.

t-

-

-

I

.:

-

i I Hi

SPECTRUM ANALYSISBY

JOHN LANDAUER,Member ofthe Imperial

LL.D.,Naturalists.

German Academy of

AUTHORIZED ENGLISH EDITIONBY

J.

BISHOP TINGLE,Chicago,III.

PH.D.,

F.C.S.,

Instructor of Chemistry in the Lewis Institute,

FIRST EDITION.FIRST THOUSAND.

NEW YOR K.JOHN WILEY & SONS. LONDON CHAPMAN & HALL, LIMITED.:

1898.

Copyright,BYJ.

1898,

BISHOP TINGLE.

ROBERT DROMMOND. ELECTROTYPRR AND PRINTER, NEW YORK.

AUTHOR'S PREFACE.THIS work originated as a reprint of an article on Spec" trum Analysis recently contributed to Fenling-Hell's Neues " it was published as a sepaHandworterbuch der Chemie rate book at the request of a number of competent authorities, but not without some hesitation on the part of the author,;

because in treating a subject in an encyclopedic article regard

must be paid to the whole plan and scope of the work, whilst in a separate book the author is quite independent. The favorable reception accorded to the book whenpublished givesrise

to the

hope that shortcomings

arising

some extent counterbalanced by a origin fulness of contents brought together in small space, by the strictly historical treatment of the subject adopted throughfromits

are to

out the book, by tolerably full bibliographical references, and by the care which has been bestowed on the numerical tablesIn order to secure a degree of uniserving for reference. formity hitherto wanting, the older measurements have beenrecalculated so as to bring system of wave-lengths.

them

into accord with

Rowland's

THE AUTHOR.BRAUNSCHWEIG,1898.iii

ABBREVIATIONS.following abbreviations have been used in the bibliographical references:

THE

A. B. A.

=

Abhandlungen

Koniglichen Akademie der VVissenschaften zu Berlin.et

der

A.B.C.

c.

p.

A. R.N.r.

C.

= = = =

Annales de chimieChemical News.

de physique.

British Association Reports.

Comptes rend us hebdomadaires des Seancesde 1'Academie de Sciences, Paris.Nature.

N. N. A. S.P.

= U ==

Nova ActaPoggendorff's

Regiae

Societatis

Scientiarum

Upsaliensis.

A.

Annalen

der

Physik

und

Chemie.P.P.

M. M.

P. S.

= = = = = = =

Philosophical Magazine. Proceedings of the Manchester Literary andPhilosophical Society.

P.

R. S. R. S. E.T.

Proceedings of the Royal Society.Proceedingsburgh.of

P.

the

Royal Society,

Edin-

P.

Philosophical Transactions.

P. T. E.

T. R. S. E.

Philosophical Transactions, Edinburgh. Transactions of the Royal Society, Edin-

W. A.

=

burgh.

Wiedemann'sChemie.

Annalen

der

Physik

und

TRANSLATOR'S PREFACE.

curriculum

claim of spectrum analysis to a place in a chemical is steadily obtaining increased recognition, and its is generally admitted both for students importance preparingfor teaching,

THE

calits

work.pursuit

who wish to engage in technologiThe subject may rightly demand a wider field since furnishes so many opportunities for an excellentandfor those

training in accuracy of observation and manipulative skill that it might, with great advantage, find place in a general sciencecourse.

The expense

is

by no means

prohibitive,

and

is

almost entirely confined to the first cost of the instruments, which, with proper care, last for years, and even with the cheaper and smaller ones, such as Browning's "Students'Spectroscope," which costs about $30, much interesting work can be done and valuable discipline obtained.

Of the works on spectrum analysis hitherto published in English, none are suitable as text-books, either on account of their size and consequent cost, or from the manner in whichthe subjectIt isis

presented.this little

hoped that

book may,

in

some degree, sup-

ply this lack.

There has been no attempt to treat the subject exhaustively, but rather to indicate the more salient points of theory, etc., leaving it to the teacher to complete and expand themat his

own

discretion.it

No

doubt

would be well

if

all

students were compelledvii

to take a course in general physics before attacking chemistry,

Vlll

TRANSLATOR'S PREFACE.

but at present this is a state of things not realized in practice; to those who have followed such a course the physical sectionbut it may serve to call of the book should be superfluous the attention of the others to matters on which they should;

obtain more instruction.

The

tables of wave-lengths will,

it

is hoped, be useful in the practical work which would probThe posiably constitute the greater portion of the course. tion of the more prominent lines and bands can, by their

help, be at once ascertained, identification facilitated.

and

their actual occurrence

and

LEWIS INSTITUTE,CHICAGO,ILL., Dec. 1897.

CONTENTS.

CHAPTERINTRODUCTORYHISTORICALi.

........2.

I.

PAGEr

Introductory,

Historical,8.

Bibliography of

Works on

Spectrum Analysis,

CHAPTERPHYSICAL PROPERTIES OF LIGHTWave-length,persion,16. 14.

......Prisms,13.

II.

nDis-

ii.

Reflection; Refraction, 12.

Abnormal Dispersion, 15. Pure Spectra; Gratings, Diffraction, 17. Comparison of Diffraction and Refraction19.

Spectra,

CHAPTERSPECTROSCOPES..-

III.... .

.

.

k

.20

Spectroscopes with Angular Vision, 20. scopes, 28. Grating Spectroscopes, 31.

Direct-vision Spectro-

CHAPTER

IV..

SPECTROSCOPIC INSTRUMENTS FOR SPECIAL PURPOSES

.

.

