S P E C T R A L S T U D I E S OF M O L E C U L E S OF A S T R O P H Y S I C A L

I N T E R E S T *

M A R K A N D E Y S I N G H

Department of Physics, University of Gorakhpur, U.P., India

(Received 19 October, 1987)

Abstract. This paper has been devoted to a minimum amount of literature necessary for an understanding of the spectra of molecules, molecular ions, and radicals present in the cosmic objects such as Sun, comet, interstellar medium, Earth's atmospheres, long period variables, peculiar late-type stars, nebulae, and planetary atmospheres.

1. General

In recent years the discovery of the presence of various types of molecules, from simple to relatively complex form, in various astronomical objects has excited all scientists alike. It is realised that the chemical processes not only go in stars but also played an important role in the formation of galaxies in the early phase of evolution of the Universe. It is obvious from the diverse nature of objects and their respective conditions that the similar chemical processes are not responsible for occurrences of a particular molecule in different objects. The underlying physics, however, cannot be different, but differences in physical and dynamical situations in different objects have their effect on the observed molecules. The compilation of 583 references available in literature of 153 diatomic molecules/molecular ions/radicals of astrophysical interest and the respective scientific information contained therein has been presented for each molecules such as the dissociation energy, spectral region, transition levels, astrophysical objects where the respective molecule have been detected (say, comet, meteorite, Sun, planet, star, interstellar matter, Galaxy, etc.); computed theoretical parameters (i.e., FCF's, tran- sition probabilities, ~-centroids, PE-curves), and available laboratory data with the respective references in one of our papers (Singh and Chaturvedi, 1987).

2. Laboratory Astrophysics

Laboratory astrophysics is usually defined as those branches of physics which are of interest, or of potential interest, for astrophysics. All information about the composition of stellar atmospheres, of planetary nebulae of the interstellar medium, of comets, and until very recently, also of planetary atmospheres, has come from the study of the spectra of these objects. It is not surprising to find that there has been a close connection and interaction between laboratory spectroscopy and astrophysics. Thus the interplay is obvious.

* Astrophysics and Space Science Review Article.

Astrophysics and Space Science 141 (1988) 75-101. �9 1988 by Kluwer Academic Publishers.

76 M. SINGH

Spectra of astrophysical origin are due to conditions which are usually widely differ- ent from those normally existing in terrestrial sources. Variations in pressure, tempera- ture, concentration - one or more of these - will give rise to differences in spectra. Attempts have been made to simulate these conditions in the laboratory with a view" to identifying the spectra of astrophysical sources and studying the mechanism that leads to a particular spectrum. Our knowledge of extra-terrestrial sources is solely through the study of the radiation, we receive from them. The spectroscopic features of this radiation, compared with those of terrestrial sources lead to the identification of the emitter and also to the structure and composition of the celestial source and the analysis of this radiation gives also information about the mass, size, pressure, temperature, magnetic field, and motion of extra-terrestrial bodies as well as the intervening space. However, these conditions are very often different from those normally existing on the Earth. This leads to changes in the spectral features and line intensities and the consequent difficulty in identification. It has been found that by a suitable adjustment of the experimental conditions spectra occurring in extra-terrestrial sources can be produced in the laboratory also. Different temperatures and pressures, at which matter in astrophysical sources exists, are responsible for the observation of spectra of simple polyatomic and diatomic molecules or radicals, their ions and neutral atoms as well as their ions (sometimes a dozen or more electrons removes). For purposes of identifi- cation it is, therefore, necessary to obtain these spectra in the laboratory and compare the two sets. One can easily realize that the laboratory sources have to be of widely different types so as to be able to give the wide range of spectra. The problem of obtaining the desired spectra in the laboratory becomes difficult in view of the fact that many of the spectra of astrophysical origin involve forbidden transitions and it is not always possible to obtain them in the laboratory conditions favourable for their appear- ance. The study of spectra in the infrared, visible, and ultraviolet regions, the vacuum ultraviolet and soft X-ray region having now becoe important with the development of satellite and lunar observations. It may be mentioned that such studies are useful not only from the point of view of astrophysics but also for obtaining better understanding of matter and of the laws of atomic and molecular physics. The discovery of the C 3 molecule may be cited as an example. And there are many others, which show how study of astrophysical and laboratory spectra have helped in the overall growth of knowledge.

It is noteworthy that astrophysical molecular spectra are very often due to radicals or forbidden transitions of simple molecules. Quite a few of the spectra of astrophysical origin have not yet been identified and explained. Notable among these are the diffuse interstellar bands which have been known for more than thirty-five years but have not been identified. It is necessary to obtain them in the laboratory so that they may b e

identified. Study of laboratory spectra and determination of atomic and molecular constants are

also often helpful in identification of a spectrum of astrophysical origin even though the actual transition involved may not be obtained in the laboratory - one can build the spectrum from the constants determined and then compare the two. Moreover, knowledge of the constants can be used in explaining abundances or absence of certain

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 77

species. Thus accurate determinations of excitation and dissociation energies and ionization potentials become necessary. If one knows two of the three quantities, oscillator strength (i.e., f-value of a transition) the observed intensities in absorption and its abundance, then the third can easily be determined. If f-values, transition probabilities, P-centroids, and PE curves are determined from laboratory studies, abun- dances in astrophysical sources can be found from their observed absorption spectra.

One of the most exciting recent developments in astronomy has been the discovery of molecules in the cosmic space. These have been detected by searching for spectral bands and lines at appropriate wavelengths.

The intensity distribution of molecular bands in an electronic band system depends upon source conditions and molecular constants. The importance of the laboratory molecular spectra in view of its relevance to many astrophysical problems involving the estimation of the physical conditions (viz., temperature, pressure, density, and abun- dance) of the emitter in the various cosmic sources (i.e., stellar atmosphere, aurora, night sky, and compact objects such as pulsars and neutron stars) from the relative intensities of the bands of a molecule and radical, it is necessary to have a theoretical knowledge of the corresponding vibrational transition probabilities, P-centroids, and PE curves for the respective bandheads in a band system. The knowledge of vibrational transition probabilities, ?-centroids and PE-curves is essential for the interpretation of spectral intensities in terms of source conditions. A number of workers have, therefore, under- taken theoretical studies to provide vibrational transition probabilities, ~-centroids and PE-curves, etc., of the molecules which are of importance in astrophysics, gas kinetics, molecular spectroscopy, and combustion processes.

3. Spectral Studies of the Sun

The foundation of a quantitative study of molecular lines in the stellar spectra was laid by Russell (1934) who could satisfactorily explain the intensities of molecular bands in the solar spectrum.

These ideas were further elaborated and improved upon in subsequent investigations by Roach (i939), Babcock (1945), Hunaerts (1947), and Minnaert (1949). The theory of weighting functions proposed and used for weak atomic lines was extended by Minnaert (1949) to the interpretation of molecular spectrum also. Prior to this work, all the absorption lines were assumed to originate in a 'thin reversing layer' lying just above the photosphere. This was the first publication about molecular lines in which the reversing layer and photosphere are replaced by a realistic atmosphere, where the temperature and pressure vary from layer to layer, hnportant information has been obtained about the physical conditions of the higher layers of photosphere by studying the C - L behaviour of molecular lines (Pecker, 1957; Laborde, 1961; Kozhevnikov and Polonskij, 1969). Owing to these efforts, a systematic method is now available with the help of which the observed equivalent widths of molecular lines can be used to know the elemental abundances of the constituent atoms. This method requires consideration

78 M. SINGH

of the number of atoms actually forming a certain molecule. Careful selection of dissociation constants, partition functions, and oscillator strengths is necessary for calculating the intensities of molecular lines. This method has been considerably improved upon and used for the interpretation of the molecular lines in the solar spectrum by various investigators (Schadee, 1964; Sotirovski, 1971).

The data of the 153 diatomic molecules and available molecular transitions have been listed in Table I of one of our papers (Singh and Chaturvedi, 1987) and some have also been observed in the solar photospheric spectrum (Pande, 1972; Mallia, 1974).

Also the Lyman bands (B 1 Z+ _ X leg ) of H 2 molecule have recently been identified in emission in the ultraviolet spectrum (2 ~ 1400 ~) of the solar photosphere (Jordan et al., 1977). The intensity of these lines is enhanced in the umbral spectrum.

In Table I by Singh and Chaturvedi (1987) we have included the identification of CH + by Grevesse and Sauval (1971) and that of Phillips bands ofC 2 by Lambert and Mallia (1974). The main difficulty in the two investigations is that the considered lines are very weak and also that wavelength coincidences have been based on the calculated wavelength of uncertain accuracies. Grevesse and Sauval (1971)have obtained some coincidences for CH + fines but the difference between the solar wavelengths and the wavelenghts calculated on the basis of molecular constants is very high (> 0.2 A).

Although Dunham (1937) and Adams (1939, 1941) detected the first molecules (CN, CH ~-, and CN) in interstellar space, and also in photospheric spectrum, the field was for the most part ignored until Weinreb et al. (1963) discovered the 18 cm transitions of OH.

Mallia (1974) has listed the molecule SH as amongst those already detected in the photospheric spectrum but the details of the identifications are yet unpublished.

In Table I by Singh and Chaturvedi (1987) we present a list of molecules observed in the sunspot spectrum. In a recent publication, Wing et al. (1977) have attempted to confirm the presence of the Fell molecule in the sunspot spectrum on the basis of the probable (0-0) and the (1-0) bands. A laboratory analysis regarding the electronic transition and the band designations is not available. The authors conclude that about 80 ~o of the Fell wavelengths obtained for these bands in the laboratory can be detected with certainly as unblended solar lines in the sunspot spectrum.

