transitions with a 1 MHz and 250 kHz filter bank Some blended spectral features are marked by lsquolsquo( b)rsquorsquo Severalspectra for example HCN show a large difference in line width between 1MHz and 250 kHz This is mostly due to blending at 1MHz spectra and hence fitting errorsare expected to be significant but errors for some molecules eg C34S are probably due to uncertainties Line identifications were made after correcting Doppler shiftwith the velocity of local standard of rest of 46 km s1 for W3(OH) and 39 km s1 for W3 IRS 5

KIM ET AL188 Vol 162

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 31: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

stronger than that in W3(OH) Helmich amp van Dishoeck (1997)attribute it to a more active outflow in W3 IRS 5 than that inW3(OH) The column density of SO2 in the three sources ap-pears to be less abundant than that of SO by a factor of 4ndash10

The molecule OCS (J frac14 7 6) would be detected in W3(OH)and G343+015 but not in W3 IRS 5 On the other hand twoweak lines of OCS were observed in W3 IRS 5 from submilli-meter observations by Helmich amp van Dishoeck (1997) The

OCS (8ndash7) transition was also seen in W3(OH) but it was notdetected inG343+015where only 1MHz observations are avail-able Since G343+015 was identified as a site rich in moleculescompared with W3(OH) and W3 IRS 5 OCS (8ndash7) should havebeen detected if the observations had been carried out deep withthe high-resolution spectrometer Our measurements for OCS(7ndash6) in W3(OH) were performed twice on different dates be-cause of its weak signal-to-noise ratio The weak line intensity ofOCS implies a low abundance compared with other S-bearingspecies detected in our observations The probable reason is thatthe local temperature is not high enough for the active formationof OCS via a chemical reaction of CS with atomic oxygen Inaddition according to a reaction process postulated by Prasad ampHuntress (1982) the nondetection of H2CSmdasha precursor of OCSin the reaction schememdashis another possible reason accountingfor the nondetection of OCS in W3 IRS 5

53 HCN and HNC

Both HCN and HNC have the same precursors HCNH+ andH2CN

+ and have almost the same dipole moments from whicha similar HNCHCN abundance ratio is expected HCN has threehyperfine transitions (F frac14 0 1 F frac14 1 1 and F frac14 2 1) andtheir intensity ratio (ie statistical weight) of those componentsin the LTE and optically thin conditions is 135

The chemical model of Brown et al (1989) predicted an HNCHCN ratio of 005 whereas Herbst (1978) inferred 09 Herbstrsquosratio is consistentwith the one derived in cold clouds where chem-ical reactions are dominated by gas-phase reactions (Wootten et al

TABLE 3

High-Resolution (250 kHz) Observations of Missing Lines (Kim et al 2000) and New Observations of Interesting Molecules

and Isotopic Variants of G343+015

(obs)a

(GHz)

TA

(K)

(km s1) Species Transition

(rest)(GHz)

RTA dv

(K km s1) Notes

863397 054 1063 H13CN 1ndash0 F = 1ndash1 863388 608

1ndash0 F = 2ndash1 863402

1ndash0 F = 0ndash1 863423

867544 059 485 H13CO+ 1ndash0 867543 305

870908 034 439 HN13C 1ndash0 F = 0ndash1 870907 156

HN13C 1ndash0 F = 2ndash1 870909

HN13C 1ndash0 F = 1ndash1 870909

879256 033 445 HNCO 4(0 4)ndash3(0 3) 879252 158

886303 034 088 HCN 1ndash0 F = 1ndash1 886304 031

886322 120 280 HCN 1ndash0 F = 2ndash1 886318 356

886343 128 230 HCN 1ndash0 F = 0ndash1 886339 407

891889 254 299 HCO+ 1ndash0 891885 803 Self-absorbed

906639 291 363 HNC 1ndash0 F = 0ndash1 906635 1120 (b)

HNC 1ndash0 906635

HNC 1ndash0 F = 2ndash1 906636

HNC 1ndash0 F = 1ndash1 906637

919794 007 521 CH3CN 5(2)ndash4(2) F = 6ndash5 919801 040

919863 019 1212 CH3CN 5(1)ndash4(1) 919853 240 (b)

CH3CN 5(0)ndash4(0) 919871

971721 021 343 C33S 2ndash1 971721 076

975826 031 275 CH3OH 2(1)ndash1(1) A 975828 089

979812 458 442 CS 2ndash1 979810 2144

1040291 013 593 SO2 3(1 3)ndash2(0 2) 1040294 084

1070615 027 136 SO2 27(3 25)ndash26(4 22) 1070603 040 ()

1088939 051 546 CH3OH 0(0)ndash1(1) E 1088939 293

1099056 023 646 HNCO 5(0 5)ndash4(0 4) 1099057 160

NotesmdashResults of detected molecules and their transitions with a 250 kHz filter bank Some blended spectral features are marked by lsquolsquo(b)rsquorsquoThe SO2 27(3 25)ndash26(4 22) transition marked by lsquolsquorsquorsquo is a probable detection (see Fig 4)

a Observed frequencies are obtained after correcting Doppler shift with the velocity of the local standard of rest 58 km s1

Fig 6mdashLine profiles of a point source Cygni in the SiO (v frac14 1J frac14 2 1) line obtained with a 250 kHz and a 1 MHz spectrometer

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 189No 1 2006

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 32: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

TABLE4

ColumnDensitiesforW3(OH)W3IRS5andG343+015

ColumnDensity(1013cm

2)

OpticalDepth()

Tex

(K)

N(X

)N(H

2)a(101

0cm

2)

Molecule

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

W3(O

H)

W3IRS5

G343+015

From

OCS

gt136

gt88

164

164

lt160

55

J=7ndash6

gt28

210

gt32

J=8ndash7

C3H2

gt06

gt09

70

70

gt07

SO

gt107

gt147

gt168

193

193

193

gt126

gt123

gt08

2(2)ndash1(1)

gt146

gt01

gt141

92

92

92

gt172

gt01

gt88

3(2)ndash2(1)

SiO

gt01

gt04

71

gt02

gt02

HNCO

gt02

gt17

gt19

70

70

70

gt03

gt14

gt12

4(04)ndash3(03)

22

106

gt14

5(05)ndash4(04)

HCN

[2ndash1]

76

03

36

42

33

38

57

38

47

89

25

23

HCN

[1ndash1]

39

16

01

38

27

03

46

33

50

46

14

01

HCN

[0ndash1]

32

29

31

33

32

39

38

37

48

38

25

20

H13CN

03

02

21

01

02

06

57

38

47

04

03

13

04

05

19

01

04

05

46

33

50

05

04

12

05

03

20

03

02

05

38

37

48

06

03

13

HCO+

123

59

67

44

32

44

68

38

68

144

50

43

H13CO+

03

03

07

02

02

03

68

38

04

02

04

HNC

55

04

130

34

41

47

44

74

65

82

HN13C

01

05

01

01

47

01

03

N2H+

gt06

gt21

45

45

gt07

gt13

CS

200

gt37

451

36

41

67

71

99

236

gt31

285

292

628

55

60

67

344

397

259b

563

48

54

67

305

356

13CS

07

12

01

01

67

09

08

13CS

C33S

02

04

01

01

67

03

02

C33S

C34S

15

27

02

02

67

18

17

C34S

SO2

gt12

gt19

gt22

52

52

gt14

gt16

gt14

3(13)ndash2(02)

gt224

246

27(325)ndash26(422)

HC3N

gt01

240

gt01

CO

69309

81502

50

49

189

180

81606

69400

27(325)ndash26(422)

75564

88950

55

55

189

88972

75786

13CO

68462

80493

49

48

189

80610

68581

C17O

13CO

4834

6612

24696

06

06

48

114

5692

5634

15599

C18O

36309

69

22936

C17O

227

120

490

00

00

01

267

103

310

C18O

451

430

1622

01

01

04

531

366

1025

From

rotationdiagram

HC3N

04

01

15

01

168

1

189

1

05

09

CH

3OH

155

05

200c

97

1

99

183

125

H2CS

112

15

1148

80

125

1574

132

718

CH

3CN

02

10

11

01

362

7

350

3

03

07

CH

3CCH

51

12

112

01

255

05

270

02

51

132

aThedensity

NH2frac14

85

1022cm

2forW3(O

H)NH

2frac14

12

1023cm

2forW3IRS5andNH2frac14

16

1023cm

2forG342+015derived

from

NH

2frac14

27

1021cm

2R T

R(1

3CO)dv(Sanderset

al1984)

bDerived

from

1MHzdataofC34S

cFrom

Kim

etal(2000)

190

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 33: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Fig 7mdashSpectral lines of HCO+ H13CO+ HCN and HNC observed toward G343+015 IRAS 183350711 IRAS 184500200 W3(OH) and W3 IRS 5

191

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 34: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

1978) A ratio higher than 10 is also observed in TMC-1 (IrvineampSchloerb 1984) The chemical environment seems to be differentin warm and giant molecular clouds in which the HNCHCN ra-tios are observed to be between 0015 and 040 (Goldsmith et al1981) Since the small ratio in the warm clouds could not be ac-counted for solely by the gas-phase reaction with a precursorH2CN