.35

Kriiss' Universal Spectroscope, 38. Spectrometer, 35. SpecSolar and trophotometer, 39. Sorby's Microspectroscope, 40. Stellar Spectroscopes, 41. Stellar Spectrometers; Spectrographs, Rowland's Concave Grating Spectrograph, 45. 43.

CHAPTERSPECTROSCOPIC ADJUNCTS

V.51

Flame Spectra, 51. Electric Arc, 54. Electric Spark, 55. Geissler or Pliicker's (Vacuum) Tubes, 58. Observation of the InvisibleRegions of the Spectrum; Ultra-violet,vation of Absorption Spectra, 63.Scales, 65.60.

Infra-red, 61.

Obser-

Measuring Appliances andix

Drawings

of Spectra. 68.

X

CONTENTS.

CHAPTEREMISSION-SPECTRALine Spectra,71.

VI.PAGE

69

Lockyer's Long Current, 75. Absorption Spectra; Kirchhoff's Law, 76. Influence of the Temperature and the Physical State of Substances, 77. Influence of the Optical Density; Influence of the Solvent, 78. Fluorescence and Absorption, 79. Relationship between the Linesof an Element. 80.

Influence of Temperature and Pressure, 72. Influence of Magnetic and Short Lines, 73.

Relationship between the Spectra of Differ-

ent Elements, 87.

CHAPTERSPECTRA OF THE ELEMENTSUnit of Measurement,Reactions, 95.96. 92.

VII.

.92Scale of Colors(in;

Spectra of the Elements

Delicacy of Spectrum Alphabetical Order),

CHAPTERABSORPTION SPECTRA

VIII.174

Absorption by Gases and Liquids, 175. Different Salts of the same Colored Base or Acid, 177. Relationship between Molecular Structure and Absorption Spectrum, 178. Absorption in theviolet, 182.

Visible Portion of the Spectrum, 179. Absorption in the Absorption in the Infra-red, 184.

Ultra-

CHAPTERTHE SOLAR SPECTRUMThe FraunhoferSun, 190.Lines, 186.

IX.186

The Chemical Composition

of the

Rowland's Table of Wave-lengths of the Fraunhofer

Telluric Lines of the Solar Spectrum, 201. Limits of Lines, 191. the Investigation, 202. Physical Condition of the Sun, 203. Solar

Nucleus

;

Photosphere

;

Sun-Spots. 204.

Solar Faculae205.

;

Reversing207.

Layer

;

Chromosphere and Prominences,

Corona,

CHAPTEROTHER CELESTIAL BODIESFixed Stars, 208.ites

X.208

Planets and Moon, 209.;

Comets;

Meteor-

and Shooting Stars

Light; Lightning, 211.

Aurora Borealis; Zodiacal Nebulae, 210. Displacement of the Lines, 212.

SPECTRUM ANALYSIS.CHAPTERINTRODUCTORY.SPECTRUM ANALYSISmeansofis

I.

HISTORICAL.

which

it is

a chemico-analytical method by possible to determine the constituents of

a substance, by observing the refraction (dispersion), or the Its further development offers a diffraction of light-rays. means of investigating the molecular structure of matter.

produced when light-rays are refracted is termed a spectrum. White-hot solid bodies emit rays of all refrangibility, and give a continuous spectrum; glowing gases or vapors emit rays of definite refrangibility, and therefore

The image which

is

yield a discontinuous spectrum consisting of bright lines which are characteristic of each substance, and which consequentlyit occurs alone, or together the rays from a white-hot solid pass through a colored medium some of them are retained giving an absorption spectrum, which varies with the chemical

serve for

its

identification

whether

with other bodies.

When

composition of the medium.far

Spectra-reactions are characterized by an extreme delicacy exceeding that of chemical tests, and therefore theirled to the discovery of a

employment has

number

of

new

elements which occur only in small quantity. Since the distance of the source of light has little effect on a spectrum, the method can be employed for the investigation of celestial

2

SPECTRUM ANALYSIS.it

bodies:

has extended our knowledge of their nature to an1

extent which was previously entirely unattainable.hoff

Spectrum Analysis was founded by Kirchand Bunsen in 1859, an d subsequently developed. Other observers had previously noticed spectrum lines, and had suggested the application of spectroscopic observations to chemical analysis, but their efforts were fruitless, as at that time it was not certain whether the bright lines of a glowing gas were solely dependent on its chemical composiThe sodium reaction was particularly misleading as it tion. was often observed when the presence of this metal was not suspected, and was therefore variously ascribed to sodium, to The yellow sodium flame was first sulphur, or to water. Thomas Melville in 1752, but he was unable to noticed byHistorical.2

determine

1822 investigated the spectra of many colored flames, particularly those given by strontium, copper, and -boric acid, and in 1827 showed that by this means the substances giving the colors could beits

origin.

John Herschel

3

in

recognized even when present only in extremely small quanFox Talbot 4 in 1826 expressed himself still more tity. definitely, stating that if his theory that certain bodies gavecharacteristic lines should prove to be correct, then a glance at the prismatic spectrum of a flame would suffice to identify

substances which would otherwise require a tedious chemical In 1834 he correctly described analysis for their detection.the lithium and strontium spectra, and again pointed out thatKopp, Entwickelung der Chemie in der neuren Zeit (Miinchen, 1873), Kirchhoff, Zur Geschichte der Spectralanalyse. P. A. 118, 102. Brewster, C. r. 62, 17. Kahlbaum, Aus der Vorgeschichte der

1

pp. 215, 642.94,

{Braunschweig,2 8

Rosenberger, Geschichte der Physik, 3 Spectralanalyse (Basel, 1888). P. M. [4] 25, 250. Stokes, N. 13, 188. Talbot, 1890). P. R. S. E. 7, 461.Edinb. Phys. andLit.