A similar analysis for the photospheric spectrum, however, gave them a negative result in contrast to the conclusions by Carroll et al. (1976).

The strengths of the 8690 and 9890 A bands of the Fell molecule are measured in the spectrum of a large sunspot (W6hl et al., 1983). The strongest lines attain central depths of about 20~o of the continuum intensity. The Fell lines are not detectable in the photospheric spectrum (W6hl et aL, 1983).

The band structure of the Fell molecule has defined analysis. The spectrum contains a large number of highly spaced lines which is typical for hydrides. The laboratory spectrum is well observed and contains evidently several bands (McCormack and O'Connor, 1976; Klynning and Lindgren, 1973; Davis, 1980). The blue (224865 - 4973 A) and green (~,5286 - 5355 A) were tentantively identified in the solar spectrum (Carrol et al., 1976) using wavelengths tabulated by Moore et al. (1966). The

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 79

infrared bands starting at 8690 A were reported to be present in sunspots (Wing et al., 1977). No detailed measurements have yet been made.

In the laboratory spectra of McCormack and O'Connor (1976) the 'green' band appears substantially stronger than the infrared band. This seems to be the case also in the spectrum of a sunspot available to the authors, WOhl et aL (1983) which shows Fell lines of medium strength in the neighbourhood of 25250 A. The strengths of Fell lines have presently been measured in a high-resolution spectrum of a large sunspot (Roma number 6751). The observations were made with the FTS (Fourier Transform Spectrometer) of KPNO (Brault, 1978). The spectral scans were done on 31 March, 1976, when the spot was located nearly in the centre of the solar disk. Further details of this spot are given by Engvold et al. (1980).

The umbral spectrum is contaminated by stray light from the penumbra and from the photosphere. The infrared part of the spectrum is least affected by stray light, where it evidently is less than 20~o of the umbral intensity (Engvold et aL, 1980).

3 .1 . A S T R O P H Y S I C A L S I G N I F I C A N C E O F T H E C 2 M O L E C U L E

Sinha (1978) deals with the possibility of detecting the various electronic transition of the C 2 molecule in solar spectrum, he gives a brief resume of the latest developments concerning the observation of these transitions in different astrophysical sources.

In spite of the familiarity of astrophysicists with the C 2 molecule for well over hundred years the interest in this molecule has been renewed from time to time. Molecular equilibrium calculations carried out by Schadee (1968) indicate that in the atmospheres of oxygen-rich subgiant and giant stars of spectral classes later than G5 the concen- tration of C 2 decreases with temperature due to increased depletion of carbon atoms to form the CO molecules. Branch (1969) argued that since the sunspot spectrum is similar to that of a late K dwarf, it might be expected that the bands of C 2 would appear even weaker in the sunspot spectrum than in the photospheric spectrum. Subsequently he showed that the identifications of Swan bands of C 2 molecules due to Richardson (1931) and Laborde (1961) in the sunspot spectrum are erroneous and that the said bands are actually absent. Further investigations have supported this viewpoint (Schadee, 1970; Sotirovski, 1971; Harvey, 1972; Lambert and Mallia, 1973; Sinha and Pande, 1974) and it has been indicated that isotropic MgH lines could give a misleading impression about the presence of the Swan bands in the sunspot spectrum.

It is interesting that till recently no band other than the Swan band of C 2 molecules was discovered in the spectra of a variety of celestial objects. Based upon the identifi- cation of XIZg state as the ground electronic state of the C 2 molecule, Ballik and Ramsay (1963) suggested the desirability of an investigation of the Phillips bands in the spectra of comets and interstellar medium. It is only now that this band has been detected in the interstellar cloud Cyg OB2 No. 12 (Souza and Lutz, 1977). Efforts for its identifications in the cometary spectra were made but no positive results have come up (Wyller, 1962; Krishnaswamy and O'Dell, 1977). Another transition which arises from the ground state of the C 2 molecule is the Mulliken band system. Smith has tentatively concluded the presence of this band in the spectrum of comet Kohoutek

80 M. SINGH

(Feldman, 1977). The Ballik-Ramsay bands form another interesting transition of C 2 molecules. These bands have been detected in the spectrum of carbon stars.

3.2. O N THE PHILLIPS SYSTEM IN THE SOLAR SPECTRUM

Sinha (1978) theoretically found that under LTE, weak spectral features of the (0, 0) band of the Phillips system might be present in the solar atmospheric spectrum. How- ever, having failed to detect these features in the infrared atlas due to Hall (1974), he infer that the calculated equivalent widths are probably exaggerated. The cause for larger predicted equivalent widths than found in the infrared tracings by H all (1974), then may lie in the uncertainties in the choice of the band oscillator strength in the atomic abundances and in the model atmospheres chosen (Smith, 1969). The values of the carbon abundance as reported by various investigators are conflicting and also the available electronic transition moment measurements through which band oscillator strengths can be calculated cover a wide range of values.

The fact that the (0-0) band of the Phillips systems is not identifiable in the photospheric spectrum, leads us to doubt the tentative identification by Lambert and Mallia (1974) of a wear (1-0) band. The viewpoint of Sinha (1978) is further supported by the fact that for wavelength calculations, Lambert and Mallia (1974) have used molecular constants whose dependability for such calculations involving vibrational states with v' < 2 is not beyond doubt. However, the solar features identified by Lambert and Mallia (1974) would require an alternative explanation.

To resolve the difficulties encountered in the identification of the Phillips bands, Sinha (1978) suggest that more precise and reliable data on molecular constants, band oscillator strengths, and the solar carbon abundance be obtained. Furthermore, low noise centre-to-limb scans of the spectral region occupied by Phillips bands may also prove helpful.

3.3. MULLIKEN BANDS IN THE SOLAR SPECTRUM

The problems regarding the identification of the Mulliken system in the photospheric spectrum are LTE calculations carried out by Sinha (1978) show that weak features, attributable to the lines of the (0-0) band of this sytem might be present in the ultraviolet spectrum of the solar photosphere around 230 nm. In the calculations, Sinha (1978) used the/Re(?)[ 2 value reported by Cooper and Nicholls (1975) which is one-third of the value reported by Smith (1969). The former authors, however, do not elaborate upon this point. Sinha (1978) feel, therefore, that a further laboratory verification of ] Re(~l 2 value is desired.

In a crowded region, such as the ultraviolet part of the photospheric spectrum is, the detection of such weak features would demand much caution. Sinha (1978) suggest that prior to such an effort, a synthetic spectrum of the region under investigation should be constructed to facilitate the identifications. It is encouraging to note that S auval et aL

(Biemont, 1978) report to have identified features as weak as 0.3 mA in the adjacent spectral region around 300 nm.

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 81

3.4. O N B A L L I K - R A M S A Y A N D F O X - H E R Z B E R G B A N D S Y S T E M I N T H E S O L A R

S P E C T R U M

For the K = 35 line of the Q3-branch of the (0-0) band belonging to the Ballik-Ramsay system, Sinha (1978) obtain W = 2 mA, utilizing the VAL-76 model. The results for the other models, viz., HSRA-71 and BCA-68 are successively higher. However, for the band under consideration, Hall's (1974) infrared tracings do not cover Q-branch lines which originate from K-values in the range 25 to 65; the K-value for maximum population of rotational levels is N 35. So, Sinha (1978) made an effort to identify the lines of the R-branch which is half as intense as the Q-branch and which is rather well covered in these tracings. His effort gave a negative result (Sinha, 1978). Also the position on the solar disc at which Hall (1974) obtained his tracings are not known and Sinha (1978) assume that they are for positions near to the centre of the solar disc.

Since molecular line formation takes place essentially in the upper photospheric layers, these lines get enhanced in the C - L observations. Sinha (1978), therefore, recommend low noise centre-to-limb observations for the Ballik-Ramsay bands in the photospheric spectrum.

Recently, Sauval et al. (Biemont, 1978) have reported the identification of weak features of the Q-branch lines of the (0-0) band of the Ballik-Ramsay system in the photospheric spectrum. They, however, do not give the observed intensities of these features and he believes that it should be near to the threshold for detection in Hall's (1974) tracings, i.e., 1 mA. The reported identification seems to be in agreement with his interference regarding the absence of the weaker lines of the R-branch. It may be mentioned that these authors too compute a maximum equivalent width of 2 mA for the lines of the Q-branch. They further state that the computed equivalent widths based upon Cooper and Nicholls (1975) measurements for ]Re(PI 2 are in agreement with observations. Sinha (1978) is, however, of the opinion that since such calculations are sensitive to the chosen carbon abundance and the model atmosphere, this agreement should be accepted cautiously and that a laboratory check upon the Re(P[ 2 value given by Cooper and Nicholls (1975) is desirable.

Based on the semi-qualitative discussions Sinha (1978) concludes that the Fox-Herzberg band system is too weak for detection in the photospheric spectrum.

3.5. SULPHUR MOLECULES IN THE SOLAR ATMOSPHERE

The principal result obtained by Sinha (1978) is that out of a probable list of SH, CS, SO, NS, and $2, the molecule SH can show weak features in the near-ultraviolet part of the photospheric spectrum. The calculations are based upon a theoretical estimate of band oscillator strength made by Henneker and Popkie (1971). The dependability of this estimate has been discussed and suggest that a laboratory determination of the oscillator strength is also needed so as to enable one to match the model based equivalent widths with observations.