+ Goldsmith et al (1981) proposed a neutral-atom molec-ular reaction CH2+HHCN+H Talbi et al (1996) introducedan exchange reaction H+CNH$HCN+H to account for the lowratio with increased temperatures The abundance variation ofHCN and HNC between giant molecular clouds and cold cloudsis due the particular neutral-neutral reaction which is importantfor HCN formation (Millar amp Freeman 1984) This neutral-neutral reaction has a small activation barrierWhile the reaction isswitched off in cold clouds it proceeds efficiently in the warmergiant molecular clouds leading to the enhancement of HCN

In the LTE and optically thin conditions the three hyperfinestructure (hfs) ratio (F frac14 0 1 F frac14 1 1 and F frac14 2 1) in HCNis 135 and therefore deviations from the intensity ratio al-lows one to investigate optical depth effects Emission of H13CNis assumed to be optically thin and the hyperfine ratio is expectedto retain the LTE ratio Zinchenko et al (1993) observed a peculiarHCN hfs ratio toward an H2Omaser source S76E where the F frac141 1 intensity is stronger than those of F frac14 2 1 and F frac14 0 1They attributed it to a shielding by blueshifted low density andcold gas in the line of sight GuilloteauampBaudry (1981) estimated frac14 1 3 from the HCN lines they studied In contrast to coldclouds where the F frac14 1 1 component is observed to be strongerthan F frac14 2 1 or F frac14 0 1 in warmer clouds the F frac14 1 1 com-ponent is observed to be weaker than the two other components

(Walmsley et al 1982) To account for the widespread HCNanomalies Cernicharo et al (1984) proposed a scattering processby a diffuse envelope for emission from optically thick hot coresThe first two components are optically thick lines whereas the last0ndash1 component is an optically thin line Radiation from the hotcore of the first two components becomes scattered by the sur-rounding envelope whereas the emission of the optical thin com-ponent (F frac14 0 1) penetrates without being scattered Thereforethe 0ndash1 component is observed to be relatively stronger than theintensity predicted from LTEAmong three sources we observed an hfs ratio close to LTE

was observed in W3(OH) and the F frac14 0 1 component inG343+015 was observed to be the strongest among the hfs com-ponents (Fig 7) Our 250 kHz spectrum of G343+015 shows anevident absorption that was not seen in W3(OH) and W3 IRS 5Hence aswill be described later in detail the absorptionscatteringby cold gas of the optically thick component (F frac14 2 1) inG343+015 may be responsible for the observed hfs ratio The hyperfineratios of H13CN as was supposed are found to be consistent withthe LTE ratios in the three sources supporting our assumption thatthe hfs anomaly seen in the HCN lines can be explained by opticaldepth effectsAs shown in Figure 7 typical hyperfine ratios are found from

our high-resolution (250 kHz) observations of G343+015W3(OH) W3 IRS 5 IRAS 184500200 and IRAS 183350711 (S-J Kim et al 2006 in preparation) and from Moprasurveys of IRAS 174702853 (Kim et al 2002)

1 F frac14 0 1 lt F frac14 1 1 lt F frac14 2 1 which is the LTE ratioand is seen in W3(OH) and IRAS 1747028532 F frac14 1 1 lt F frac14 0 1 lt F frac14 2 1 which is seen in W3

IRS 53 F frac14 1 1 lt F frac14 2 1 lt F frac14 0 1 which is seen in

G343+015 and in IRAS 1845002004 F frac14 2 1 lt F frac14 0 1 lt F frac14 1 1 which is seen IRAS

183350711

We also observed a strong absorption feature in the HCO+

spectrum in G343+015 From this result we speculate theremight be a strong relationship between the HCN hfs ratio and thedegree of HCO+ absorption an optical depth effect proposed byCernicharo et al (1984) Thermal overlap effects introduced byGottlieb et al (1975) and developed by Guilloteau amp Baudry(1981) have been used to account for the anomaly The modelpredicts the thermalization occurs forN gt1016 cm2 and the LTEratio deviates at N gt 3 1012 cm2 However the model fails topredict R02gt1 seen from our observations in IRAS 183350711(Fig 7) and S76E (Zinchenko et al 1993) up toN ffi 1016 cm2 akinetic temperature of 30 K and nH2

frac14 105 cm2Though HN13C was observed in G343+015 and W3(OH)

it was not observed in W3 IRS 5 The nondetection indicatesthat the HN13CHNC ratio in W3 IRS 5 is lower than that in theother two sources Since three hyperfine components of HCNwere resolved by the 250 kHz observations we derived columndensities for three hyperfine components separately As a resulta high abundance in three HCN components was obtained inW3(OH) The derived T

A(HNC)TA frac12HCN(F frac14 2 1) ratios are

07 and 04 for W3(OH) and W3 IRS 5 respectively We ob-tained only a lower limit column density of HNC in W3 IRS 5sinceHN13Cwas not detected The derivedHNC column density(N frac14 568 1013 cm2) in W3(OH) is close to that from sub-millimeter observations by Helmich amp van Dishoeck (1997)This similarity indicates that HNC is an extended componentThe highest abundance and optical depth of HNC were found inG343+015 (Table 4)

Fig 8mdashLVG model calculations of CS (2ndash1) for G343+015 and W3(OH)with a kinematic temperature of 20 K which was obtained from CO (1ndash0) Thehorizontal lines represent observed T

R solid lines represent excitation tem-perature dotted lines represent the model variation of T

R and dashed linesrepresent the variation of optical depth

KIM ET AL192 Vol 162

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 35: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Intensities of CS and HCN (F frac14 2 1) in molecular clouds arequite similar but Dickel et al (1980) observed a strong intensityof HCN in W3 IRS 5 where weak CS lines were observed Ourresults for T

A(CS)TAfrac12HCN(F frac14 2 1) are 14 and 13ndash14 for

W3 IRS 5 and W3(OH) respectively which imply a similardensity distribution for the species in the regions as proposed byDickel et al (1980) The ratios are very close to each other eventhough two sources seem to be in a significantly different evo-lutionary stages The similarity in the ratio might be widespreadin the interstellar molecular clouds although it might be pre-mature to conclude this

54 Carbon-bearing SpeciesmdashCCH CN C3H2 and HC3N

Unsaturated carbon-chain species such asCCHCN andHC3Nare common components of cold molecular clouds and they havebeen widely observed in dark clouds Among them HC3N is auseful tracer of density of both cold clouds and hot cores

We observed three transitions of HC3N in W3(OH) whereaswe have just one clear detection of the 10ndash9 transition inW3 IRS5where it is weak compared to that in W3(OH) With the non-detections of the other two transitions in W3 IRS 5 it implies arelatively low abundance of HC3N in this region and suggests thatthe HC3N abundance is enhanced in the later evolutionary phase(eg the hot core H ii phase) rather than in the earlier phase Fromthe rotation diagram analysis for HC3N of W3(OH) a rotationtemperature of 17 K and a column density of 5 1012 cm2 wasderived (Table 4) In comparison with those of G343+015 tem-perature and density are similar when 3mm transitions are used inthe rotation analysis but the analysis using 2 mm transitions re-sults in a somewhat large difference in the column density Thederived column densities and temperatures from the three sourcesare summarized in Table 4 in which the column density of C3H2

is a lower limit

55 Symmetric and Asymmetric SpeciesmdashCH3CN CH3OHCH3CCH H2CS and HNCO

The molecules CH3CN and CH3CCH have been widely usedto obtain physical and chemical information in hot molecularcores as they have several transitions with wide excitation en-ergies in close frequency ranges Therefore simultaneous obser-vations of several K-ladder transitions of a molecule can reduceobservational errors arising from uncertainties in telescope point-ing beam efficiency and calibrations

The molecule CH3OH is a slightly asymmetric top moleculeand is one of the most common species in the envelopes and thehot cores of star-forming clouds It is a useful molecule for in-vestigating the evolutionary stage of molecular clouds becauseit has many transitions with a wide range of energies fromlt10 Kto gt100 K in the millimeter wavelength range Transitions withenergylt20K above ground state are usually detected in the 3mmwavelength range allowing detailed investigations of cold enve-lope regions rather than warm hot core regions A few transitionswith large energy differences are also detected

For a given T (rotational excitation temperature) line intensi-ties can be calculated using the following relation (Loren et al1981 Townes amp Schawlow 1975)

frac14 g(K )8Nf ij

2 20

3ckT eth1THORN

where g(K ) is the K level statistical weight f is the fractionalpopulation of each rotational level and jijj2 frac14 2 J 2 K 2eth THORN

J (2J thorn 1)frac12 The statistical weight g(K ) for K frac14 3 6 9 istwice those of other K lines (Townes amp Schawlow 1975)

The molecule CH3CCH has a lower dipole moment thanCH3CN so the relative population of CH3CCH should be muchcloser to being thermalized as its critical density is n 104 cm3The relatively high dipole moment of CH3CN requires densitiesmuch higher than 104 cm3 to approach thermalization (Askneet al 1984 Andersson et al 1984) CH3CCH is abundant in halossurrounding hot cores but CH3CN is not seen in halos due to thelow excitation temperature there (Andersson et al 1984) OurmeasuredCH3CN column density therefore is presumed to be af-fected by beam dilution resulting in a low measured density