Essays, 2,

12.

T. R. S. E. (1823) 9.

P. A. (1829) 16.P.

On

the theory of light (Lon3.

don, 1828) 4 Brewster's Journ. of Sci.

5.

M.

(1833) [3] 3, 35; (1836), 9,

INTRODUC1 'OR Y. HIS TO RICA L

.

3

such optical methods permitted of the identification of these elements with a minimum quantity of substance, and with anexactitude equalling, if not excelling, that attained by any Doubt was, however, cast on this conclusion other process. by contradictory statements in the same communications, and

the method of analysis was rendered fundamentally dubious, because, in opposition to Herschel, Talbot maintained that the reactions could be produced by the simple presence of thesubstance in the flame,its

volatilization not being necessary.

1845 an investigation on tne spectra of the alkali metals; diagrams were given, but the results did not constitute any great advance, as he had

W.

A. Miller

1

published in

employed a luminous flame, and was therefore unable to determine what was characteristic of any particular metal. In 1856 Swan definitely proved that the yellow line which is almost always present is peculiar to sodium compounds, and that the frequency of its occurrence is due to the almost In his work on the universal distribution of sodium salts.8

prismatic spectra of the hydrocarbons Swan showed that the lines observed are constant in position; he thus made a valuable contribution towards the solution of the question as to whether the bright lines of a glowing gas are exclusively

dependent on its chemical composition. The definite and general answer to this problem was, however, not given by Swan, but by Kirchhoff and Bunsen. The spectra of the electric spark had been under observation simultaneously with those of flames; Wollastondetected a large number of bright

3

lines, but without offering He was also the first to describe clue to their origin. any the dark lines in the solar spectrum, and he improved the

apparatus employed by substituting a narrow slit for the circular opening which Newton had used to admit the light.1

2 3

P. M. [3] 27, 81. B. A. R. 1845T. R. S. E. 3, 376; (1857)21,353. P. T. 1802. p. 365.

4Fraunhofer1

SPECTRUM ANALYSIS.

was scarcely more successful than Wollaston so far as the origin of the bright lines was concerned: his fame rests on the discovery of the diffraction grating, the measurement of wave-lengths which its use permitted, and on the observation of the dark lines in the solar spectrum which bear He drew 350 of these, and finding that they his name. varied from those observed in stellar spectra, he concluded that they originate in the sun and stars, and are not due to 2 Wheatstone in 1835 found that the earth's atmosphere.with the use of different metallic electrodes the spectra vary, but they remain constant no matter whether the dischargetakes place in air, hydrogen, or in a vacuum; he therefore concluded that the metal is volatilized, but not burnt, by the

He published drawings of the spectra passage of the spark. of sodium, mercury, zinc, cadmium, bismuth, tin, and lead,and recommended the method

The

spectra

of

various

for analytical purposes. metals volatilized in air were3

in 1849; he studied, although less thoroughly, by Foucault also observed the dark /7-line, since known as the reversed

sodium line, but failed to draw the important conclusion from this which Kirchhoff subsequently made. Masson, who improved the method of working, using condensers charged by induction-currents, investigated the spark-spectra of iron, tin, antimony, bismuth, copper, lead, cadmium, and carbon; in all these cases he noticed that the lines due to moist air were present, although he was ignorant of their origin. This was indicated by Angstrom's important work published in He showed that the lines which occur in the space 1853. between the electrodes are due to air or to any other gas4 a1

Denkschriften der Miinchener Akad., 1814, 1815; Gilbert's Ann. 74,B. A. R. 1835.C. N. 3, 198.P.

3372

M.

[3] 7.

3

Institut. 1849, p. 44.

4b

A.

c. p.

(1851) [3] 31, 295.p. 335.

K. Vetenskaps Akad. Handl. (Stockholm, 1853),

P. A. (1855)

94, 141.

INTRODUCTORY. HISTORICAL.

5

which may be present, whilst those close to the electrodes are Angstrom also drew and described the given by the metals. of a large number of metals and non-metals, and spectra almost discovered the relationship between the emission and absorption of light, since he stated, in accordance with a suggestion of Euler, that at a common temperature bodies absorb the same vibrations which they are capable of producing.

In 1858 Pliicker commenced his investigations of the spectra produced by the passage of an electric current through highly He found that the elementary gases, or rarefied gases.1

the constituents liberated fromacterized

compound

gases,

are char-

by

bright lines.8

van der Willigen, who in electrodes moistened with a salt solution give the spectrum of the salt, and that it is therefore unnecessary to use theorder to obtain its spectrum. In the same year and Bunsen published their work " Chemische Kirchhoff Analyse durch Spectralbeobachtungen" their results were

Similar work was pursued by 1859 a l so showed that platinum

metal

itself in

3

;

obtained to some extent independently of previous investigators who, whilst frequently on the right path, had failed to reach the goal. They reduced spectral phenomena to achemical-analytical method, and definitely proved that the bright lines produced by a glowing gas are dependent only

This law still forms the basis chemical composition. but their second proposition has been of spectrum analysis,its

on

in

subsequently considerably modified; it states that the manner which the constituents of a substance are combined is with-

out influence on their spectra, and that these are also almost entirely unaffected by the temperature and pressure of the vapor. After Roscoe and Clifton had called attention to the difference between the spectrum of an element and those of4

its

compounds, A. Mitscherlich1

*

showed,

in 1863, that

every

*

P. A. 103, 88; 104, 113, 622; 105, 67; 107, 77, 415. 3 P. A. 106, 610; (1859) 107, 473. P A. 110, 167.

4

P.

M.

P. S.

1862.

5

P. A. (1863) 121,

3.