On the basis of a comparison of abundance calculations in a photospheric and a sunspot model, Sinha (1978) expressed the possibility of detecting the vibration-rotation

82 M. SINOH

transitions of SH and SO molecules in the infrared umbral spectrum. For the identifi- cation of these transitions one needs the vibration-rotation spectrum of SH and SO molecules and also the corresponding transition probabilities. A low noise tracing for the concerned infrared region of the umbral spectrum is also needed.

For the molecules CS, Sinha (1978) obtain equivalent widths ~ 1 m~ for the centre of the solar disc. This result is based upon the calculations for the photospheric HSRA-71 model and the carbon abundance cited therein. The use of the cooler VAL-76 model reduces the number of free carbon atoms available for the formation of CS molecules. Also the carbon abundance recommended by Mount and Linsky (1975) and used in the chapters concerning the various band systems of C a molecules, is substan- tially lower than the HSRA-71 value. It may be added that recently, Tsuji (1977) has obtained a good agreement between observations and predictions for the CO lines by utilizing the VAL-76 model and still lower carbon abundance logN(C) = 8.28 which, if accepted, will further reduce the expected equivalent widths of the considered lines of CS. In brief, Sinha (1978) feel that it may be very hard to detect the lines of the CS molecule in the photospheric spectrum.

However, it may again be emphasized that Sauval et al. (Niemont, 1978) have reported an identification of features as weak as 0.3 m,~ in a nearby region of UV spectrum, i.e., near 300 nm. Therefore, low noise C - L observations alone can resolve the problem of the presence or the absence of CS molecule in the photospheric spectrum.

The molecules SO, NS, and S z appear too weak for detection in the photospheric UV spectrum, since they have much lower molecular abundances as compared to SH and CS. However, to derive the dependence of the electronic transition moment upon the internuclear distances, band oscillator strengths for the (B3E- - X3~]- ) transition of the SO molecule involving lower vibrational states are needed. Such an investigation might help in the study of absorption produced by SO molecules in the solar spectrum. Also the dependendability of the molecular constants in the calculation of the wave numbers for the CS molecule should be assessed through more detailed laboratory investigations. Sinha (1978) found discrepancies to the extent of ~ 5 cm-1 in the calculated and the observed wave numbers utilizing the molecular constants given by Crawford and Shurcliffe (1934) and also those compiled by Suchard (1975).

Sinha (1978) would like to mention that in a search for molecular lines which includes SH lines also in the near-ultraviolet spectrum, Biemont (1978) conclude, ". . . all the bands analyzed are surely not absent and are very probably present in the solar disc spectrum". This supports predictions about the presence of SH molecule in the photospheric spectrum made by Sinha (1978) earlier.

4. Cometary Physics

In the past decade or so, scientists have studied the light emissions of numerous comets in various portions of the electromagnetic spectrum.

Comets are generally supposed to have three distinct parts: (i)A nucleus of small dimensions generally of the order of a few kilometres or less. It is a collection of rocky

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 83

and metallic particles coated with frozen ices of CH4, CO2, H20 , and NH 3 . The nucleus of the comet has been looked upon as a collection of meteorites. (ii) A head or coma surrounding the nucleus and extending to perhaps hundred or thousands of kilometres. (iii) A tail directed away from the Sun, the length of which may well be the order of millions ofkilometres. The nucleus is composed of rocky or metallic material and frozen water, ammonia, methane, and carbon dioxide. This slushy material is similar to the one believed to be present in the original solar nebula of which the planets were formed. The nucleus is the central bright part of the comet and it is usually too small to allow an exact measurement of its size. Halley's comet is believed to have a nucleus of diameter about 6 km. The coma is like a halo that surrounds the nucleus and can have a very large size. In the case of Halley's comet, the diameter of the coma at the perihelion is nearly 450000 km (Bhatti, 1985).

The halo is formed from the nucleus, as the comet approaches the Sun, due to heating effect and the evaporation of frozen matter. The spherical luminous envelop halo (coma) that forms the head of a comet - which scientists have termed a 'chemical factory' where molecules from the nucleous are continuously being converted into chemical compounds, ions, atoms, and other substances like H + (UV, V), O + (UV, V), C § (UV), C + + (UV), S + (UV), Na + (V), K + (V), Ca + + (V), Cr + (V), Mn + (V), Fe +

(V), Ni + (V), Cu + (V), C 2 (UV, V), C 3 (V), CH (V, R), CH3CN (R), CN (V), CO (UV, V), CS (UV), u c n (R), NH (V), NH 2 (V), OH (UV, V, R), S 2 (UV), U20 (R), NH 3 (R), CH + (V), CN + (UV), CO + (UV, V), C02 (UV, V), H20 + (V), N f (V), OH + (V), etc. (Basu, 1986).

The chemical processes start only as the comet approaches the Sun when the Sun's heat starts heating up its surface. Soon, gases start escaping from the nucleus taking solid dust along with them. It is these escaping gases and dust that make up the coma which often reaches a diameter of 100 000 km or more. In the course of time, as more gases and dust escape due to heat, their interaction with the solar winds and the pressure

of solar radiations gives rise to the spectacular 'tail' of comets. The tail length for Halley's comet often stretches from 65 to 130 million km. The solar winds consist of charged particles like electrons, protons, and nuclei of light elements, give offby the Sun at very high speeds. These winds and the high pressure of the other radiations emitted by the Sun split the neutral molecules or atoms of materials in the coma to produce glowing streams of charged or ionized particles. These include hydroxyl, cyanide, nitrogen, carbon, and calcium ions. Besides, silicate dust is also detected in comet tails.

By analyzing the light reflected from the tail it is possible to determine its chemical constituents, but that does not necessarily provide a clue to the actual composition of the nucleus. This is because, after ejection from the nucleus, the molecules are greatly modified by solar radiation which makes the composition of the tail different from that of the original ejected matter. For example, H20 molecules have never been observed in the tail; its presence has been inferred only from the discovery of H20 + ion and the study of the abundances of O, H, mad OH. The relative abundance of carbon also remains unexplained; younger comets seem to have more carbon than older ones which

have traversed the neighbourhood of the Sun several times. The origin of minor

84 M. SINGH

components discovered in the tail, such as CN, C2, and C3, etc., still remains a total mystery.

As the comet moves away from the Sun the gases condense. So that the tail disappears.

Comets have always been a challenge for spectroscopists physicists and astronomers. The major interest of cometary research comes from the fact that comets are among the most primitive objects of the solar system. Since cometary nuclei were probably formed in the outer region of the Solar System, and have presumably not evolved since their formation, these objects may be considered as samples of the primitive solar nebula. It is generally accepted that cometary nuclei, unlike planets and asteroids, remained cool during and after their accretion process. This has allowed these objects to retain much of the original content of the solar nebula, and presumably of the interstellar medium. Indeed, their cold environment and their size prevented any physical fractionation or chemical evolution. Thus any comet entering the inner solar system is a unique opportunity for cosmogonical studies. Therefore, one of the main goals of cometary investigations is the determination of the chemical composition and physical state of these fossil bodies. Besides in situ observation by mass spectrometers aboard the space probes which recently flew by comets Giacobini-Zinner and Halley, or in a remote future, by nucleus sample return missions, the only remote-sensing technique which can achieve this goal was - and remains - spectroscopy. Composition determinations have to be based on spectroscopic observations of the cometary atmosphere which develops when a comet visits the inner Solar System. Over the past century, a considerable amount of information has been obtained by visible spectroscopy of comets, and more recently by UV spectroscopy, but these observations concern almost uniquely the radicals, atoms, and ions produced by photodissociation or photoionization of the parent molecules, which are directly sublimed from the nucleus or formed by chemical reactions in the very inner coma. These parent molecules generally cannot be observed in the visible or UV spectral ranges, because their electronic bands either are weak or lead directly to dissociation rather than fluorescence when they are excited by solar light. Therefore, these parent molecules have to be searched for through their vibrational transitions in the infrared or their rotational lines in the microwave or submillimeter ranges. Since the energy involved in these transitions is small (Crovisier, 1986).

Many molecular species were observed in visible cometary spectra well before they could be available in terrestrial laboratory. When a comet visits the inner solar system, its nucleus is heated by the Sun, volatile materials is sublimed and an atmosphere - the coma - develops. Gravity is not sufficient to retain this atmosphere, and the coma expands. Stable molecules directly sublimed from the nucleus ('parent molecules') may undergo chemical reactions in the inner coma, where the density is high. Solar ultraviolet radiation dissociates these molecules, which form radicals, atoms, and ions. These secondary products are called 'daughter molecules'. The reviews of physical processes in comets have been given by Wilkening (1982) and Mendis et al. (1985). Solar radiation induces fluorescence processes and helps to study molecules present in the coma. Because of its small dimension and low temperature, the nucleus of a comet is extremely

SPECI'RAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 85

difficult to observe, apart from recent radar measurements (Kamoun et al., 1982), this has been no direct investigation of cometary nuclei at the present time. Up to now, most of the cometary spectroscopic observations deal with 'secondary' material, i.e., products

of the photodissociation of parent molecules outgassing from the nucleus. All these radicals are observed in the UV, visible or radio ranges. There is also a little evidence of the nature of the parent molecules. Although H20 seems to be the most plausible candidate - on the basis of the measured OH and H production rates and the detection of H20 + - there is no complete model for any distribution of parent molecules which can account for the observed relative abundances of radicals.