We detected CH3CN and CH3CCH in W3(OH) and G343+015 but neither was detected in W3 IRS 5 in our survey As-suming an early evolutionary phase for W3 IRS 5 according tocurrent chemical models the nondetection of the two species in-dicates that their precursors CHthorn

3 CH4 and H2O may be highlydepleted on the grain surface On the other hand according toequation (1) intensities ofK frac14 3n (where n is an integer) lines ofCH3CN based on its statistical weight are expected to be strongerthan K frac14 2 components Contrary to the theoretical predictionhowever the 6(3)ndash5(3) transitionwas found to be equal or weakerthan theK frac14 2 components however all the transitions observedbeing weak prevents a definite conclusion Similar differencesbetween the observed line intensities and the statisticalweights arealso found in the prestellar object IRAS 174702853 (Kim et al2002)

A column densityN frac14 (21 10) 1012 cm2 and a rotationtemperature (36 7 K) of CH3CN were obtained in W3(OH)using rotation diagram analysis (see Appendix) While a full rota-tion diagram analysis for the 3 and 2 mm transitions of CH3CNobserved in the G343+015 was performed (Kim et al 2001) wealso made a rotation diagram with the 3 mm data only to compareits column density to the result from W3(OH) (Fig 9) This en-ables a direct comparison of the physical and chemical charac-teristics of the two regions In the G343+015 a column densityN frac14 (11 01) 1013 cm2 and a temperature 350 3Kwereobtained (Table 4) While we find a good agreement between therotational temperatures the column densities differ by1 order ofmagnitude implying that the physical and chemical states of thewarm or hot cores are not identical though they seem to be ina similar evolutionary stage From CH3CCH we obtained N frac14(51 12) 1013 cm2 and Trot frac14 255 05 K in W3(OH)and N frac14 (11 01) 1014 cm2 and Trot frac14 270 02 K inG343+015 (Table 4) The results of W3(OH) clearly supportour assumption a low temperature and a high column densityof CH3CCH relative to CH3CN In comparison with OSO-20 mobservations by Askne et al (1984) who obtained a column den-sity N frac14 (28 04) 1014 cm2 and a rotational temperature40ndash60 K for CH3CCH toward G343+015 and W3(OH) ourobservations provide lower column density values possibly aresult of beam dilution because of the larger beam

The scatter of the data points in the CH3CN and CH3CCH ro-tation diagrams is not very significant in the two sources (Fig 9)except for one data point of CH3CN and two points of CH3CCHof W3(OH) These all have higher excitation energies and areweak lines Three points are supposed to correspond to very hotcore emission as they deviate from themodel fit which is close toa warm core model There might be however some possibilityof line misidentification These lines are weak and moreoverthey were not designated as detections in G343+015 and onlya 1 MHz spectrum is available (Kim et al 2000)

Seven CH3OH transitions were observed in W3(OH) butnone of themwere observed inW3 IRS 5 The 0(0)ndash1(1) E and

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 193No 1 2006

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 36: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

2(1)ndash1(1) A transitions of CH3OH which were not reported inthe previous study of G343+015 (Kim et al 2000) were de-tected in W3(OH) Rotation diagram analysis for the lines of Aand E types yielded a rotational temperature (97 1 K) and acolumn density [(16 05) 1014 cm2] from W3(OH) data(Table 4) The rotational temperature and the column densityappear to be in good agreement with those obtained from thelower energy transitions (Euk lt 45 K) of CH3OH observed inG343+015 We employed molecular parameters calculated byAnderson ampDe Lucia (1990) for the rotation diagram As shownin Figure 9 all the transitions except for a transition with a largeenergy level difference are aligned in the linear fitting Most ofthe detected transitions have upper energy states lt20 K andtherefore they tend to be excited easily at low temperatures in thecold outer halos of hot molecular cores However the 8(0)ndash7(1)A+ transition with a high energy (84 K) deviates from the linearfit The detections of CH3OH transitions with high excitation en-ergies as also in CH3CN tend to confirm the existence of a hotcore in W3(OH) For G343+015 a two-component fit in therotation diagramwith a energy value of 45 Kwas more success-ful than a one-component fit (Kim et al 2000 2001)

Apart from the nondetection of CH3CN in W3 IRS 5 thenondetection of any CH3OH transitions inW3 IRS 5 is a remark-able result From rotation diagram analysis for CH3OH tran-sitions in W3(OH) and G343+015 many rotational transitionswith low excitation temperatures are found to be excited in theregions far away from the hot cores Therefore the nondetection

of methanol implies that methanol or its reaction partners areprobably to be frozen on dust grains which is similar to the casesof CH3CN andCH3CCHAllamandola et al (1992) found a highabundance of solid state methanol in W3 IRS 5 4 orders of mag-nitude larger than the gas-phase abundance supporting our resultsand the evolution scenario The derived column densities with-out consideration of the interaction of dust and gas phase in thedense clouds must underestimate the total column densities (egSandford amp Allamandola 1993)Thioformaldehyde (H2CS) like H2CO has two distinct rota-

tional energy states ortho with parallel nuclear spins and parawith antiparallel They cannot be interconverted by radiative tran-sitions (Gardner et al 1985) While three transitions of H2CStwo of which are ortho were observed in W3(OH) and G343+015 there was no detection in W3 IRS 5 An ortho-to-para ratioof 081 in W3(OH) and 048 in G343+015 were found Withnuclear spin degeneracy gI frac1414 for para and gI frac14 34 for ortho(Turner 1991) we obtained a rotation temperature of 125 K anda column density of 11 1014 cm2 for W3(OH) using the ro-tation diagram analysis In G343+015 a model fit of three 3 mmtransitions gave a very high temperature of 1575K and a columndensity of 12 1014 cm2 was derived However since the fit asshown in Figure 9 relies on one para transition 3(0 3)ndash2(0 2)column densities and temperatures derived fromW3(OH) or fromG343+015 may not be very accurateThe HNCO [4(0 4)ndash3(0 3)] transition was observed in

W3(OH) W3 IRS 5 and G343+015 but the [5(0 5)ndash4(0 4)]

Fig 9mdashRotation diagrams for CH3CN CH3CCH H2CS and CH3OH for W3(OH) and G343+015 We have used 3 mm transitions exclusively to make directcomparisons of the excitation conditions of the molecules between G343+015 and W3(OH)

KIM ET AL194 Vol 162

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 37: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

line was only detected in G343+015 The [4(0 4)ndash3(0 3)]transition of W3(OH) was weak just above the detection limitThe column density derived fromHNCO [4(0 4)ndash3(0 3)] inW3IRS 5 was close to that of G343+015 and more abundant by 1order of magnitude compared to that in W3(OH) However theline width in W3 IRS 5 appears to be significantly broader thanthat of other species Such a broad line width was unexpectedconsidering the nature of thermal emission for the lines Potentialimplications are either a severe contamination by noise or tur-bulence caused by the strong outflows (Churchwell et al 1986)in W3 IRS 5 However with the assumption that HNCO line isthermalmdashusing the normal line width seen in other speciesmdashweperformed Gaussian fitting to derive the line width and intensityThe derived column density ignoring nonthermal effects is 17 1013 cm2 which is comparable with that of G343+015

Jackson et al (1984) surveyed HNCO in 18 molecular cloudsincludingW3(OH) using 404ndash303 and 505ndash404 transitions but thedetections of the two transitions towardW3(OH) were negativeSolomon (1973) analyzed that the intensity ratio of T

A(5 4)TA(4 3) is 156 in the case of TR 32hBJ k While his obser-

vations yield the ratio of 044 we have found the ratio of 07toward G343+015

6 CHEMICAL VARIATION OF W3(OH)W3 IRS 5 AND G343+015

We have so far investigated chemistry of several species ofinterest by categorizing them into four groups ionized sulfur-bearing HCN and HNC and symmetric and asymmetric mole-cules They were grouped in accordance with characteristics oftheir excitation conditions reaction processes and internal struc-ture In this section we discuss their chemistry and summarize ourfindings through considering the chemistry of G343+015 wherecomplete surveys were carried out (Kim et al 2000) and com-plementary observations were performed during this season

Our observational results show remarkable chemical and in-tensity variations in several molecules observed inW3(OH)W3IRS 5 and G343+015 (Fig 10 and Table 4) In comparisonwith W3(OH) high abundances of saturated molecules such asCH3CN and CH3CCH were derived in G343+015 Moleculesdetected in W3 IRS 5 are mostly simple molecules such as COand CN but many S-bearing species were observed The linewidth of SO2 is about 2 times larger than that of W3(OH) in-dicating that W3 IRS 5 is experiencing strong shocks occurringin the earlier evolutionary stage The most striking difference inthe list of detected molecules in the three sources is the non-detection of any methanol transitions in W3 IRS 5 We expectedto see a fewmethanol transitions because many transitions at the3mm region have low upper state energies andmany of them aredetected both in G343+015 and in W3(OH) and of course inthe other star-forming regions In the submillimeter observationsconducted by Helmich et al (1994) however many methanoltransitions are detected inW3 IRS 5 although the intensities of thetransitions compared with corresponding transitions observed inW3(H2O) and W3 IRS 4 are weak