6

SPECTRUM ANALYSIS.

compound has its own peculiar spectrum, and that the exhibition of identical spectra by the various salts of an element In their first is caused by these undergoing dissociation.of the metals of the alkalis

communication Kirchhoff and Bunsen described the spectra and alkaline earths, and showed the great delicacy of the method, which permits of the recognition of substances

present in quantity far too small for detection by the ordinary processes; they also pointed out the great extension which it gives to our knowledge of thedistribution of the elements,

when

and indicated that

it

would

The correctprobably lead to the recognition of new ones. ness of this view has been proved by the discovery of caesium, rubidium, thallium, indium, gallium, and many metals of therare earths,all

by meansof

of

spectrum analysis.1

The developmentimpulse fromits

spectrum analysis received a special application to astronomy. Kirchhoff proved

between the emissive and absorptive powers

mathematically that for every ray of light the relationship of all bodies is

alike at uniform temperatures; this explained the origin of the Fraunhofer lines, and led to the investigation of the chemical

The discovery composition of the sun and its atmosphere. of this law of exchanges induced Kirchhoff to prepare moreexact drawings of the solar spectrum, and to accurately compare the positions of the Fraunhofer lines with those in thespectra of

many

terrestrial substances.scale, as did also

He employed2

for this

purpose an arbitrarythese observations.

Huggins, who extended

belongs the credit of subthe wave-length for the scale as a means of determinstituting ing the position of the lines, and his measurements, and atlas of the solar lines, remained for twenty years the foundationo

To Angstrom

3

of

all

spectroscopic investigations.27, 1859.

Angstrom's work was

1

Monatsber. Berl. Akad., Oct.P. T. (1864) 154.

2

3

Recherches sur

le

spectre normal du soleil avec atlas de 6 planches.

Upsala, 1868.

IN TR OD UCTOR Y. HIS TO RICA L.

f

confined to the visible portion of the spectrum; it was comresearches on the ultra-violet, and by pleted by Cornu's'

Langley's and Abney's on the infra-red. After Angstrom's death, Thalen showed that the metre he had employed was incorrect, and that consequently his wave-length determinaThis was confirmed by Muller and tions were too small.4

2

3

5

Kempfcarried

in

1886: their measurements of 300 solar lines were

out with great care, and became the basis of the Potsdam system. All these determinations were, however, 6 exceeded in accuracy by Rowland's Atlas of the solar spec-

trum, and his reproductions of normal lines, published in 1888 and 1893 respectively. His discovery of the concave grating " " method of determining the coincidence in 1891, and hisrelative position of lines, has greatly aided spectroscopic work, since it admits of the production of photographs without the

use of a lens, thus insuring a high degree of comparativeaccuracy. For a considerable time the measurement of the spectra of terrestrial substances did not keep pace with that of thesolar spectrum; KirchhofT's

and Huggins' determinations were

7 duly superseded by the more accurate ones of Thalen, but these were confined to the visible spectrum. Apart from

W. A.

Miller's9

8

incomplete work on the ultra-violetfirst

in 1862,

Lockyer1

in

1881 was the

to accurately investigate the22.

2 3

Partie ultraviolette (Paris, 1881), p. Spectre normal du soleil. P. M. [5] 21, 394; 22, 149; 26, 505. W. A. Beibl. 4, 375; 5, 507. C. r. 90, 182. P. T. 1880, p. 653.

4

Spectre du.9,

fer.

Acta R. Soc. Sclent. Upsala, (1884)

[3], p.

49.

W. A.

Beibl.5

520.

Publ. d. Astrophys. Obs. zu Potsdam (1886), 5. Photographic Map of the normal Solar Spectrum, Johns Hopkins Univ., Baltimore. Astronomy and Astrophysics (1893), 12, 321. P. M.6

(1894) [5] 36, 49.7

N. A.

8 9

S. U. (1868) [3] 6. P. T. (1862) 152, 861. P. T. (1873) 163, 253, 639; (1874) 164, 479, 805.

P. R. S. 25, 546; 27,

49, 279, 409; 28, 157.

8

SPECTRUM ANALYSIS.it,

subject, but he soon quitted

and1

its fuller

examination was

reserved for Hartley and Adeney, and Liveing and Dewar.' 3 Since 1888 Kayser and Runge have met with great success in their important task of measuring the emission-spectraof terrestrial substances

by Rowland's method.

They com-

order to determine the relationship of the various lines of an element, and also that of the lines ofin

menced the work

different elements.

Attempts had been made

in this direction

shortly after the discovery of spectrum analysis by Kirchhoff and Bunsen; it was at first believed that the relationship of

the lines was similar to the sound-waves of a vibrating string, which consist of a fundamental note and harmonic overtones. This view was shown by Schuster in 1880 to be incorrect, and in 1885 Balmer discovered a formula which accurately These invesreproduces the hydrogen lines in wave-lengths. with the observations of Liveing and tigations, together Dewar 6 on harmonic series of similar lines, are naturally connected with Kayser and Runge's work, which has led to the4

5

discovery of the methodical structure of a series of spectra. 7 Rydberg, working independently of Kayser and Runge, has

obtained similar results.

Investigations of this nature have tended to greatly widen the domain of spectrum analysis.

BIBLIOGRAPHY OF WORKS ON SPECTRUM ANALYSIS.

CAPRON.CAZIN.DlBBLTS.

Photographed Spectra.

1877.

La

spectroscopie.

Paris, 1878.Paris, 1895.

DEMARgAY.

Spectres e"lectriques.Spectraal-Analyse.

De

Rotterdam, 1869.