An important feature of comets is the plasma tall which is composed of photoioni- zation products dragged by the solar wind. Species such as H20 +, CO2 + , and CO + are currently observed in the visible spectrum of plasma tails (Crovisier, 1986). Micro- wave observations of these species could yield important informations on their kinematics and reveal how they are accelerated by the solar wind.

In the inner coma, the most important chemical reactions are likely to occur through charge-exchange reactions between neutrals and ions. The H30 + ion is presumably abundant due to the predominance of water and should be an important key to cometary chemistry.

The comet is in orbit with respect to the Sun. However, its path is such that it spends most of its time at great distances from the Sun. In recent years considerable popular interest occurs when its path in the orbit brings it sufficiently lose to be seen from the Earth. When the comet's position in its orbit is such that it swings in towards the Sun, its surface begins to sublime or evaporate, and, as a consequence, gas flows into space. Since the nucleus of the comet exerts only a weak gravitational effect, the outgoing molecules and atoms carry with them solid particles. Hence, by these processes the gaseous and dusty clouds are created.

The brightness of the comet may be attributed to (i) a process where the ultraviolet radiation from the Sun dissociates molecules, and then the resulting species can absorb solar radiation and emit it; (ii) sunlight is scattered by dust in the comet.

The spectra observed from comets results from the gases in the head and tail. These gases are released from the solid particles under the influence of solar radiation. Cometary spectra consists mainly of a large number of strong emission bands. However, when a comet approaches the Sun, the sodium D-lines are observed from the comet and may have a very high intensity when the comet-Sun distance is small.

The band spectrum from the head and nucleus is very different from the occurring in the tail. The species observed from their band spectra in the heads of comets are C2, CN, C3, OH, NH, CH, CH +, and NH 2, while in the tails CO +, N~, CH +, and CO, + have been identified. The identification of the species OH, NH, CN, Cz, and C 3 in comet 1940I may be seen from Straughan and Walker (1976).

The intensity distribution within these cometary bands differs from those found in the laboratory sources even at low temperatures. The irregular distribution is very evident in, for example, the CN band near 3880 ,~ and also less prominent for the CH, OH, and CH + bands. Instead of showing a smooth rotational intensity distribution, cometary

86 M. SINGH

spectra reveal complex band profiles with intensity minima in the branches. These minima appear even when the bands are incompletely resolved. The positions of the minima correspond perfectly to the wavelengths of strong absorption lines in the exciting sunlight, the absorption being due to the solar atmosphere. In addition, the radial velocity of the comet relative to the Sun affects considerably the rotational profiles, since the positions of the minima are altered owing to the change in wavelength of the solar absorption lines in accordance with Doppler principle.

From a study of the rotational structure of a band, estimate of the effective rotational temperature may be made. However, allowance has to made for the effect of solar absorption. When this is done it is found that CH, OH, NH, CN, and CH + have, on the whole, low rotational temperatures in the order of 200-400 K depending on the distance of the comet from the Sun. A study of the rotational intensity distribution in a C 2 band, however, lead to much higher rotational temperatures of about 3000-4000 K. These differences in rotational temperatures may be appreciated from the following facts: (i) The rotational distributions are not brought about by collisions since collisions will be very infrequent at the low pressures prevailing. (ii) The distribution of rotational energy is a result of the continued absorption of sunlight, and whereas the low rotational temperature type may undergo transitions in the far infrared, since they have a permanent electric dipole moment, the Ca molecule being homonuclear cannot dissipate its rotational energy in this manner. C 2, therefore, has a high rotational temperature, while the CH, OH, NH, CN, and CH + species have a low rotational temperature. The comet-Sun distance will have an important effect on the rotational distribution.

The comparison is made between the spectrum of comet 1940I and acetylene flame spectra (Straughan and Walker, 1976). The spectrum of the comet was obtained by Swings using an f/1 quartz spectrograph when the comet-Sun distance was 0.87 AU. By comparing spectrogram, the various bands due to OH, CH, NH, CN, C 2, and C 3 may be identified in the comet. In addition, it can be seen that the rotational and vibrational intensity distributions in the bands of OH, CH, NH, and CN are extremely different in the comet from those in the flames, while they are similar for the bands of C2. The observed cometary fines within each band of OH, CH, NH, and CN correspond to electronic transitions from the lowest rotational levels. The details of the acetylene flames are as follows:

(i) The spectrogram is a result from a rich acetylene flame with some addition of ammonia. The main features are bands of OH, NH (Q-branch), CN (violet system), C3, C2, and CH.

(ii) The spectrogram is a result from an acetylene flame with traces of ammonia. The following bands are present in the inner cone of the flame: CH, NH, CN, CH, and C 2. In the outer cone of the flame are present bands due to OH, NH, CN, CH, and C 2.

(iii) The spectrogram is a result from an acetylene flame with ammonia. The rotational structures of OH, NH, CH, and CN are clearly shown.

The abundance of molecules in comets can be estimated spectrophotometrically, although little work has been done along these lines. A very approximate estimate of C 2 molecules in the head of a comet gave 105 molecules per ml, while in the tail an

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 87

estimate of CO + was found to be one molecule per ml. The effective vibrational and rotational temperature values may differ considerably and have little physical meaning. However, for systems at equilibrium these values should correspond. The lack of agreement serves to show that the prevailing densities are too low that thermal equilibrium in cometary systems cannot be established by collision process.

The upper limit for the total density is of the order of 106 molecules per ml. The fundamental part of a comet is the collection of solid meteorites in the nucleus

where almost the entire mass is concentrated, and from which the head and tail of the comet originate. The head and tail of a comet are formed by liberation of gas and dust from the nucleus, and since the nucleus has a small mass, that is possessed little gravitational attraction, the head and tail need to be continuously replenished; otherwise they would disappear within a few days. As the comet approaches the Sun the nucleus emits large quantities of gas and dust and eventually disintegrates. Sometimes the whole comet breaks up, and if fragments of it (meteorites) come under the gravitational attraction of the Earth, then an opportunity of examining the fragments of former comets in the laboratory results. Molecules in a comet have a limited life in the radiation field of the Sun, depending on the possibilities of photodissociation and photoionization and on the amount of solar radiation available in the spectral region where dissociation or ionization occurs.

The spectra of comet Cunningham (1940c-1941I) in the region 3000-6400 ~ are given by Straughan and Walker (1976). The spectrograms 1 to 10 inclusive were obtained at the following heliocentric distances: 2.24, 1.85, 1.47, 1.18, 1.03, 0.87, 0.75, 0.63, 0.50, and 0.48 AU, respectively. The species detected are marked along the bottom edge of the figure. The resolution of the rotational structure of the OH and NH bands in spectrograms enabled for the first time identification of these two radicals in comets. A study of comet 19481 showed that the diameter to which radicals extend in the comet head varied with the heliocentric distance, e.g.:

f the diameter of CN is 62000 km, F 0.90 AU

(the diameter of C 3 is 4140 km ;

~the diameter of CN is 166000 km,

r = 2.21 AU (the diameter of C 3 is 22000 km.

Over the comet-Sun range 0.65-2.21 AU the profile of the CN band shows appreciable variation; the main systematic variation in profile of CN is due to its decrease in rotational temperature.

One interesting more recent study was of comet Bennett in 1970 by the orbiting Astronomical Observatory when it was noted that: (a) a large cloud of hydrogen radiated at the wavelength of the Lyman a-line; (b) the far ultraviolet absorption of the hydroxyl radical is much stronger than had been suspected; (c)there appeared to be approximately equal number of hydrogen atoms and hydroxyl radicals, which was taken to indicate that ice was a major constituent of the comet's nucleus.

88 M. SINGH

Not by any means are all the features of comets understood; even the formation of the ion tail is still open to speculation. One view is that, in addition to ionization brought about by solar radiation, the high-energy electrons in the solar wind also ionize the molecules in the coma. In addition the solar wind sets up a bow wave around the coma. Chaotic magnetic fields are created which act as a magnetic rake that selectively detaches ions from the coma. The non-ionized molecules and atoms are not so influenced and are left behind.

4.2. EXCITATION AND EMISSION PROCESSES OF MOLECULES IN COMET

The prevailing molecular excitation mechanisms of parent molecules in comets are governed by (a) radiative excitation of the fundamental bands of vibration by the solar infrared field; (b)thermal excitation by collisions in the inner coma; (c) excitation by solar radiation field (plus the nucleus and dust radiation field in the inner coma); (d) excitation by collisions; and (e) spontaneous decay, which yields the fluorescence emission (Crovisier and Encrenaz, 1983; Crovisier et al., 1982). For abundant species, large optical depths may be encountered in both rotational and rotational vibrational transitions, so that line transfer must be included. The energy distribution within the vibrational bands (for ground-based high-resolution infrared observations) or the pure rotational line emission (for microwave of submillimeter observations) then one must take into account the mechanisms (a), (b), (c), (d), and (e) for the whole set of fluorescent lines. The energy distribution within each band may be different, depending upon the nature and the lifetime of the molecule and depending upon the diameter of the region considered.

Following are the possible emission processes for cometary molecules (Crovisier and Encrenaz, 1983):

(a) Resonant fluorescence by the radiation field, which includes direct solar radiation, solar radiation scattered on the nucleus and dust particles, thermal emission of the nucleus and dust particles.