From the results of our and Helmichrsquos observations we havecome up with the following scientific questions on the chemicalproperties ofW3 IRS 5Why is methanol so rare inW3 IRS 5 Isthis direct evidence of adsorption onto grain surfaces If so whyis the molecule still residing on the surface Werner et al (1980)argued thatW3 IRS 5 is in a stage just prior to the formation of anH ii region and estimated an age of (1 3) 105 yr Based on theirestimation of the age and the contemporary evolutionary scenarioof molecular clouds we would expect that hot cores (reservoirs

of heavy species) may exist in it However our observationalresults are not consistent with the scenario If hot cores are em-bedded in W3 IRS 5 some molecules that are known as a probeof hot cores eg CH3OH or CH3CN should be observed butthosemolecules have also not been observed Thus does the onsetof theH ii phasewhich is considered the next phase of the hot corestage such as G343+015 andW3(OH) play a critical role in theformation of complex molecules including methanol Or are themolecules CH3CN and CH3OH already destroyed or converted toother species in W3 IRS 5

The CH3CN column density inW3(OH) is found to be low by1 order of magnitude compared to G343+015 where only 3mmtransitions were used for a consistent comparison of the columndensities with W3(OH) (Table 4) Since its emissions arise fromhot or warm regions the lower column density in W3(OH)compared to that in G343+015 reflects different formation en-vironments through collisional rates hydrogen densities hotcore size etc In the submillimeter (345 GHz) wavelengthregion MacDonald et al (1996) found a high CH3CN columndensity in G343+015 1 order of magnitude higher than inW3(OH) determined by Helmich amp van Dishoeck (1997) Onthe other hand it appears that the CH3OH column densities andtemperatures are similar in W3(OH) and G343+015 (Table 4)This may support an idea that the steady-state gas-phase reactionof CH3OH which was inferred from a low excitation tempera-ture (10K) derived from our rotation diagram analyses producesuniform molecular abundances in the regions in a similar evo-lutionary stage

A significant difference in the column densities of HC3N wasalso found betweenW3(OH) andG343+015 (Table 4) Althougha similar temperature (15 K) for HC3N was found in the tworegions our calculation yields a column density in G343+015which is 2 orders of magnitude higher than that in W3(OH) im-plying an HC3N abundance like CH3CN highly dependent onexcitation processes in the regions

Fig 10mdashHistogram showing the variations of integrated intensities ofG343+015 andW3 IRS 5 relative to W3(OH) In this plot we considered boththe 3 mm survey of this work and a previous survey (Kim et al 2000)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 195No 1 2006

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 38: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

Dickel (1980) found a decrease of CS abundance inW3 IRS 5from CS (2ndash1) mapping observations and accounted for it bytime-dependent ion-molecule chemistry Our CS (2ndash1) map re-vealed that the CS peak was displaced from the W3 IRS 5 posi-tion by 6000 west which is close to the strong radio emissionobserved by Dickel et al (1980) W3 IRS 5 is not bright in radioemission but it is bright in far-infrared as observed by Werneret al (1980) suggesting that W3 IRS 5 is deeply embedded indense dust W3(OH) seems to be a good site to test an LTE con-dition derivations of isotopic ratios and chemical clocks becauseno absorption features of CO or HCO+ were seen in W3(OH)spectra and the derived HCN hyperfine ratio is very close to theLTE ratio which was not found in the other sources we observed

As the variation of chemical abundances in the sources wasdetermined an examination of isotopic ratios and their depen-dence on evolutionary phase or chemical properties should be ofinterest (Table 5) In W3(OH) isotopic ratios of CS13CS CSC33S and CSC34S are found to be 27 120 and 17 respectivelyWhile 13CS was not detected in 1 MHz observations C34S wasdetected in W3(OH) However both of them were detected in250 kHz resolution with high intensities in C34S indicating thatthe C34S line is more optically thick than that of 13CS On theother hand the C34S line taken with 250 kHz resolution appearsto be weaker than that taken with 1 MHz suggesting observa-tional errors in the 250 kHz data Therefore we used the 1 MHzdata for C34S and derived an isotopic ratio of 17 for CSC34S InG343+015 isotopic ratios of 37 for CS13CS 165 for CSC33Sand 21 for CSC34S have been derived None of isotopic lines of13CS C33S and C34S were detected in W3 IRS 5 HCNH13CNratios of 22 and 10 are found in W3(OH) and G343+015respectively

It was found that the HCO+H13CO+ ratios were 36 21 and 6in W3(OH) W3 IRS 5 and G343+015 respectively The lowervalue in W3 IRS 5 compared with that in W3(OH) is probablydue toweak absorption feature in theHCO+ lines caused by inter-stellar matter in the line of sight This absorption in G343+015is also shown in CO (Kim et al 2000) While the absorption inthe HCO+ line in W3 IRS 5 is not strong enough to give a sig-nificant effect on its column density its absorption strength inG343+015 is significant and results in a lower value for HCO+H13CO+ ratio

Derived isotopic ratios of 12CO13CO forW3(OH)W3 IRS 5and G343+015 are listed in Table 5 The following isotopic ra-tios have also been derived COC17O frac14 333 COC18O frac14 152and C18OC17Ofrac14 22 inW3(OH) COC17Ofrac14 739 COC18O frac14187 and C18OC17O frac14 10 in W3 IRS 5 The strong and weakabsorptions in the CO emission lines observed from G343+015and W3 IRS 5 respectively make a direct comparison difficult

Wilson amp Rood (1994) derived [12C][13C] 20 [16O][18O] = 250 and [32S][34S] 22 in the direction of the Galacticcenter According to their study the isotopic ratios of C and S are

found to depend on the positions sampled Our results seem to bemuch closer to those in the Galactic center compared to those inthe local ISM or in the solar system

7 MAPS OF CS HCOthorn HCN AND HNC EMISSIONS

We mapped W3(OH) and W3 IRS 5 in the transitions of CS(2ndash1) and HCO+ (1ndash0) to determine their distributions and sizesfor chemical modeling We also carried out similar observationsfor HCN and HNC to obtain their spatial distribution towardW3(OH) as the chemical reactions of these species are known tobe closely related to each otherThe CS map (Figs 5 and 11) of W3(OH) shows several small

clumps to the northwest northeast andsouth of W3(OH) In theCS map ofW3 IRS 5 the peak (24 K) is seen at6000 west fromthe central position of our observations and the peak positioncorresponds to W3 IRS 4 Dickel et al (1980) saw that the CSpeak falls to 05 K in the direction of W3 IRS 5 The peak posi-tion for the CS emission agrees well between this and our obser-vations but the peak intensity (09 K) of CS we measured isstronger than that of Dickel et al (1980) The CS intensity of W3IRS 5 was observed to be weaker thanW3(OH) by a factor of 25The HCO+ intensity distribution of W3(OH) shows a more

diffuse emission than in CS The HCNmap of W3(OH) shows asouthwest-northeast elongated distribution whereas the HNC

TABLE 5

Isotopic Ratios of W3(OH) W3 IRS 5 and G343+015

ISM HC0+H13CO+ CS13CS CSC33S CSC34S CO13CO COC17O COC18O HCNH13CNa HCNH13NC

W3(OH) 35 270 120 17 14 333 152 22 62

W3 IRS 5 21 12 739 187 10

G343+015 6b 37 165 21 c 27

a From the HCN (1ndash0 F frac14 2 1) transition Severe absorption in HCN is seen in G343+015b Absorption in HCO+ results in a low value in G343+015c Absorption in CO is seen and 13COC17O = 50 and 13COC18O = 22 are derived in G343+015

Fig 11mdashOverlaid images of HCO+ and CS in W3(OH) The solid anddashed lines are for CS and HCO+ respectively This plot shows a slight de-viation of the peak positions of the species and HCO+ is more widely distrib-uted than CS

KIM ET AL196 Vol 162

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 39: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

distribution appears to be more compact in addition with one ortwo embedded clumps in the northeast and southwest directions(Figs 5 and 12) The HCO+ distribution is the most extended ofthe three molecules

8 COMPARISONS WITH CHEMICAL MODELS

Chemical model calculations have been performed by severalauthors to explain different chemical characteristics of Orion-KLand TMC-1 (Caselli et al 1993Millar amp Freeman 1984) Chem-ical models have been improved as various laboratory measure-ments of molecular reactions have provided proper reaction ratesbetween important reaction chains (Caselli et al 1993) Recentlythe ultracompact hot core G343+015 has been comprehen-sively studied with model comparisons by Millar et al (1997)Thus far chemical model calculations have been concentrated oneither cold or innermost hot cores We have carried out the calcu-lations for a warm core with an intermediate temperature35 Kand for a halo with low temperatures 10 K because the twocomponents are found from our observations and analyses

Our observations at the 3 mm wavelength seem to provide abetter opportunity in the study of chemistry of the warm and coldhalo regions Most emissions at this wavelength have lower

energy than those of submillimeter wavelength providing betterconditions for deriving physical values such as temperature anddensity which are input parameters in the model calculationsThe telescope beam size at this wavelength can cover warm andhalo regions at the same time Although model calculations fora hot core with typical temperatures of gt100 K and core sizelt01 pc may have some significant inaccuracies we carried outhot core model calculations to show that low-temperature mod-els are more consistent with the 3 mm observations than 100 Kmodels lt01 pc