P. T. 1884, p. 63. P. T. 174. 187.

A. B. A. 1888-1894. B. A. R. 1880.

P. R. S. 34, 119, 123. W. A. Beibl. 6, 934; 7, 849. Runge B. A. R. 1888, 576.

W. A. (1885)25.P. T. (1883) 174, 208.

C.1890).

r.

(1890) 110, 394.

K. Vetenskaps Akad. Handl., 23 (Stockholm,

INTRODUCTORY. HISTORICAL.

9

DIETERICI. Spectralanalyse in Ladenburg's Handworterbuch der Chemie. Breslau, 1892.

DRAPER. Catalogue of Stellar Spectra. Cambridge, 1895. GANGE. Die Spectralanalyse. Leipzig, 1893. DE GRAMONT. Analyse spectrale directe des mineraux.Paris, 1895.

GRANDEAU.Paris,

Instruction1863.

pratique

sur

1'analyse

spectrale.

HlGGS.

Photographic Atlas of the Normal Solar Spectrum,1894.

HUGGINS.

Results of Spectrum Analysis Heavenly Bodies. London, 1870.

applied

to

the

KAYSER.

Lehrbuch der Spectralanalyse. Berlin, 1883. measurements of spectra and a very complete (Containsreview of the literature.)

Spectralanalyse in Winkelmann's Handbuch der Physik Breslau, 1894. (Encyclopadie der Naturw.).

Die Spectralanalyse und ihre Anwendung der Astronomic. Berlin, 1879. V. KOVESLIGETHY. Grundziige einer theoretischen Spectralin

KONKERFUES.

analyse.

Leipzig,' 1890.

KONKOLY. Handb. der Spectroskopiker. Halle a. S., 1890. G. AND H. KRUSS. Colorimetrie und quant. Spectralanalyse. Hamburg and Leipzig, 1891. LECOQ DE BoiSBAUDRAN. Spectres lumineux. Paris, 1874.LlELEGG.

LOCKYER.

Die Spectralanalyse. Weimar, 1867. The Spectroscope and its Use. London, 1874. - Studies in Spectrum Analysis. New York and London. 1878.

LORSCHEID.

Die Spectralanalyse.

Munster, 1875.1888.

MACMUNN.

PROCTOR. ROSCOE. Spectrum Analysis.tains popular lectures

The Spectroscope. London, The Spectroscope. London.

the author and A. Schuster.

Fourth edition, revised by London, 1885. (Con-

on the subject supplemented by

IO

SPECTRUM ANALYSIS.

extracts from the more important original memoirs r and a good bibliography.) SALET. Traite elementaire de spectroscopie, I. Fascicule.Paris,

1888.

Die Spectralanalyse der Gestirne. Leipzig, a comprehensive bibliography.) (Contains 1890. SCHELLEN. Die Spectralanalyse in ihrer Anwendung auf die Stoffe der Erde und die Natur der Himmelskorper.

SCHEINER.

Braunschweig, 1883. C. Lassel, edited by

English translation by

J.

and

W. Huggins.

New

York, 1872.

SECCHI.

Die Sonne, German by Schellen.Spectralanalyse

Braunschweig,

1872.

THALEN.

expose och

Historick,

med

en

Spectralkarta.

TUCKERMAN.VlERORDT.

Upsala, 1866. Index to the Literature of the Spectroscope.

Washington, 1888.

Anwendung

des Spectralapparates zur

Photo-

VOGEL H. W.Berlin,

metric und zur quant. Analyse. Tubingen, 1873. Prakt. Spectralanalyse irdischer Stoffe.1889.

(Deals chiefly with practical analysis,

and particularly with absorption-spectra.) WATTS. Index of Spectra. Manchester, 1889. (Contains complete measurements of spectra and a very full

Supplements to the index appeared in bibliography. the B. A. R., London.) YOUNG. The Sun. New York and London, 1897.Seealso

text-booksd.

on

physics,

amongst

others:

A.

Lommel, Lehrb.Miiller-Pouillet's

Experimentalphysik (Leipzig-, 1893); Lehrb. d. Physik, 9. Aufl. v. Pfaundler1894);

(Braunschweig,

(Breslau, 1893); Kelvin

&

Winkelmann, Handb. d. Physik Tait, Elements of Natural Philos-

ophy; Tait, Light; Tyndall,

On

Light; Wright, Light.

CHAPTER

II.

PHYSICAL PROPERTIES OF LIGHT.

1

light consists of

Huygens' universally accepted theory, wave-motions of the ether, the vibrations being transmitted from particle to particle with an extremely

ACCORDING

to

high velocity in straight lines; the vibrations of the particles On of ether are at right angles to the path of the ray. of the ether, and the ease with account of the great elasticity

which the vibrations are further propagated, single rays cannot be obtained, but only pencils consisting of a number of rays, which may be considered to be parallel if it is assumed that the vibrations are very small, or at a great distance from

The varying frequency of the vibrations produces the eye the effect of color; the number of vibrations is constant for each color, but in a given medium the waveSince all light-rays are transmitted with a length differs.the source.in

uniform velocityso inair,

in

the

number

the free ether or in a vacuum, and almost of vibrations is small or great in propor-

waves are long or short. Wave-length. It is possible to directly determine the wave-length corresponding with a given color in air, and it istion as the

found that(A) of the

at the

=1

A-Yme = 0.00076 mm., 0.000589 mm., and that of the

extremity of the visible red the wave-lengththat of the yellow Z?,-line^f-lirie

at the limit of theis

visible violet

=

0.00039

mm

-

The

velocity (v) of light4, 87,

Comp. Fehling-Hell's Handworterbuch,

and text-books

of

Phy-

sics.ri

12

SPECTRUM ANALYSIS.to be about 300,000 kilometres per second; the(n)is

known

num.