(b) Thermal emission and emission due to collisional excitation. The energy distribution within each band may be different, depending upon the nature

and the lifetime of the molecule and depending upon the diameter of the region considered (Encrenaz et aL, 1983). The emission bands and lines of molecule and radical depend on the corresponding probabilities within a band system.

5. Spectral Studies of the Interstellar Medium

The matter which is dispersed in space between the stars is known as interstellar matter. So far, in the interstellar medium in the gas between the stars in our Galaxy, twenty-six different types of molecules have been detected. The number of such molecules is most dense where the dust is densest, and they have been discovered in the regions where the stars appear to be forming and in the outer atmosphere of cool stars. The spectral data from such molecules give an insight in to physical conditions in such regions. New techniques have been developed to gain information from the interstellar medium, and

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 89

space vehicles have provided a mean of overcoming the interference of absorption by the Earth's atmosphere. One useful step which has been developed recently has been to employ a lightweight Michelson interferometer which can be flown successfully aboard a space vehicle with a view to studying to emission spectra in the 10-100 micron region. Such ventures have potential in studying the composition of interstellar medium, and studied should yield information on its heating and cooling mechanisms and its dynamic processes. They may lead to evidence on the formation of photostars, the nature of the galactic nuclei, and eventually to extragalactic objects such as quasars.

Many distant hot stars show in their spectra certain absorption lines with a shift differing from that shown by the remaining lines of the spectrum. This is illustrated clearly by certain spectroscopic double stars which give CaII lines with a constant velocity shift. In the spectrum of b Orionis it was demonstrated that the H and K lines of Call give a constant velocity shift, while the shift associated with the remaining lines went through a cycle of variation every few days owing to the orbital motion of the system. By the use of high-dispersion spectrographs other stationary lines have been detected including those of Na~, KI, FeI, TiI~, and CaL

These stationary lines are really superimposed on the stellar spectrum and are due to the passage of light from the star through material dispersed in the space between the stars (interstellar material). Some of the identified interstellar absorption lines have been shown by Straughan and Walker (1976). Since all the atomic lines are resonance lines, the absorbing atomic species may be considered to be in the ground state.

Although the atomic lines were readily identified, certain other lines caused more difficulty. However, Swings and Rosenfeld and also McKellar showed that the lines were due to interstellar molecules in their lowest rotational states. For example, the lines at 4300.3, 3874.6, and 4232.6 A were shown, respectively, to belong to CH, CN, and CH +. The observed absorption lines for these radicals in interstellar space are listed by Singh and Chaturvedi (1987) and Straughan and Walker (1976).

The lack of rotational energy of the interstellar molecules can be appreciated by a consideration of the conditions in interstellar space. Owing to the low density prevailing in these region collisions are very infrequent, and excitation by radiation is either small or absent. Should any higher rotational energy levels become occupied, then depopu- lation can occur in the cases of CH, CN, and CH + by rotational energy emission.

Hydrogen is by far the most abundant element in the interstellar gas (as it is also in normal stars). However, its detection by optical methods is not possible except in emission from those clouds that are close enough to hot stars for the hydrogen to be ionized. The ionization is caused by ultraviolet radiation emitted by the hot stars, and such regions where this process occurs are called H II regions. The hydrogen atom emission results from the proton capturing an electron and becoming an excited hydrogen atom. These atoms then lose their excess energy by emitting radiation some of which lies in the visible region. At a certain distance from a star, depending on its size, temperature, and on the density of the gas, radiation is no longer available to ionize the hydrogen, and from these regions (called HI regions) no optically detectable hydrogen atom emission can be observed. H H regions are well defined volumes forming

90 M. SINGH

a sphere around a hot star and occupy only about one-tenth of the space occupied by the H I regions.

Neutral hydrogen in the H I regions can, however, be detected by radio astronomy

from the 21.1 cm (1420 MHz) H I. This emission results from transitions between the

F = 1 and F = 0 levels in the ground state of the atom (n = 1), where F is the total

angular momentum quantum number of the whole atom. The two values o f F arise since S = 1 = j and I = �89 where F takes the values:

J + L J + I - 1 . . . . . ] J - I I .

The transition F = 1 to F = 0 corresponds to a reversal of the direction of the electron

spin relative to that of the proton spin and takes place only very rarely, the mean life of an atom in the F = 1 state being approximately 11 x 106 yr. Since, however,

interstellar gas clouds are so extensive, detectable amounts of the 21.1 cm radiation are emitted.

From the Doppler velocity shift of the 21.1 cm line and on the assumption of galactic

rotation it has been found possible to determine the spatial distribution of the neutral hydrogen clouds and to confirm the spiral structure of the Galaxy.

The Galaxy is composed of the Sun, with its attendant planets, and of the stars

nebulae, and gas clouds, and the stars. The boundaries of the Galaxy are not defined but it is estimated that the Galaxy is about 30 000 parsecs across and about 3000 parsecs

thick. The Sun is near the edge of the disc at a distance of about 10 000 parsecs from the centre. An important property of the Galaxy is that it is rotating - otherwise it could

not exist - and the hypothesis of rotation is supported by extensive observations of stellar velocities in the line-of-sight. On its outer parts it rotates not as a solid body but

in the same manner as the solar system, the outer planets rotating more slowly than the inner ones. The distribution of interstellar matter and hot, bright stars in the Galaxy is

now known to lie along the arms of a spiral near the central plane of the disc, the width of the arms being of the order of 500 parsecs.

Optical studies of the light from distant stars are restricted owing to attenuation by interstellar dust. Radio waves, on the other hand, can penetrate the dust and wider regions become available for investigation. Radioastronomy has the additional

advantage in extragalactic research of being less handicapped by red shifts brought

about by the rapid recession of the Universe. In fact, the use of the 21.1 cm line emitted from cool hydrogen clouds surrounding the intense extragalactic radio source in Cygnus showed that the recession velocity of these clouds deducted from the Doppler shift of this line was in agreement with the value of 16700 km s - 1 measured in the optical spectrum. This result confirms the view that the red shift observed in spectral lines of distant galaxies is a true velocity shift.

One big step forward in radio astronomy occurred in 1963 when the OH was detected by absorption at 18 cm. Most of the OH sources lie near the H n regions and young stars. The hydroxyl radical is observed in both absorption and emission. In fact, the OH gas is considered to be pumped into an excited condition and may be regarded as an interstellar hydroxyl maser. One of the features characteristic of OH emission is that the medium itself is very dense.

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 91

Another big step in interstellar space studies took place in 1967 when water was

identified in microwave emission at 1.35 cm. Ammonia was also detected, and studies

on these two types of molecule established that: (a) the clouds are quite dense compared with other known interstellar reaction; (b) the physical conditions of the molecules are quite different from those on Earth.

Recently, radio-emission has been observed for the J 1 ~ 0 transition of HlZC14N and H13C14N at 88.6 and 86.4 GHz, respectively. In addition formaldehyde has been

detected in interstellar space by strong absorption at 6 cm and this molecule is considered to afford one of the most promising means for explaining the denser regions of the Milky Way.

The work on identification of interstellar molecules appears to be developing quite rapidly and a study of the species present gives direct information on the location and velocity of interstellar clouds and leads to information on the physical conditions and the dynamics within the clouds. Studies of spectral linewidth may be related to the temperature and turbulence or large-scale motions within the cloud. Ammonia and carbon monoxide have proved most useful in temperature determinations.

Two important features which have emerged from the study of molecules in interstellar space are: (i) the formation of molecules from the number of atoms available in the interstellar clouds must be a most efficient process, (ii) there seems to be a tendency to favour the formation of organic molecules, and many simple organic molecules have been identified already, whereas a number of expected diatomic inorganic species have not been detected.

About 90 ~ of the interstellar medium consists of hydrogen atoms while atoms of helium make up the major constituent of the remaining 10 ~o. The proponents of the 'big bang' theory consider that nearly all the hydrogen and helium in the Universe interstellar dates back to that event, which is considered to have taken place about 10 billion years ago.

Striking studies have been made of interstellar molecules containing different iso- topes, and it has emerged that the interstellar isotope ratios appear to be the same as on the Earth. This suggests that interstellar chemistry has changed very little since the Earth was formed.

In recent years one of the most fascinating areas in interstellar studies has been the speculation as to how the molecules are formed. Various theories have been proposed and include: (i) formation of the molecules on the surface of dust particles, (ii) the molecules may be built up by atoms colliding and then sticking to one another in the gaseous phase, (iii) the molecules may come about as a result of evaporation or decom- position of dust particles when they encounter heating phenomena such as energetic particles or photons.

Thus, altogether the study of interstellar space by mainly molecular spectroscopy provides a challenging area for both physicists and chemists which is rapidly developing.