In the W3(OH) model calculations we used distinct shelllayers with certain temperatures and densities because two dis-tinct layers were revealed from the analyses of the observed dataWe used a model code developed for G343+015 byMillar et al(1997) In themodel we used temperatures and densities derivedfrom our observations and rotation diagram analyses Our linesurvey shows that W3(OH) has at least two distinct componentshaving different temperatures and densities They are a cold halowith a temperature of 10 K a volume density of 2 104 cm3and a radius of 15 pc and a warmhot core with temperatures inthe range of 35ndash100K densities 2 105 5 105 or 3 106 cm3and radii ranging from 01 to 10 pc The input parameters for themodel calculations are summarized in Table 6

The visual extinction values adopted in ourmodel calculationsare somewhat different from those used for hot cores such asG343+015 by Millar et al (1997) We used an H2 and Av re-lation [Av frac14 N (13CO)(015 1016) cm2] that is more valid inthe cold outer region than the inner hot region It is known thatthere is 16 mag of extinction per a column of hydrogen of3 1021 cm2 The uncertainty in the visual extinction does notaffect the chemical model because 72 mag of visual extinc-tion as used in Figure 13 is equivalent to about 300 mag in theUV and therefore there are no UV photons to affect thechemistry

Figure 13 shows results of the calculations considering grain-surface reaction routes where the calculations were performedwith a temperature of 100 K and a density of 3 106 cm3 Theinput values for the 100 K model are assumed as they are foundto be typical in hot cores The chemical model code was set as adefault to perform calculations using desorption of moleculesfrom grains in the innermost region (lt01 pc) where evapo-rations of mantle components and then gas-phase reactions areknown to be significant Initial values used in the model calcu-lations such as hydrogen density cloud temperature extinctionand cloud radius are given in the plot A sharp increase of abun-dances occurs at an early time 3 104 yr in this model Theabundances then drop rapidly after 105 yr The rapid abundance

Fig 12mdashOverlaid images of HCN and HNC in W3(OH) The solid anddashed lines are for HCN and HNC respectively It shows an identical peak forHCN and HNC and a more dense distribution of HNC than HCN

TABLE 6

Physical Parameters Adopted for Model Calculation of W3(OH )

Component

Mass

(M)n(H2)

(cm3)

T

(K )

Outer Radius

(pc)

N(H2)

(cm2)

AV

(mag) Comments

Warmhot core 50a 107andash3 106b 100cndash35d 01a 4 1023b 73e W3(OH)W3(H2O)

Extended halo 2 103andash6 103f 105andash104g 35ndash6h 1ndash15i 85 1022j 32e

a Wilson et al (1991)b Helmich amp van Dishoeck (1994)c Helmich amp van Dishoeck (1994) Mauersberger et al (1988) Wilson et al (1991)d From CH3CN rotation diagrame From AV frac14 N (13CO)015 1016 [Strohacker 1978 N(13CO) from Helmich amp van Dishoeck (1994)] and from this workf From the relation Mvir frac14 125FWHM2

HCOthornD (Strong-Jones et al 1991)g From this workh From the rotation diagrams of CH3OH CH3CCH and CH3CN from the excitation temperature of HCO+ and CS and from this worki From CS (2ndash1) and HCO+ (1ndash0) maps and from this workj From the relation NH2

frac14 27 1021RTR (

13CO) dv (Sanders et al 1984) and from this work

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 197No 1 2006

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 40: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

decrease in hot cores is related with the change of fractionalion density which is inversely proportional to the square root ofgas density (Caselli et al 1993) The plot of N2H

+ and HCO+

shows that N2H+ is more abundant than HCO+ and sharply in-

creased at 104 yrThe H2S molecule is known to be critical in the formation of

S-bearing species in a region having a temperature gt100 K H2Sis evaporated from mantle and it seems to survive about 103ndash104 yr Molecules O2 and OH are more abundant than H2S al-though the plots for these reactants are not presented here Ac-cording to Millar et al (1997) the SH radical that is liberatedthrough reaction with atomic hydrogen is destroyed by an Hatom creating an S atom The S atom reacts with O2 and OH toform SO followed by the formation of SO2 from SO with thereaction with OH In a longer timescale CS is the most abundantspecies after SO2 The plot shows that SO2 is most abundant be-tween 103 and 104 yr but after 104 yr CS is most abundant fol-lowed by OCS SO and SO2

The HCN and HNC plot demonstrates that HNC is moreabundant than HNC for lt103 yr due to the recombination ofH2NC

+ which is formed by the reaction of C+ and NHthorn3 Their

abundances become almost the same up to 108 yr as neutral-neutral reactions become efficient for producing HCN (Millaret al 1997)

The plot of C-containing species shows that C2H ismost abun-dant for lt103 yr followed by CN C3H2 and HC3N The abun-

dances of the species drop at the core age of 104 After 105 yr CNis most abundant followed by C2H C3H and H3CN The HC3Nmolecules are effectively produced in the hot gas through thereaction of C2H2 and CN Its abundance has a peak at 5 103 yrand its reactant C2H2 drops sharply at the age of 10

4 yr althoughthe C2H2 model is not included in this plotThe plot of CH3OH CH3CN CH3CCH and H2CS is most

interesting as these molecules are known to form effectivelythrough reactions with mantle components In the hot gas theformation of CH3CN is mainly driven by radiative recombina-tion of CHthorn

3 with HCN which is also the major formation in thehalo (Millar et al 1997) H2CS becomes the most abundant mol-ecule after 104 yr and the abundances of CH3CN and CH3OHare roughly equalFigure 14 presents plots of 60 Kmodels for several molecules

adopting a hydrogen volume density of 5 105 cm3 The 60 Kmodel calculations have been performed including the assumedinput parameters but with a lower density of hydrogenmoleculeand with an increased source radius compared with those of the100 K model The models exclude grain-surface reactions to bedifferent from the 100 K models because the return of mantlemolecules should be less important in cooler gasThe HCO+ and N2H

+ plot shows that HCO+ is abundant bymore than 1 order of magnitude compared with N2H

+ (Table 4)The plot shows that the difference of the abundances is significantin the earlier stage in contrast to the 100 K model

Fig 13mdashTime evolution of fractional abundances of molecules resulting from 100Kmodel calculations which include grain-surface reactions and effects of mantleevaporation

KIM ET AL198 Vol 162

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 41: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

The plot of S-bearing species shows that CS is still the mostabundant forlt105 yr After 105 yr CS SO SO2 is roughly equalwhereas the abundance of OCS is very low The calculation ofcolumn density (Table 4) indicates that SO2 is the least abundantmolecule followed by OCS the abundance of which was de-rived from the 8ndash7 transition The OCS abundance derived fromthe J frac14 7 6 transition in the table seems to have a somewhatlarge error bar as the line is weak and contaminated by noise

The HCN and HNC plot shows that their abundances arealmost the same throughout the evolutionary time but HNC be-comes more abundant after 105 yr Table 4 shows that columndensities for HCN and HNC are very close each other and con-sistent with the model

The plot of C-containing species shows that their abundancesare almost the same for lt105 yr and then CN is the most abun-dant for gt105 yr as also seen in the 100 K model We were notable to derive the column densities of CN and C2H due to the com-plexity of their hyperfine structures We could obtain a more ac-curate excitation temperature of HC3N from rotation diagramanalyses than an approximate value used for C3H2 (see Appen-dix) Therefore the column density of HC3N seems to bemore re-liable than C3H2

The plot of CH3OH H2CS and CH3CN shows that there areminor differences in the fractional abundances before 104 yr butafter 104 yr H2CS is most abundant followed by CH3OH andCH3CN The model shows an increase of fractional abundances

smoothly up to 104 yr compared to the 100 K model Among thethree species CH3CN is least abundant (Table 4) which is con-sistent with the model The largest abundance of H2CS is found inG343+015 whereas CH3OH and H2CS abundances are almostthe same in W3(OH)

As the maximum temperature of 35 K is derived from theCH3CN rotation diagram analyses a boundary of two compo-nents with a warm core and cold envelope was assumed to existaround 35 K with a hydrogen volume density of 2 105 cm3

(Fig 15) In fact the temperature range of a warm core is knownto be 30ndash100 K (Garay amp Rodriguez 1990 Andersson amp Garay1986) Therefore we expect that the 35 and 60 K models wouldproduce similar results and may agree with molecular abun-dances of a warm core for example CH3CN in our observa-tions The hydrogen volume density adopted in the 35 K modelcalculation is very close to that found from the LVG calculationof CS This 35 K model generates similar results to those fromthe 60 K model as expected since there are only minor differ-ences in temperature and density as well as a minor difference inAv The decrease of CS and increases of SO2 and SO are seen inthis model as well as seen in the 60 K model

Figure 16 presents results of the model calculations taking atemperature of 10 K for the envelope component which wasderived from the rotational analyses of methanol lines TheHCO+ and N2H

+ plot shows a sharp decrease after 106 yr butthe plot is similar to those in the 35 and 60 Kmodels The plot of

Fig 14mdashTime evolution of fractional abundances of molecules resulting from 60 K model calculations in which pure gas-phase reactions were included but theeffects of mantle evaporation were not included

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 199No 1 2006

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 42: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