ber of vibrationsIn this mannerit

is

obtained by the expression n == A

found that the number of vibrations of

= 395, 509, and 763 billions per second These numbers are inconceivably great, and respectively. awkward to write, and it is therefore usual to define the color by the wave-length, although this varies with the medium. In dealing with wave-lengths measured in a vacuum, thethe above three linesmillionth part of a millimetre

=

o.ooi mikron

is

taken as theit is

unit, and, in accordance with Kayser's suggestion, sented by the symbol /i/ 3 -line of

lend

confirmation

cleveite,

helium; but atmospheric argon contains at least three bright lines in the violet which are not shown by the gas fromcleveite;

hence Ramsay concludes that atmospheric argon is Berthelot obtained a fluorescent specprobably a mixture.1

trum by the action of a moderately strong induction-current on a mixture of argon, benzene vapor, and mercury in a Geissler tube; the spectrum differs from that given by any other gas, and the yellow and green rays were perfectlyvisible in the spectroscope infull daylight. He considers that the spectrum is that of a compound of argon and mercury with the constituents of benzene, but Dorn and Erdmann 2

found

that

nitrogen.

spectrum between A = 5060 and 3320/1/1, using a powerful concave grating, and Kayser has published a preliminary list of the lines in the blue spectrum, the gas being obtained from4

some of the lines were those of mercury and Eder and Valenta have photographed the argon3

the atmosphere; the lines observed are not given in Rowland's

Atlas and reproductions of the Fraunhofer1

lines.

C.

r.

(1895) 120, 062, 797, 1049. 1386; (1897) 124, 113.

7 1

Ann. (1895) 287, 230. Wiener Akadem. AnzeigerLieb.

(1895),

No.d.

21.

C. N. (1895) 72, 99. Newall, C. N. 71, 115.

4

Sitzungsber.

Berl.

Akad.

(1896) 24.

See also

IOO

SPECTRUM ANALYSIS.Trowbridge and Richards1

find that the oscillatory dis-

is an important factor in producing charge of the condenser The pure red spectrum is of argon. the blue spectrum of an obtained if the tube is connected with the terminals

electric

machine; but if the spark-gap trum changes at once to blue.

is

interposed, the spec-

Red spectrum7723-

V

SPECTRA OF THE ELEMENTS.3376.61

1OI

102

SPECl^RUM ANALYSIS.Arc and spark spectra:6170.7*4467.0!3825.1!2991.11

6111.2*4459-4t3785-of

5652.1*

5559-2*4036.7!

5499-1*3949- 2 t

5332-1*3931-4! 3053.o!

4494-7t3922. 3t

443 I -7t

3119692860.542492.98

3075.442830.2!2456.61

3057-7t

3032-962601.2!2370.85 2157.1!

2898.832526.4!

2780.302437.302228.77 2067.26

2745.092381.28

2528.3!

2369.752148.2!

2349922r.i4.2i

2288.192133.92

2271.46

2165.642009.31

2113.14

BARIUM.barium has been investigated by 4 Kirchhoff, Huggins, Thalen, and Lecoq de Boisbaudran; the arc-spectrum by Lockyer, Liveing and Dewar/ and, most accurately, by Kayser and Runge, who employed theof1

The spark-spectrum2

3

6

7

chloride and carbonate,

and measured

162 lines.

Barium

compounds are gradually dissociated in a hot Bunsen flame, and all exhibit the band-spectrum of the oxide, together withlineA,

=

5

535. 69 of the metal.

Immediately on their introduc-

produce their own peculiar fugitive these can always be obtained with certainty if a wire spectra; holding ammonium chloride is placed in the flame below thetion the haloi'd derivativessalt under examination. For prolonged experiments hydrogen chloride, hydrogen bromide, or iodine vapor must be introduced into the flame. The flame-spectra

specimen of barium

of these

compounds have been studied by Mitscherlich

8

and

Lecoq de Boisbaudran.1

A. B. A. 1861.P. T. (1864) 154, 139.

a

3

N. A.

S.

U. (1868)

[3] 6.

46 *'

Spectres lumineux (Paris, 1874). P. T. 163, 369; 164. 806.Ibid. (1883)

8

174, 216. A. B. A. 1891. P. A. (1862) 116, 419;

(1863) 121, 459.P. A. 110, 161.

also

Bunsen and Kirchhoff, Only Only

For the flame-spectrum see Bunsen, P. A. (1875) 155, 366.

*!

visible in the spark-spectrum.

visible in the spark-spectrum.

(Thalen.) (Hartley and Adeney.)

SPECTRA OF THE ELEMENTS.Arc and spark spectra:6497 07

IO3

104

SPECTRUM ANALYSIS.BISMUTH.

spark spectrum is obtained by the use of bismuth 2 electrodes, and has been measured by Huggins, Thalen, and1

The

the arc-spectrum by Liveing and and recently, commencing at 6i8/f/w, by Kayser and Dewar, The spark-spectrum exhibits many lines that are Runge. Bismuth salts, moistened with absent from that of the arc. hydrochloric acid, produce in the Bunsen flame a fugitive

Hartley and Adeney;4

3

6

band-spectrum of the oxide. The spectra of the compounds themselves are obtained by volatilizing them in a hydrogen 6 flame. They have been drawn by Mitscherlich.

Arc and spark spectra:6493.8*

^

SPECTRA OF THE ELEMENTS.!

105

third,

have also been observed by Hartley in the sparkfound, however, fourteen spectrum; Eder and Valenta2

additional lines, the majority of which are double, and confirmed the presence of four that had been detected by

have recently photothe arc-spectrum between A 2 100-4400. graphed Onl)' the double line could be detected; the numerous bands areCiamician.