6. Absorption Spectrum of the Earth's Atmospheres

The opaqueness of the Earth's atmosphere at wavelengths less than 3000 A is partly due to ozone, which has a maximum concentration at a height of 20-30 km above the

92 M. S1NGH

Earth's surface. The ozone formation is dependent on the photodissociation of the oxygen molecule by the absorption of wavelengths of less than 2400 A. The 0 3 ab- sorption extends to approximately 2200 A but at about 2400 ~ the A3y,, + ~X3Zg - bands of oxygen are found. Below 1950 A intense 0 2 absorption (the Schumann-Runge system B3Y, u - +-- X 3 1 ~ ) produces opaqueness down to about 1300 A. Below this wave-

length other systems of 0 2 and N 2 cause complete absorption. At the infrared end of the electromagnetic spectrum intense rotation-vibration

transitions of water vapour occur in the regions 0.94-1.85 #, 2.5-3.0 #, and 5-7.5/~ and, in addition, rotational transitions of water vapour seriously deplete the solar spectrum at wavelengths greater than 16 # and completely between 24 # and ~ 1 mm, other localized absorptions also occurring in the visible and infrared regions are due to

CO2, N20, CH4, and 0 2. On moonless nights the Earth's atmosphere is still found to give a feeble illumination,

and efforts have been made to account for this emitted electromagnetic radiation. Owing to the weakness of the source of the type of instrument employed must have a high photographic speed, and consequently the dispersion and resolution are poor. The

spectrograph described on page 203 (Straughan and Walker, 1976) has been used for this type of investigation. The following are some of the difficulties encountered.

(i) It is not easy to determine by inspection of such spectrograms whether lines or

bands are being observed. (ii) The extent of any individual band may be small, possibly owing to the low

temperature of the source. (iii) Forbidden lines are observed which are not easily reproduced in the laboratory. From the work done so far, the greater part of the spectra of the Earth's atmosphere

is thought to be molecular in origin, though part of it is attributed to atoms. The atomic lines which have been definitely identified are the O I D 2 - 1S o transition

at 5577 A, the O I 3p _ 1.2 1D 2 transition at 6300 and 6364 A, and the sodium D-lines.

Oxygen above 100 km is in the atomic state. At night recombination of the O atoms occurs mainly between 90 and 120 km. Oxygen is more easily photodissociated than nitrogen, and this together with the absence of nitrogen lines in the spectrum is taken to indicate that nitrogen in the upper atmosphere is mainly in the molecular state.

From the twilight glow the violet N f bands may be detected in emission. The sodium D-lines and the red O I lines are also present, their intensity being greater than that

observed in the night glow. The spectra from auroral displays differ considerably from those of the twilight and

night glows. Forbidden lines of O I and N I and molecular bands of N2 and N2 + have been identified in auroral spectra. In addition the Balmer lines of hydrogen, the D-lines of sodium, and some He lines are occasionally found.

For the OH bands at 10440 ~ the height at which the radiation is emitted has been estimated at about 70 km. The rotational structure of these bands has been resolved and M einel (1950, 1951) determine the temperature of the OH from the rotational intensity

distribution and obtained a value of 260 + 5 K.

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 93

7. Long-Period Variables with H20 Emission

There are five long-period variable stars: U Ori, W Hya, S CrB, U Her, and R Aql. They are all relatively close, bright, late M and, except for W Hya, Mira Ceti-type variables. Compared to other H20 sources, they are not very intense: the H20 source in W49 is more than one million times brighter, intrinsically, than the brightest source in this group, R Aq 1. Of particular interest is the correlation between the H20 emission velocity and the velocity of the star as indicated by the absorption fines. In all cases, the relationship Vop t a b s > VH20 > Vopt em is satisfied (Schartz and Barrett, 1970). Since all of the stars are later than M5e, one would expect to find infrared H20 absorption bands in their spectra during part or all of a period and, in fact, they all do show this (Frogel, 1970).

These stars are also anomalous OH sources but the characteristics of the OH deviate from those usually found in late-type star, type liB, OH sources (Wilson et al., 1970). The strongest emission from these stars occurs at the main lines, rather than at 1612 MHz and the stars tend to show one rather than two emission velocities (Wilson, 1970c). Wilson suggests that these stars represent a distinct class of OH sources which he classifies as type Ib - main line OH sources associated with stars which show H20 emission (Wilson, 1970c). These stars also appear to have systematically smaller I -K colour indices than type IIb sources (Wilson, 1970b).

Like all other H20 sources, the emission from these stars is time variable. The velocity of the H20 line from long period variables indicates that the emitting region is closely associated with the star although the lack of velocity variations possibly argues against it being in the stellar atmosphere. Emission from a circumstellar shell, as has been proposed for OH, is possible although there is evidence that such a shell is less dense in H20 sources than in stars with 1612 OH emission (Wilson, 1970a). The H20 emission region of such a shell would have to be expanding at about 0-15 km s- 1 and possess a projected H20 density of at least 1016 cm -2 to give the observed velocities and support the masering mechanism. Frogel (1970) has found that the H20 infrared absorption bands from stars with H20 emission not only tend to be intense but also undergo larger variations in intensity as a function of phase. This observation probably supports the idea of stellar mass loss to a circumstellar shell which is indicated by the H20 velocities. It should, however, be pointed out that an essential contradiction exists between the infrared and microwave HzO observations. The H20 emission and absorption fines vary out of phase which virtually eliminates the possibility that the H20 maser is pumped by infrared radiation in a simple minded way. As the time variations indicate, the actual inversion mechanism must depend upon phase but, perhaps, the spectral energy distribution is more important than the actual intensity or, perhaps, collisional processes are involved. If the H20 emission originates from a circumstellar shell, the anticorrelation of H20 emission and absorption variations is less important since the absorption is most likely not a circumstellar phenomenon.

94 M. SlNGH

8. Peculiar Late-Type Stars with H20 Emission

There are H20 emissions from three late-type stars VY CMa, RX Boo, and NML Cyg. RX Boo is in this group because of tis short period and the small amplitude of its light variations. It is interesting to note that RX Boo violates the velocity relationship. The other two sources, VY CMa and NML Cyg, are probably similar objects - highly reddened late M stars. There is, however, a great disparity in the strength of the H20 emission from these two sources, VY CMa being over one thousand times more intense.

In addition to being a well studied optical and infrared object and one of the brighter H20 sources, VY CMa has the distinction of being the only source in which a confirmed velocity shift in a line has been observed. Sullivan (1970) has suggested that, in addition to shouring a slow decrease in intensity, the 36 km s - 1 feature is shifting upward in velocity, his observations confirm this shift. During the 14 month period shown, the 36 km s - 1 feature has moved upward about 0.3 km s - 1, retaining its line shape but

decreasing in intensity. VY CMaq is also among a group of bright sources which have been studied with a

very long baseline interferometer. The present upper limit of the angular size ofVY CMa is 0.003 arc sec which implies a linear size of about 1 AU, if the interferometer results represent the true angular size of the emitting region, and an H20 brightness temperature greater than 1014 K (Burke et al., 1970). The position of the H20 source has been found to be coincident with that of the star to within the accuracy of the experiment (Johnston

et al., 1970). In future work, he hopes to be able to refine the position to better than a sec of arc

and to measure the relative angular separation of the three strongest velocity

components.

9. Brief Comments on the Lack of H 2 and H20 Lines

Astronomers have not found any convincing Doppler-displaced lines of stellar steam in the 2 # spectrum of ~ Ori (Schwartz, 1970; Wilson, 1970a). He looks for both the possible low-excitation stellar lines on the wings of the telluric H20 features near 1.9/~ and also for the high-excitation lines selected from the list of H20 flame emission transition of Benedict et al. (1954) near 2 #. His negative result is in agreement with the recent aircraft spectra of ~ Ori discussed by Johnson et al. (1968) and reviewed in terms of stellar O/C by Wing and Spinrad (1970).

The fundamental band of the rotation-vibration quadrupole spectrum of H 2 also lies in his region of interest: he did not detect any of the S(0), S(1), or S(2) H 2 lines in the

Ori spectrum (Spinrad and Wing, 1969; for the implications). However, F and M Querci and Spinrad have recently probably found two or three weak H 2 lines in the

connected spectra ofo Cet at its 1966 light minimum. At this phase Mira is much cooler

than ~ Ori.

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 95

I0. Spherical Studies of the Nebulae and Forbidden Transition

A nebula is defined as a cloud of gas or of fine solid particles which normally envelops

a very hot star. A typical nebula embodies a star of type-O (temperature ~ 50 000 K) which is surrounded by a shell of gas: this has an emission spectrum containing the Balmer lines of H and prominent lines He~ and HeII. In the Orion nebula (NGC 1976), the hydrogen and helium lines are quite apparent. Among weaker lines of the nebula are those attributed to CII, CxII, and NeIII and sometimes also to OII and NIII. In the near-infrared region some hydrogen lines belonging to the Paschen series have been observed, and in one particular nebula (NGC 6572) a line from the neutral oxygen atom has been detected. Lines from O III and N III have also been observed in the spectra of nebulae and are often quite intense.

The unfamiliar conditions of extremely low density in nebulae (for example, the density of the Orion nebula has been estimated at ~ 300 atoms ml - 1 of that of air at

normal temperature and pressure) give rise to spectra which are not normally encountered in the laboratory.

In addition to the already indicated lines, other lines are observed which were not readily identified initially. The wavelengths of these lines have been measured quite accurately and the values obtained did not correspond to any known lines. These unknown lines were at first attributed to an unknown element, nebulium. Since, however, not suitable gaps were available in the periodic table it appeared more probable that nebulium was a known element under unfamiliar conditions.

The problem was solved by Bowen (1928, 1936) who showed that the frequencies were due to O, N, Ne, and S in various stages of ionization. Some of these lines have now been produced in the laboratory, although the majority have not yet been obtained, since their emission is forbidden by the selection rules for electric dipole radiation. They are consequently called forbidden lines.

The selection rules for atoms for electric dipole radiations are: (i) A J = 0 , + l , J = 0 - ~ J = 0 .