S-bearing species shows that the abundances of SO and SO2 dropsharply after 106 yr and SO2 seems to be almost destroyed Thedecrease of SO is very different from the results of the 35 and 60Kmodels where SO increases as CS decreases HCN and HNCremain almost the same throughout the time span The plot ofC-containing species shows that the abundance of these species areroughly equal before 104 yr since then CN is the most abundantmolecule whereas HC3N becomes the least abundant molecule

By examining fractional abundances derived from the theo-retical modelings and the observations we may be able to esti-mate the age ofW3(OH) The CO is best fitted in the 10 Kmodelwith a cloud age between 104 and 105 yr HCO+ shows steadyvariation in density regardless of models except at 100 K whichis far from consideration for this molecule The amount of ion-ization is determined by cosmic rays and therefore affectsoverall chemical evolution Therefore ionized species includingHCO+ is significantly influenced by the increase of cosmic raysWe took the same cosmic-ray value for the model calculationsresulting in steady variations of HCO+ andN2H

+ Since 10Kwasobtained from the rotation diagram of CH3OH we expected thata 10 K model of the molecule would produce a good agreementwith observations However the resultant 10 K model value isfound to be 4 times higher than observations Among conceiv-able possibilities accounting for this difference the hot coreemission seems to be one of major reasons The best fits ofCH3CN have been found at the 35 K model and the 60 K model

However H2CS shows a large difference between the model andobservation as occurs for other S-bearing species The rotationdiagram of HC3N results in an excitation temperature of 17 Kwhich comes between the warm core and halo The best agree-ment of HC3N is found at the 35 K model Our models producesimilar density variations over the ages of the clouds and the bestfits were seen at 35 K or at 10 KSince Orion-KL is the most extensively studied star-forming

region through observations and model calculations our obser-vations and models have been compared with those of Orion-KLThree distinct components having different physical and chem-ical characteristics are found which are hot core plateau andridge (extended and compact Blake et al 1987) The physicalcondition of the hot core of Orion-KL is not consistent with thewarm and cold component of W3(OH) Emission lines from theplateau are broad indicating outflow effects Our survey showsevidence of outflows in W3(OH) from the observations of SiOSO2 and SOWe are not sure whether these emissions are in factcoming from the interaction with outflows whereas emissionsfrom the molecules inW3 IRS 5 seem to be highly influenced byoutflows evidenced by large line widths compared withW3(OH)(Table 1) Then the most interesting region among the threecomponents of Orion-KL is the ridge (extended and compact)because it shows relatively close agreement in temperature withthe warm core of W3(OH) considering the analyses by Blakeet al (1987)

Fig 15mdashTime evolution of fractional abundances of molecules obtained from 35 K model calculations Pure gas-phase chemical reactions only were included butwith different input parameters as shown in the plot

KIM ET AL200 Vol 162

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 43: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

The extended ridge of Orion-KL is dominated by speciescontaining many carbons whereas the compact ridge has morecomplex species includingCH3CNCH3OHHCOOCH3 etc Thetemperature of the compact ridge is 3 times higher and the ex-tended ridge has a temperature 2 times higher than the warm com-ponent of W3(OH) Despite the fact that the detected moleculesin W3(OH) and in the ridge components of Orion-KL are sim-ilar there are differences in heavy molecules such as HCOOCH3

and CH3OCH3 These undetected heavy species are known toform by ion-molecule reaction starting from CH3OH and H2COwhich are abundant in the compact ridge (Caselli et al 1993)CH3OH is detected from our survey but none of the heavierspecies such as HCOOCH3 and CH3OCH3 are detected evenfrom the G343+015 survey of Kim et al (2001) This impliesthat a temperature of 35 K for a warm component is not suitablefor efficient formations of the next generation of CH3OH andH2CO although Caselli et al (1993) assume evaporation of hotcores to occur at 40 K to explain well-known early-time peaks at3 104 yr The possible reasons are (1) precursors CH3OH andH2CO are not abundant in the warm region of W3(OH) (2) thetemperature of 35 K not enough for the efficient evaporation ofthose species from the grain mantles (3) heavy species may beconfined to small clumps making it hard to detect them with ourlarge-beam telescope as the estimated core size of W3(OH) isvery small compared to that of Orion-KL

On the other hand the 10 K model shows relatively highabundances of reactive C-bearing species than other models

These carbon species are observed to be abundant in cold darkclouds such as TMC-1 where gas-phase reaction is dominantWe could not note any significant difference among three models(10 35 and 60 K) because those models assume that gas-phasereaction is particularly dominant in the intermediate- or low-temperature region rather than grain-surface chemistry as ap-plied to our 100 K model However we could obtain highabundances of CH3CN and H2CS in the 60 K model which arepresumably due to thermal effects known to be particularly ef-ficient in star-forming regions (Caselli et al 1993)

Overall the 35 60 and 10 K models with cloud ages of 104ndash105 yr show better agreement with the observations compared tothe 100 K model (Table 7) Helmich et al (1994) estimated thatthe age forW3(H2O) is less than 10

5 yr and is probably of the or-der of several times 104 yr which appears to be close to our es-timation However a significant disagreement in the abundancesof sulfur-bearing species such as SO CS OCS SO2 and H2CSis found Similar substantial disagreements between models andobservations had also been found for several hot cores (Hatchellet al 1998) The chemical abundances of sulfur-bearing speciesare highly dependent on initial injections of H2S from grain sur-face to gas phase (Hatchell et al 1998) but our chemical modelusually adopts the fractional abundance relative to hydrogen mol-ecule to be 106 previously found in the Orion hot core and pla-teau (Minh et al 1990) Reduced initial abundances for H2S from106 to 108 or 107 produce good agreements with observa-tions (Hatchell et al 1998) Our model calculations and those

Fig 16mdashPlot of fractional abundances derived from 10 K model calculations A temperature of 10 K was derived from the rotation diagram analysis of methanol

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 201No 1 2006

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 44: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

TABLE7

TheRatiosofModelFractionalAbundancetoObservedFractionalAbundance

100K

Model

60K

Model

35K

Model

10K

Model

Molecule

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

10

104yr

10

105yr

10

106yr

CO

005

065

404

079

452

631

219

625

631

109

525

628

HCO+

006

007

083

039

076

127

071

129

148

107

195

155

N2H+

031

034

109

017

122

357

102

420

607

178

894

740

SO

021

1667

14286

001

009

36611

005

9921

30635

026

590

1247

CS

013

165

4661

2524

10890

6945

5792

15097

7102

3006

7699

1084

OCS

000

004

044

023

262

038

084

418

080

018

458

145

SO

2

021

32143

435714

000

009

180643

003

13343

100571

010

362

586

HCN

054

2022

236

4061

1869

045

3253

102

060

1766

404

059

HNC

554

4308

446

3857

2472

164

3658

271

207

1980

442

069

C3H2

215

6308

3692

4780

889

003

8231

321

005

143738

13809

123

HC3N

185

10192

4423

782

644

003

1406

447

002

6735

2331

040

CH

3CN

26776

23497

2787

019

105

000

156

002

000

394

034

001

CH

3OH

001

396

4800

19404

30424

003

39944

019

004

14788

432

004

H2CS

288

3960

9600

11632

90160

2479

36412

46240

2404

5612

40760

2512

NotesmdashWhilefractional

abundancesofmodel

calculationsarecalculatedfrom

asingle

lineofsightobservational

resultsareobtained

withthebeam-averaged

columndensitiesDetaileddiscussionsonmodel

calculationsandbeam

averaged

columndensity

arediscussed

inThompsonamp

MacDonald(1999)

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 45: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

conducted by Hatchell et al (1998) therefore suggest that thereduced initial abundance of H2S should be adopted in the chem-ical models for molecular clouds other than Orion-KL Sgr B2and Sgr A

The model code involves major assumptions in the chemicalreaction rates particularly in the desorption of molecules fromthe grains so that they can participate in gas-phase chemistryWhile basic input physical parameters in the model calculationsare properly obtained uncertainties still exist not only in the ob-servations but also in the model calculations in which dynamicaleffects for example are not included (Millar et al 1997) Never-theless we could obtain reasonable agreements between the ob-servations and the warmhalo models in which we turned offgrain-surface reactions

9 SUMMARY

A total of 45 transitions from 17 species in the 1 MHz reso-lution observations ofW3(OH) have been detected in the TRAO3 mm surveys toward W3(OH) On the other hand 68 transi-tions from 28 species are detected from detailed high-resolution(250 kHz) observations inW3(OH) whereas 33 transitions from17 species are detected inW3 IRS 5 Several molecules and tran-sitions missing from an earlier survey do in fact exist such asHNCO [4(0 4)ndash3(0 3)] [5(0 5)ndash4(0 4)] CH3OH [(2(1)ndash1(1)A 0(0)ndash1(1) E ) and SO2 [3[(1 3)ndash2(0 2) 27(3 25)ndash26(4 22)] toward G343+015 These are observed to be weak(03K) except for CH3OH [0(0)ndash1(1)E]] transition indicatingthat those weak lines might have been observed in the marginaldetection limit of previous 1MHz observations (Kim et al 2000)