3

Rowland and Tatnall

4

=

probably due to some compound, such as boric anhydride. Boric acid and its salts produce a characteristic band spectrumin the

Bunsen flame. Arc and spark spectra:3450.8*(2497.8212496.867)5

2267.0!

2266. 4f

BORIC ACID.

The wave-lengthFlame-spectrum63985193:

is

measured6032 4722

at the

middle of the bands.

6211

5808

5481

5440

4912

4530

BROMINE. Bromine vapor gives a line-spectrum with the electric spark, but the measurements of it are only approximate.6

Its

absorption-spectrum at the ordinary temperature has been7

accurately investigated by Hasselberg;1

when

a high disper-

P. R. S. 35, 301.

2 3

Denkschr.Sitzber. d.

d. Wien. Akad. (1893) 60, 307. Akad. d. Wiss. zu Wien. [2] 82,r.

425.

See also Troost and

Hautefeuille, C.45

(1871) 73, 620.

Salet, A.

c. p.

(1873) [4] 28, 59.

Astrophys. Jour. (1895) 1, 16. Lecoq de Boisbaudran, Speqtres lumineux (Paris, 1874). Thalen, Upsal. Universit. Arsskrift. 1866. Also Salet and Eder, and Valenta, as above. 6 A. c. p. [4] 28, 26. Plucker, P. A. Salet, Spectroscopie (Paris, 1888). Plucker and Hittorf. P. T. 155, i. Ciamician, Wien. 105, 527; 107, 87.Ber. 767

[2],

499;

77

[2],

839;

78

[2], 867.

K. Svensk. Akad. Handlingar. (1891) 24, No. 3. Mem. de 1'acad. de St. Petersb. (1878) 26, No. 4. See also Daniell and Miller, P. A. 28, 386. Roscoe and Thorpe, P. T. 167, 209. Moser, P. A. 160, 188.* Visible only in the spark-spectrum. f Visible only in the spark-spectrum.

(Hartley.)

(Eder and Valenta.)

io6sionfine

SPECTRUM ANALYSIS.employed lines groupedisit

seen to consist of a large into bands.is

number

of

The spectrum obtained with a continuous discharge differs from that produced when a condenser is included in thecircuit.1

Spark-spectrum of bromine vapor:7000*6148

6780*5876(5450

6630*58305423)

65835723(5327

65595595305

6546(5509

63535497

54905166)

5240

5184

5060

49303980:

(4816

4788)

4705

4677

4618

4366

Absorption-spectrum

SPECTRA OF THE ELEMENTS.Group

lO/

108chloride,

SPECTRUM ANALYSTS.

90 Cadwhich have been measured by Hartley and Adeney. mium chloride and bromide are dissociated in the Bunsen flame, and exhibit the lines X = 5086, 4800, and 4678,1

but usually the metal. The arc-spectrum differs latter exhibits a pair considerably from that of the spark; the 5378.8 and 533$. 3> of lines of the highest intensity of \ absent from the former, and the same applies to which are 4215 and 2111 lines of inferior brightness between A

=

Arc and spark spectra:6467.3*4678. 37f

6438.8*4662.69 3261.172639.632239.93

5378.8*4413.23(3252.63

5338.3*3613 043133.292573.122144.45

5154-85 3610.663081.03)

5086.06f3467.762980.75 2312.95

4800.09f 3466.332880.88

3403.742763.992267.53

2601.992194.67

2329.35

2288 10

CESIUM.,

2

Bunsen and KirchhofT discovered caesiumof

in

1861

by

means

spectrum analysis.

Its salts are all dissociated in

the Bunsen flame, and exhibit the lines of the metal; the more prominent ones are X 4555 and 4593 in the blue, andA.

=

6010 in the orange. Arc, spark, and flame spectra:6973.96723.6

6213.43888.83

6010.63876.73

5845.1

5664.0

5635.1

4593-341

4555-44

3617-08

3611.84

hoff,

P. T. (1883) 175, 98. See also Thalen, N. A. S. U. (1868) [3] 6. KirchA. B. A. 1861. Mascart, Annales de 1'Ecole normale (1866), 15.

Lockyer, P. T. (1873) 163, 369. Cornu. Journ. de Phys. (1881) 10, 425. Liveing and Dewar, P. R. S. (1879) 29, 482. P, T. (1888) 179, 231. Ames, P. M. (1890) [5] 30, 33. Bell, Am. Journal oi Sciences, June, 1886. 2 Kayser and Runge, A. B. A. 1890. Bunsen, P. A. 119, i; 155, 366. Johnson and Allen, P. M. [4] 25, 199. Thalen, N. A. S. U. (1868) [3] 6. Lecoq de Boisbaudran, Spectres lumineux (Paris, 1874). Lockyer, P. R. S.(1878) 27, 280.

Liveing and Dewar,

Ibid. (1879) 28, 352.

*

Only

visible in the spark spectrum.

(Thalen.)

f Visible

also in

the flame

spectrum of the chloride and bromide.

(Lecoq de Boisbaudran.)

SPEC7^RA OF

'1

HE ELEMENTS.

109

CALCIUM.been measured by 4 Lecoq de Boisbaudran, Huggins/ 6 5 Lockyer, Cornu, Liveing and Dewar, and more recently by who employed the electric arc and Kayser and Runge, The faint bands which occasionally appear calcium chloride. in the yellow and red when the arc is used are considered byline-spectrum1

The

of

calcium3

has7

Kirchhoff,

Thalen,

8

be due to oxide. Many of the and are therefore always calcium produced, H. BecquereP visible when carbon electrodes are employed. observed bands from A = 8880-8830 and from A = 8760-8580 The haloid compounds have been investiin the infra-red. 10 gated by Bunsen, Mitscherlich, and Lecoq de Boisbaudran; in the Bunsen flame some bands peculiar to each compound are visible, together with those of the oxide and the blue The oxide bands are also line, A 4226.91, of the metal.