(ii) A L = 0 , _+l, a n d A l = _+1. (iii) AS = 0. (iv) Even terms combine only with odd terms and vice versa.

However, for magnetic dipole and electric quadrupole radiation the selection rules are in some cases quite different, and although the normal conditions on Earth are generally unsuitable for such transitions, in nebulae the reverse is true. Table I summarizes the selection rules for electric and magnetic dipole and for electric quadrupole radiation.

The Einstein transition probability of spontaneous emission Anm is given by

Anm - - 64~4v3m ]R~,,,]2 ; 3h

An,, gives the fraction per second of atoms in the initial state n undergoing transitions to the state m where n is the excited state. The mean life of state n is

J = I/A,,,,,.

96 M. SINGH

TABLE I Comparison of some of the selection rules for atoms for electric dipole, electric quadrupole,

and magnetic dipole radiation

Electric dipole Electric quadrupole Magnetic dipole radiation radiation radiation

AJ 0 + l O+l+2(J'+J">_2) 0+1 Terms + ~ - + ~ + + ~-~ +

AL 0_+1 0 + 1 + 2 0_+1 AS 0 0 0

For a permitted electric dipole transition J is approximately 10 -4 s, that is the average

lifetime of the state n is 10 - 3 s. If, however, no electric dipole transition is allowed from

state n to state m, A,m will be very small, and J will consequently be large. It has been caIculated that for transitions which take place by virtue of the electric quadrupole J

is of the order of 1 s, while for a magnetic dipole transition J is approximately 10 -3 s. Under these conditions n is termed to a metastable state.

Generally, if a transition from n to rn cannot take place by the electric dipole mechanism, then the other possibilities are by electric quadrupole or magnetic dipole transitions. However, on Earth the probability of most gaseous atoms radiating by such

means is very unlikely, because even in the highest vacuum attainable the atoms would

lose their excess energy by collision with other atoms, before they had the opportunity to emit. At extremely low densities prevailing in the gaseous nebulae atomic collisions are very infrequent, occurring for a given atom at an average interval of several hours

or even days. Under these conditions emission may take place by by magnetic dipole and electric quadrupole transitions giving rise to the so-called forbidden lines. The

concept of forbidden lines is based on the definition that lines which obey the selection

rules for electric dipole radiation are the permitted ones whereas all the remaining lines are termed forbidden. This definition automatically classes any transition involving

magnetic dipoles or electric quadrupoles as forbidden transitions, even though they obey the selection rules for that type of interaction. If in any portion of the nebula the predominant ion of an element X is X n +, then usually only forbidden transition o fX n +

appear. However, recombination o fX n + with electrons produces permitted transitions of X (n 1 ) N + In certain cases, for example, H, HeII, OIII, and N m , permitted

transitions do arise. One particular well-known forbidden line - the green auroral line at a wavelength of

5577.35 ,~, which is emitted after dark in the high terrestrial atmosphere and which is of great intensity during polar aurorae - was reproduced in the laboratory by McLennan and Shrum in 1924. They diluted oxygen with an inert monatomic gas in the discharge tube and concluded from the spectrum that the green line was due to the forbidden transition 1 0 2 -- 1 S 0 of the neutral oxygen atom; the transition was considered to take place through the electric quadrupole moment.

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 97

Forbidden lines belonging to Fe II, Fe hi, FevI, and FevII have also been detected. The solar corona has forbidden lines of FexI, Fexm, Fexiv, CaxH, Caxv, Nixm, Nixv, NixvI, and Mgx. Elements in gaseous nebulae detected by their permitted lines are relatively few in number and include H, He, C, N, O, and possibly Fe, Si, and Mg. The heavier atoms detected by their forbidden lines are, for example, F, Ne, S, C1, K, Ca, Fe, and Ni.

11. Spectral Studies of the Planetary Atmospheres

A number of criteria are available which enable the astronomer to decide whether a given planet is likely to have an atmosphere. The criteria depend on the observations of (i) clouds, seasonal polar caps, twilight arcs, and atmospheric refractions, (ii)polari- zation and reflection phenomena, (iii) whether sufficient gravitational attraction exists on the planet for it to retain an atmosphere. This can be computed from the mass, radius, and temperature of the planet. These criteria are useful since they enable the astronomer to fix his attention on those planets where a definite possibility of an atmosphere exists.

The presence of absorption of emission bands in the spectrum of a planet is especially valuable in a number of ways. It gives definite proof of the existence of particular species (i.e., an atmosphere of some kind), identifies the substance(s), and enables estimates of the relative abundance of the various species to be made. On the other hand, spectro- scopic observations may be limited by (i) the transmission of the Earth's atmosphere, (ii) sources of low spectral intensity. This leads to low dispersion spectra, (iii) H2, N2, and the inert gases are not revealed and C02, CO, and N20 only if the source is bright enough to make infrared observations possible.

Gases which are common to the atmosphere of both the planet and the Earth can sometimes be detected if the source is sufficiently intense for high dispersion studies. This is achieved by virtue of the Doppler shift which results in a lack of coincidence of the two spectra.

The high percentage of water vapour and oxygen in the Earth's atmosphere makes the determination of these gases in extra-terrestrial atmospheres difficult. However, steps can be taken to avoid the terrestrial water vapour by, for example, working at high altitudes or on days of low humidity.

Table I! summarizes the major constituents in the atmospheres of some of the planets in the solar system while Table III gives a fuller account for Mars, Venus, and the Earth.

Big developments in astrochemistry have been in the use of techniques which over- come the absorption of radiation by the Earth's atmosphere. One way in which this may be achieved is to place the spectrograph in a suitable housing which is mounted on a rocket. The spectrograph could be, for example, an ultraviolet one covering the range 500-2500 A. After passing through the Earth's atmosphere and taking the necessary spectra the rocket is recovered.

Barth et al. (1971) used a spectrometer in Mariner 6 and 7 and obtained the spectrum of the Mars upper atmosphere in the 1100-4300 A region, and they obtained as well the emission spectra of CO, O, and CO2 + . Their spectra are given by Barth et al. (1971)

98 M. SINGH

TABLE II

The main constituents in the atmospheres in the solar system

Venus CO2, H20 Mars CO2, H20, 02 Jupiter CH 4, NH 3 Uranus CH4, Hz Neptune CH4, H 2 Saturn CH 4, NH3 Titan (largest satellite of Saturn) CH 4

where they may be observed, below 2600 A the Cameron bands of CO, the Fox-Duffenback-Barker bands of CO~- above 3000 A, the COz + ultraviolet doublet band system at 2890 ,~, and the 2972 .~ line of atomic oxygen. The CO~- has been produced through photoionization as a result of solar ultraviolet radiation. In the ionospheric layers of the Earth molecular ions have a prominent role, e.g., N~, O ] , NO +, H20 +, and NO~-, and as Herzberg (1971) has indicated, the upper atmospheres of other planets would be expected to contain layers in which molecular ions are an important feature.

The Soviet entry probes, Venera 4, 5, and 6, have been used to study the atmosphere of Venus. In this case a direct sampling of the planetary atmosphere showed that the mole fraction of CO2 on Venus is in the range 93-97 ~o in the total atmospheric pressure range of 1.6-2.0 atm.

TABLE III

Abundance of gaseous matter in kg cm 2

Volatile Earth atmosphere Venus atmosphere Mars atmosphere

C O 2 0.5 X 10 -3 95 + 25 (1.4 + 0.2) X 10 .2 H20 (1-10) x 10 .3 10 -z (0.5 -2 .5 ) • 10 .6 02 0.23 < 6 x 10 -3 <2.5 x 10 -5 N2 0.75 <5 <0.8 x 10 -3 Ar 1.3 x 10 -2 <5 <0.3 x 10 -2 CO (1-10) x 10 .7 3 x 10 .3 0.7 x 10 .5

A considerable amount of data is now available on the gaseous matter in the atmos- pheres of Venus and Mars, and a comparison is made with that of the Earth in Table III.

It is apparent from Table III that CO 2 is an abundant compound in the atmospheres of all three planets. At present, though it is not possible to make a comparison of the total amounts of CO2 at or near the surfaces of these planets. It is possible, for example, that the Mars polar caps may contain large amounts of solid CO2 while the surface of Venus may well have large amounts of carbonate deposits. If Venus has all its water in the atmosphere, as some thinks feasible, then the water on Venus is no more than 1/1000th of that on Earth.

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 99

Particularly interesting studies have been made of water on Mars. It was detected

spectroscopically in 1964 and later work indicated that the relative humidity might be

as high as 5 0 ~ . One worker considered that the presence of liquid water on the surface

of Mars was extremely unlikely. Another worker obtained spectroscopic evidence that

the polar caps contain water. I t is considered possible that large amounts of water may be present as permafrost or as hydrated minerals on the surface. This example of the

detection of the presence of water on Mars serves to illustrate that after the initial detection of species on a planet, further work and speculation often follow. Murray (1973) has described Mars f rom Mariner 9.

It is apparent that a study of the a tmosphere of terrestrial planets leads us to a better

appreciation of the Ear th 's a tmosphere and its stages of evolution. Perhaps the most intriguing question is that o f whether the Earth is unique or just a member of the series

of terrestrial planets and in, say, the sequence between Venus and Mars. More informa- tion is required on the degassing process of the surface of each of these planets, and

how they differ. Furthermore, it is desirable to determine to what extent their a tmos- pheres have been modified after outgassing. Theories are formulated and evidence

produced for reactions between the constituent gases of the a tmosphere and the planet 's crust. It has been speculated that a large number of compounds may be present in the

a tmosphere of Venus, and some compounds which have not yet been detected but have

been proposed, are COS, HBr, H2S, and HgBr2; such proposals take into account the high surface temperature.