We estimated column densities from isotopic ratios for mol-ecules where one transition was detected In particular the ro-tation diagram method for HC3N CH3CN and CH3OH whereseveral transitions were observed was used to derive temperatureand column density of the species Lower limits to the columndensity for species having one or two transitions were also esti-mated Significant differences in the chemical properties and abun-dances in the three regions are found G343+015 in comparisonwithW3(OH) andW3 IRS 5 shows a high detection ratio of sat-urated molecules such as CH3OH and CH3CN as well as SO2We also found significant chemical differences betweenW3(OH)and W3 IRS 5 which is consistent with the observations byHelmich et al (1994) and Helmich amp van Dishoeck (1997) whoconcluded that W3 IRS 5 was in a less evolved stage and hasa lower excitation temperature than W3(OH) The core sizes ofW3(OH) and W3 IRS 5 are estimated to be 2 and 13 pc re-spectively from the CS (2ndash1) map

Our major results are summarized as follows

1 The chemical compositions of W3(OH) and W3 IRS 5 arequite different implying they are in different evolutionary stagessupporting the results of Helmich et al (1994)

2 Variations of observed line intensities in the three regions(Fig 10 and Tables 2 and 3) seems to be evident Among thesethree sources G343+015 is found to be richest in organic spe-cies as well as other radical species

3 Species such as SO SO2 and SiO are detected inW3(OH)but SiO is not detected in W3 IRS 5 Species (SO and SO2)detected in W3 IRS 5 may result from shocks induced by out-flows However the non detection of SiO in W3 IRS 5 is a littleperplexing as its abundance is also known to be enhanced in thepostshock gas Therefore more sensitive searches for SiO inW3IRS 5 are required Most of the detected lines in W3 IRS 5 are

from simple linear molecules except for SO2 C3H2 and HNCOsuggesting thatW3 IRS 5 is in a state prior to the formation of thehot core or H ii phase There are significant abundance differencesbetween the observations andmodel calculations especially in theS-bearing species such as SO and SO2 Large line widths of SOand SO2 inW3 IRS 5 imply thatW3 IRS 5 is experiencing shocksfrom outflows in an early evolutionary phase

4 Absorption features in the HCO+ (1ndash0) and CO (1ndash0)spectra of W3 IRS 5 and G343+015 indicates the presence ofcold material along the line of sight or the presence of a sur-rounding cold collapsing envelope For W3 IRS 5 the linefeature of the CO (1ndash0) profile is consistent with the collapsingor infalling cloud model (Zhou et al 1993) whereas it is not forHCO+ (1ndash0) We attributed this to shocks generated by outflowsactive in the earlier evolutionary phase of molecular clouds butmore detailed observations are needed to test this On the otherhand a strong absorption feature in the HCN (1ndash0) and HCO+

(1ndash0) transitions is seen toward G343+015 These are consis-tent with the collapsing cloud model

5 The nondetection of N2H+ toward W3 IRS 5 considering

gas-phase chemical reactions in the earlier evolutionary stage iscontrary to our expectation It is possible that the nondetection isdue to depletion observed in cold dark clouds (Bergin et al 2002)

6 H41 is detected only toward W3 IRS 5 whereas OCSCH3CNH2CS CH3CCH and CH3OH are only seen inW3(OH)The nondetection of H41 toward W3(OH) means that the linerelative to W3 IRS 5 is optically thin Its detection toward W3IRS 5 is presumed to be due to the emission by the compactH ii regions physically associated with W3 IRS 5 A detection ofH41 having broader FWHM (21 km s1) and high energy(Eu frac14 94 K) in the direction of W3 IRS 5 and its nondetectiontoward W3(OH) is far from our expectations We tentatively at-tribute the detection of the recombination line in W3 IRS 5 toseveral compact H ii regions physically associated with W3 IRS5 based on its emission velocity (VLSR frac14 40 km s1) Theultracompact H ii regions W3F or W3M found near IRS 7 andIRS 5 (Colley 1980) are possible candidates

7 The 5 extent of the CS (2ndash1) and HNC (1ndash0) maps ofW3(OH) is estimated to be 06 pc at a distance of 22 kpcAlthough HCN distribution is similar to CS it is slightly elon-gated (Fig 5) In the direction of W3 IRS 5 a large clumpcomparable to that of central clump of W3(OH) is identified at6000 west from the central position The HCO+ (1ndash0) image ofW3(OH) compared to CS and HNC shows rather diffuse dis-tribution having a radius of about 15 pc

8 We confirm that the chemical variation of molecular cloudsin different evolutionary phases is evident as predicted by chem-ical models However we emphasize the importance of moresensitive measurements especially for H2S and for consider-ation of shock chemistry to provide more robust predictions ofS-bearing species

The authors acknowledge the TRAO staff members for theirkind help during the long-term observational period and alsoappreciate many thanks for allowing large amount of telescopetime Without their support help and allowance of enough ob-servational time this project would not be successfully com-pleted S J K acknowledges financial supports from the KoreaAstronomy Observatory and from KOSEF R14-2002-043-01000-0 Y L acknowledges financial support from KOSEFR01ndash2003ndash000ndash10513-0

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 203

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 46: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

APPENDIX

COLUMN DENSITIES

A1 LTE FROM MEASUREMENTS OF OPTICALLY THIN LINES

When optically thin lines are detected we can derive column density and excitation temperature from the relative intensity ratios

TR frac14 f frac12J (Tex) J (Tbg)(1 e ) ethA1THORN

with

J (T ) frac14 T0 eh=kT 1 1

ethA2THORN

where f is the beam-filling factor which is 1 for extended source and T0 frac14 hkFor an optically thick line ( 31)

Tex frac14 T0 lnT0

T R thorn J (Tbg)

thorn 1

1

ethA3THORN

On the other hand if the excitation temperature Tex is known then the optical depth is given by

iso frac14 ln 1 TR

J (Tex) J (Tbg)

ethA4THORN

The optical depth is related to the column density and excitation temperature by

frac14 832SN

3hvQ(exph=kTex 1) expEu=kTex ethA5THORN

The column density of an optically thin isotopic line is given by

Niso frac14 3hvisoisoQ(Tex)

832isoSiso(exp

h=kTex 1) expEu=kTex ethA6THORN

In the high-temperature limit (hBTkT ) diatomic and linear molecules have a partition function

Q(Tex)frac14 kTex

hB ethA7THORN

where frac14 1 for HCN HNC HCO+ and SiO and frac14 3 for HC3N to account for the hyperfine splitting and 3 species such as SOand C2S (Blake et al 1987 Turner 1991) In the high-temperature limit (hATkT hBTkT ) the partition function for the symmetricmolecules CH3CN and CH3OH have the form (Turner 1991)

Qrot frac14 frac12(kTex)3=h3AB21=2 ethA8THORN

where frac14 13 (Blake et al 1987) and frac14 23 (Turner 1991) The CH3OHmolecule has torsionalA andE states Assuming that theyhave the same partition function value

Qrot frac14 Qrot(A)thornQrot(E )frac14 2frac12(kTex)3=h3AB21=2 ethA9THORN

From equation (A5) we can also relate the optical depth of the main line and the isotope by

main

isofrac14 Nmainviso(exp

hmain=kTex 1) expEu(main)=kTex

Nisovmain(exphiso=kTex 1) expEu(iso)=kTex ethA10THORN

KIM ET AL204 Vol 162

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 47: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

From equation (A1) the ratio of the measured intensities R is given by

R frac14 TR main

TR iso

1 emain

1 eiso ethA11THORN

which lets us calculate main

main frac14a log f1 1 exp (iso)frac12 Rg ethA12THORN

From equation (A10) the column density of the main line is given by

Nmain frac14 Nisovmain exp (hiso=kTex 1) exp Eu(iso)=kTexfrac12 isoviso exp (hmain=kTex 1) exp Eu(main)=kTexfrac12 ethA13THORN

and the abundance ratio X is determined by

X frac14 Nmain

Niso

ethA14THORN

A2 LOWER LIMIT COLUMN DENSITY

For molecules having only one detected transition without isotopic species we derive a lower limit to the column densities byassuming that the rotation temperature of linear molecules is Trot frac14 Eu k and for symmetric or slightly asymmetric tops is Trot frac14(23)Euk (see MacDonald et al 1996) Then a lower limit to the column density of linear molecules is given by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e ethA15THORN

and for symmetric or slightly asymmetric top molecules by

Nmin frac14 3kRTR dv

832SgIgKQ(Trot)e

3=2 ethA16THORN

A3 ROTATION DIAGRAM

Column densities and rotational temperatures of molecules with several detected transitions are derived from the rotation diagrammethod with several assumptions Trot 3Tbg the lines are optically thin and all the level populations are in thermal equilibrium (Blakeet al 1987 MacDonald et al 1996)

Nu

gufrac14 3k

RT R dv

832SgIgK ethA17THORN

and

Nu

gufrac14 NT

Q(Trot)expEu=kTrot ethA18THORN

Setting equation (A17) equal to equation (A18) and taking natural logs one finds the relationship

ln3k

RTR dv

832SgIgKfrac14 ln

NT

Q(Trot) Eu

kTrot ethA19THORN

where is the transition frequency is the permanent electric dipole moment k is the Boltzmann constant gI and gK are degen-eracies Eu is the upper state energy S is the line strengthQ(Trot) is the partition function for the rotational excitation temperature TrotNT is total column density and