Thus it is evident that the inforrnation gained by spectroscopic means on a few species

in the a tmospheres of planets leads to various types of speculation.

References

Adams, W. S.: 1939, A Rep. Mt. Wilson and Palomer Obs. Adams, W. S.: 1941,Astrophys. Jr. 93, 11. Babcock, H. D.: 1945, Astrophys. J. 102, 154. Ballik, E. A. and Ramsay, D. A.:, 1963, Astrophys. J. 137, 84. Barth, C. A., Hord, C. W., Pearce, J. B., Kelly, K. K., Anderson, G. P., and Stewart, A. I.: 1971, J. Geophys.

Res. (in press). Basu, B.: 1986, Sci. Rep. (Indian) 23, 74. Benedict, W. S., Bass, A. M., and Plyler, E. K.: 1954, J. Res. Nat. Bur. Standards 52, 161. Bhatti, S. S.: 1985, Bull. Indian Ass. Phys. Teach. 2, 117. Biemont, E.: 1978, private communication. Bowen, I. S.: 1928, Astrophys. J. 67, 1. Bowen, I. S.: 1936, Rev. Mod. Phys. 8, 55. Branch, D.: 1969, Solar Phys. 10, 112. Burke, B. F., Papa, D. C., Papadopoulos, G. D., Schwartz, P. R., Knowles, S. H., Sullivan, W. I. III, Meeks,

M. L., and Moran, J. M.: 1970, Astrophys. J. 160, L63. Carroll, P. K., McCormack, P., and O'Connor, S.: 1976, Astrophys. J. 208, 903. Cooper, D. M. and Nieholls, R. W.: 1975, J. Quant. Spectr. Rad. Trans. 15, 139. Crawford, F. H. and Shurcliffe, W. A.: 1934, Phys. Rev. 45, 860. Crovisier, J., Combes, M., Enerenaz, T., and Le Bourlot, J.: 1982, 'Cometary Exploration I', Proc. Inter-

national Conf. on Cometary Exploration, November 15-19, Budapest, Hungary, p. 197. Crovisier, J. and Encrenaz, T.: 1983, Astron. Astrophys. 126, 170.

100 M. SING[-[

Crovisier, J.: 1986, Space-Borne Sub-Millimetre Astron. Mission, A Cornerstone of the ESA Long-Term Space Sei. Programme, Segovia, Spain, June 4-7, p. 57.

Davis, S.: 1980, private communication. Dunham, T. J.: 1937, Publ. Astron. Soc. Pacific 49, 26. Encrenaz, T., Crovisler, J., Combes, M., and Le Boulot, J.: 1983, in C.-I. Lagerkvist and H. Rickman (eds.),

Asteroids Comets Meteors, Proc. of a meeting held at the Astron. Obs. of the Uppsala Univ., June 20r22, p. 291.

Engvold, O., WOhl, H., and Branlt, J. W.: 1980, Astron. Astrophys. Suppl. Seri. 42, 209. Feldman, P. D.: 1977, private communication. Frogel, J. A.: 1970, Astrophys. J. (in preparation). Grevesse, N. and Sauval, A. J.: 1971, Astron. Astrophys. 14, 477. Hall, D. N. B.: 1974, An Atlas of Infrared Spectra of the Solar Photosphere and of Sunspot Umbrae in the Spectra

Intervals 4040 cm - l_5095 cm- z, 5550 cm - 1 _ 6700 cm - l, and 7400 cm - 1 _ 8790 cm - t Kilt Peak National Obs., U.S.A.

Hervey, J. W.: 1972, Solar Phys, 24, 354. Henneker, W. H. and Popkie, H. E.: 1971, J. Chem. Phys. 54, 1763. Herzberg, G.: 1971, Faraday Lecture, p. 201. Hunaerts, J.: 1947, Ann. Astrophys. 10, 237. Hyland, A. R., Beeklin, E. E., Frogel, J. A., and Neugebauer, G.: 1970 (in preparation). Johnson, H. L., Coleman, I., Mitchell, R. I., and Steinmetz, D.: 1968, Comm. Lunar Planetary Laboratory

7, 83. Johnston, K. J., Knowles, S. H., Sullivan, W. T. III, Moran, J. M., Burke, B. F., Lo, K. Y., Papa, D. C.,

Papadopoulos, G. D., Schwartz, P. R., Knight, C. A., Shapiro, I. I., and Welch, W. J.: 1970 (in prepar- ation).

Jordan, C., Br(ickner, G. E., Bartoe, J.-D. F., Sandlin, G. D., and Van Hoosier, M. E.: 1977, Nature 270, 326.

Kamoun, P. G., Pettengfll, (3. H., and Shapiro, I. I.: 1982, in L. Wilkening (ed.), Comets, Univ. of Arizona Press, Tucson, p. 288.

Klynning, L. and Lindgren, B.: 1973, Inst. of Physi., Univ. of Stockholm, Report No. 73, p. 20. Kozhevnikov, N. I. and Polonskij, V. V.: 1969, Soviet Astron. 13, 267. Krishnaswamy, K. S. and O'Dell, O. R.: 1977, Astrophys. J. 216, 158. Laborde, G.: 1961, Ann. Astrophys. 24, 89. Lambert, D. L. and Mallia, E. A.: 1973, Solar Phys. 31, 123. Lambert, D. L. and Mallia, E. A.: 1974, Bull. Astron. Inst. Czech. 25, 216. Mallia, E. A.: 1974, Molecular Spectroscopy, Specialist Publication, Chemical Society, London, Vol. 2, p. 554. McCormack, P. and O'Connor, S.: 1976, Astron. Astrophys. Suppl. Ser. 26, 373. Meinel, A. B.: 1950, Astrophys. J. 111, 555. Meinel, A. B.: 1951, Rept. Progr. Phys. 14, 121. Mendis, D. A., Houpis, H. L F., and Marconi, M. L.: 1985, Fund. Cosmic Phys. 10, 1. Minnaert, M. G. J.: 1949, Bull. Astron. Inst. Neth. 11, 51. Moore, C. E., Minnaert, M. G., and Houtgast, J.: I966, The Solar Spectrum 2935 ft to 8770,4, NBS

Monograph. Mount, G. H. and Linsky, J. L.: 1975, Solar Phys. 41, 17. Murray, B.: 1973, Mars from Mariner 9, Scientific American. Pecker, J. C.: 1957, Mere. Soc. Roy. Sci. Lidge 18, 332. Richardson, R. S.: 1931, Astrophys. J. 73, 216. Roach, F. E.: 1939, Astrophys. J. 89, 99. Russell, H. N.: 1934, Astrophys. J. 79, 317. Schadee, A.: 1964, Bull. Astron. Inst. Neth. 17, 311. Schadee, A.: 1968, Astrophys. J. 151, 239. Schadee, A.: 1970, Solar Phys. 15, 345. Schwartz, P. R. and Barrett, A. H.: 1970, in G. W. Lockwood and H. M. Dyck (eds.), Kitt Peak National

Obs. Contr., No. 554, p. 95. Singh, M. and Chaturvedi, J. P.: 1987, Astrophys. Space Sci. 135, 1. Sinha, K.: 1978, 'Molecules in Solar Atmosphere', Ph.D. Thesis, University of Gorakhpur, U.P, India, p. 7. Sinha, K. and Pande, M. C.: 1974, Bull. Astron. Inst. Czech. 25, 84.

SPECTRAL STUDIES OF MOLECULES OF ASTROPHYSICAL INTEREST 101

Smith, W. H.: 1969, Astrophys. J. 156, 791. Sotirovski, P.: 1971, Astron. Astrophys. 14, 319. Souza, S. P. and Lutz, B. L.: 1977, Astrophys. 3". 216, L49. Spinrad, H. and Wing, R. F.: 1969, Ann. Rev. Astron. Astrophys. 7, 249. Straughan, B. P. and Walker, S.: 1976, Spectroscopy 3, 208. Suchard, S. M.: 1975, Spectroscopic Data L Part A, IFI/Plenums Data Company, New York. Sullivan, W. T. III: 1970, presented at the 132 Meeting of the AAS, Boulder. Tsuji, T.: 1977, Publ. Astron. Soc. Japann 29, 497. Weinreb, S., Barrett, A. H., Meeks, M. L., and Henry, J. C.: 1963, Nature 200, 829. Wilkening, L. L.: 1982, Comets, Univ. of Arizona Press, Tucson. Wilson, W. J.: 1970a, Ph.D. Thesis, Elect. Engr. Dept., Massachusetts Inst. of Technology. Wilson, W. J.: 1970b, Ph.D. Thesis, MIT Electrical Engineering Department. Wilson, W. J.: 1970c, private communication. Wilson, W. J., Barrett, A. H., and Moran, J. M.: 1970, Astrophys. 9". 160, 545. Wing, R. F. and Spinrad, H.: 1970, Astrophys. J. 159, 973. Wing, R. F., Cohen, J., and Brault, J. W.: 1977, Astrophys. 9". 216, 659. W6hl, H., Engvold, O., and Brault, J. W.: 1983, Inst. Theor. Astrophys., Blindern, Oslo, Report No. 56. Wyller, A. A.: 1962, Observatory 82, 73.