RTR dv is the integrated line intensity The reduced nuclear spin weight gI and K-level degeneracy

gK of linear molecules are both 1 For C2v species such as CH3CN and CH3CCH however gK frac14 1 for K frac14 0 and gK frac14 2 for K 6frac14 0and gI frac14 12 for K frac14 3n and gI frac14 14 for K 6frac14 3n (where n is an integer Turner 1991)

If several lines are observed it is possible to fit a straight line to a plot of lnNugu frac14 ln NT Qexp(EukT ) against Euk where Nu isthe upper state column density to yield the temperature from the slope and the column density from the intercept Therefore the totalcolumn density can be derived by the relation

NT frac14 exp ( yp)Q(Trot) ethA20THORNwhere yp is the intercept on the y-axis obtained from the fit to equation (A19)

MOLECULAR LINE SURVEY OF W3(OH) AND W3 IRS 5 205No 1 2006

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf
Page 48: A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 ......A Molecular Line Survey of W3(OH) and W3 IRS 5 from 84.7 to 115.6 GHz: Observational Data and Analyses Abstract We have

REFERENCES

Allamandola L J Sandford S A amp Tielens G G M 1992 ApJ 399 134Anderson T amp De Lucia F C 1990 ApJS 72 797Andersson M Askne J amp Hjalmarson A 1984 AampA 136 243Andersson M amp Garay G 1986 AampA 167 L1Askne J Hoglund B Hjalmarson A amp Irvine W M 1984 AampA 130 311Bachiller R 1996a IAU Symp 178 Molecules in Astrophysics Probes andProcesses ed E F van Dishoeck (Dordrecht Kluwer) 103

mdashmdashmdash 1996b ARAampA 34 111Bally J amp Lada C 1983 ApJ 265 824Bergin E A Alves J Huard T L amp Lada C J 2002 ApJ 570 L101Blake G A Mundy L G Carlstrom J E Padin S Scott S L Scoville N Zamp Woody D P 1996 ApJ 472 L49

Blake G A Sutton E C Masson C R amp Phillips T G 1986 ApJS 60 357mdashmdashmdash 1987 ApJ 315 621Brown R D Burden F R Cuno A 1989 ApJ 347 855Campbell M F Butner H M Harvey P M Evans N J II Campbell M Bamp Sabbey C N 1995 ApJ 454 831

Caselli P Hasegawa T I amp Herbst E 1993 ApJ 408 548Cernicharo J Castets A Duvert G amp Guilloteau S 1984 AampA 139 L13Cernicharo J Guelin M amp Kahane C 2000 AampAS 142 181Charnley S B 1997 ApJ 481 396Charnley S B Kress M E Tielens A G G M amp Millar T J 1995 ApJ448 232

Churchwell E Wood D Myers P C amp Myers R V 1986 ApJ 305 405Colley D 1980 MNRAS 193 495Cummins S E Linke R A amp Thaddeus P 1986 ApJS 60 819Dickel H R 1980 ApJ 238 829Dickel H R amp Dickel J R Wilson W J amp Werner M W 1980 ApJ 237711

Dreher J W amp Welch W J 1981 ApJ 245 857Forster J R Caswell J L Okumura S K Hasegawa T amp Ishiguro M1990 AampA 231 473

Garay G amp Rodriguez L 1990 ApJ 362 191Gardner F F Hoglund B Shukre C Stark A A amp Wilson T L 1985AampA 146 303

Goldsmith P F Langer W D Ellder J Kollberg E amp Irvine W 1981ApJ 249 524

Gordon M A 1987 ApJ 316 258Gottlieb C A Lada C J Gottlieb E W Lilley A E amp Litvak M M1975 ApJ 202 655

Gregersen E M Evans N J II Zhou S amp Choi M 1997 ApJ 484 256Guilloteau S amp Baudry A 1981 AampA 97 213Hatchell J Thompson M A Millar T J amp MacDonald G H 1998 AampA338 713

Helmich F P Jansen D J De Graauw Th Groesbeck T D amp vanDishoeck E F 1994 AampA 283 626

Helmich F P amp van Dishoeck E F 1997 AampAS 124 205Herbst E 1978 ApJ 222 508Herbst E Smith D amp Adams N G 1984 AampA 138 L13Imai H et al 2000 ApJ 538 751Irvine W I amp Schloerb F P 1984 ApJ 282 516Jackson J M Armstrong J T amp Barrett A H 1984 ApJ 280 608Jaffe D T Hildebrand R H Keene J ampWhitcomb S E 1983 ApJ 273 L89Jewell P R Hollis J M Lovas F J amp Snyder L E 1989 ApJS 70 833Johansson L E B et al 1984 AampA 130 227Kawaguchi K Kasai Y Ishikawa S amp Kaifu N 1995 PASJ 47 853Keto E Proctor D Ball R Arens J amp Jernigan G 1992 ApJ 401 L113Kim H-D Cho S-H Chung H-S Kim H-R Roh D-G Kim H-GMinh Y-C amp Minh Y-K 2000 ApJS 131 483

Kim H-D Cho S-H Lee C-W amp Burton M B 2001 J Korean AstronSoc 34 167

Kim H-D Ramesh B amp Burton M G 2002 Publ Astron Soc Australia19 505

Loren R B Mundy L amp Erickson N R 1981 ApJ 250 573Lovas F J 1992 J Phys Chem Ref Data 21 181MacDonald G H Gibb A G Habing R J amp Millar T J 1996 AampAS119 333

Mardones D Myers P C Tafalla M Willner D Bachiller R amp Garay G1997 ApJ 489 719

Mauersberger R Wilson T L amp Henkel C 1988 AampA 201 123Millar T J amp Freeman A 1984 MNRAS 207 405Millar T J Herbst E amp Charnley S B 1991 ApJ 369 147Millar T J MacDonald G H amp Gibb A G 1997 AampA 325 1163Minh Y C Ziurys L M Irvine W M amp McGonagle D 1990 ApJ 360136

Nummelin A Dickens J E Bergman P Hjalmarson A Irvine W MIkeda M amp Ohishi M 1998 AampA 337 275

Persi P Busso M Corcione L Ferrari-Toniolo M amp Marenzi A R 1996AampA 306 587

Prasad S S amp Huntress W T Jr 1982 ApJ 260 590Rawlings J M C Taylor S D amp Williams D A 2000 MNRAS 313 461Roberts D A Crutcher R M amp Troland T H 1997 ApJ 479 318Sams B J III Moran J M amp Reid M J 1996 ApJ 459 632Sanders D B Solomon P M amp Scoville N Z 1984 ApJ 276 182Sandford S A amp Allamandola L 1993 ApJ 417 815Schilke P Groesbeck T D Blake G A amp Phillips T G 1997 ApJS 108301

Snyder L E amp Watson W D 1977 ApJ 212 79Solomon P M 1973 ApJ 185 L63Strohacker O C 1978 MS thesis Univ Texas at AustinStrong-Jones F S Heaton B D amp Little L T 1991 AampA 251 263Sutton E C Blake G A Masson C R amp Phillips T G 1985 ApJS 58 341Sutton E C Jaminet P A Danchi W C amp Blake G A 1991 ApJS 77 255Talbi D Ellinger Y amp Herbst E 1996 AampA 314 688Thompson M A amp MacDonald G H 1999 AampAS 135 531Townes C H amp Schawlow A L 1975 Microwave Spectrocopy (New YorkDover)

Turner B E 1989 ApJS 70 539mdashmdashmdash 1991 ApJS 76 617mdashmdashmdash 1995 ApJ 449 635Ulich B L amp Haas R W 1976 ApJS 30 247van Dishoeck E F amp Blake G A 1986 ApJS 62 109mdashmdashmdash 1988 ApJ 334 771mdashmdashmdash 1998 ARAampA 36 317van Dishoeck E F Blake G A Jansen D J amp Groesbeck T D 1995 ApJ447 760

Walmsley C M Churchwell E Nash A amp Fitzpatrick E 1982 ApJ 258L75

Walsh A J Burton M G Hyland A R amp Robison G 1998 MNRAS 301640

Werner et al 1980 ApJ 242 601Wilson T L Johnston K J amp Mauersberger R 1991 AampA 251 220Wilson T L amp Rood R T 1994 ARAampA 32 191Womack M Ziurys L M amp Wyckoff S 1992 ApJ 387 417Wootten A Evans N J II Snell R amp vanden Bout P 1978 ApJ 225 L143Wynn-Williams C G amp Becklin E E 1974 PASP 86 5Wyrowski F Schilke P Walmsley C M amp Menten K M 1999 ApJ 514L43

Zhou S Evans N J II Kompe C amp Walmsley C M 1993 ApJ 404 232Zinchenko I Forsstrom V amp Mattila K 1993 AampA 275 L9Ziurys L M amp McGonagle D 1993 ApJS 89 155

KIM ET AL206

  • University of Wollongong
  • Research Online
    • 2006
      • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
        • S J Kim
        • H D Kim
        • Y Lee
        • Y C Minh
        • R Balasubramanyam
          • See next page for additional authors
          • Publication Details
            • A Molecular Line Survey of W3(OH) and W3 IRS 5 from 847 to 1156 GHz Observational Data and Analyses
              • Abstract
              • Disciplines
              • Publication Details
              • Authors
                  • CDocuments and SettingsnkeeneLocal SettingsTemporary Internet FilesContentIE5SLAZ4DYN61219web[1]pdf