Lecoq de Boisbaudran

to

lines are readily

11

=

produced

if

the flame

is

charged with hydrogen chloride,fluoride.

hydrogen bromide, hydrogen iodide, or hydrogen Arc and spark spectra:6499.85

I I

O4355-

SPECTR UM ANAL YSIS.

SPECTRA OF THE ELEMENTS.investigated by

Ill

Swan from 1850 onwards.2

'

In3

common

with4

Angstrom, Thalen, and Liveing and Dewar, he ascribed it to hydrocarbons, but the last workers, together with Attfield,5

Morren,

and Dibbits,

8

due

to carbon, since

it is

subsequently recognized that it was produced by the combustion of pure

cyanogen in dry oxygen. This band-spectrum consists of five complex bands, with the following wave-lengths according to Angstiom and Thalen, and Watts:7

i

or orange band.'

y^jlow^band

3

or & reen band

-

4

or blue band.

5

or indigo

band

6187-5954

5633-5425

5164-5032

4736-4677

4381-4232

three medial bands have been recently measured by and the green one by Kayser and Runge; 9 for the Fievez, o8

The

others there are only the old observations of Watts, Angstrom 10 available. In addition to and Thalen, and Piazzi- Smyth the above bands others are sometimes observed in the arc; they occur in the blue, violet, and ultra-violet, and have thefollowing wave-lengths:I

Band.

II

Band.

Ill

Band.

IV Band.

V

Band.

4600-4500

4290-4150

3884-3850

3590-3550

3370-3350

in the arc is doubted by Kayser and and Dewar have ascribed them to cyanogen, Runge; Liveing H. W. Vogel, and others regard them as a whilst Lockyer,

Their existence11

1

P. T. E. (1857) 21, 411.

2

N. A. S. U. (1875)

9.

3

P. R. S. (1880) 30, 152, 494; (1882) 33, 403; (1883) 34, 123.Ibid. (1862) 152, 221.

P. T. (1882)

174, 187.P.

M.

(1875) 49, 106.

A.

c. p.

(1865) [4] 4, 305.

P. A. (1864) 122, 497. P. M. (1869) [4] 38, 249; 45, 12; (1874) 48, 369, 456; (1875) 49, 104.

Mem. de 1'Acad. roy. de Belgique (1885), 47. A. B. A. 1889. P. M. (1875) [4] 49, 24; (1879) [5] 8, Astr. Obs. Edinb. (1871) 13, 58. P. T. E. 30, 93. 107. 11 P. R. S. (1878) 28, 308; (1880) 30, 335. See also Plucker, P. A. (1858) 105, 77; (1859) 107, 533, and with Hittorf, P. T. 155, i. Jahresber. (1864) Van der Willigen, P. A. (1859) 107, 473. Huggins, P. T. (1868) p. no.IJ

1 1

2

SPE C TRUM A NA L YSIS.

second band-spectrum of carbon produced only at high tembut experiments Kayser at first shared this view, peratures.

made

A

a different conclusion. conjunction with Runge led to was directed on to the strong current of carbon dioxidein

arc,

appeared.

fainter and diswhereupon the cyanogen bands became that this was not due to a In order to prove

current of air was lowering of the temperature a still stronger substituted for the carbon dioxide: the bands immediately increased in brightness in consequence of the additional The view that the cyanogen bands are supply of nitrogen. their occurrence in essentially due to carbon is supported by last fact was long comets, and in the solar spectrum; this That the specdoubted, but was established by Rowland.

trum

of a

compound which

is

dissociated at 1000

visible in

the solar spectrum appears

should be somewhat paradoxical,

but Kayser and Runge have pointed out that the carbon molecule, as is shown by its varying specific heat, is not a constant quantity, and the "cyanogen bands" maybe the spectrum of an unknown compound of carbon and nitrogen

which

is

capable of

existence at very high temperatures.

The cyanogen bands have been measured by Liveing andDewar, and the and Runge.third, fourth,

and

fifth

ones also by Kayser

The carbon bands all have their brighter edges directed towards the red end of the spectrum; each possesses a number of edges, varying from three to seven, which become weakertowards the violet. The lines extend from the first edge of one band to the beginning of the next, so that no portion of the spectrum from 620^ to 340^ is free from carbon lines, the total number of which is at least 10,000. Metallic spectraobtained by means of the arc and carbon poles always exhibit158, 558.r.

(1871) 73, 620.

Lielegg, Wien. Ber. (1868) 52, 593. Troost and Hautefeuille, C. Wullner, P. A. (1872) 144, 481. Salet, A. c. p. (1873) [4]

28, 60.

Lecoq de Boisbaudran, Spectres lumineux (Paris, 1874). Wesendonck, Inaug-Diss. (Berlin, 1881). Hartley, B. A. R. 1883. Eder, Denkschr.

Wiener Akad.

(1890) 57.

SPECTRA OF THE ELEMENTS.

11$

the metallic lines superposed by the carbon bands, hence a

good knowledge6584.0

of the latter

is

very desirable.

Line-spectrum:6578.53920.02478.35695.1

5661.9

4266.62509.0

2837.22296.5

2836.3

Band-spectrum:I.

II.

Orange Band 6187 Green-yellow Band:i.

5954:

6188. 2

6119.95633-4

Edge:

5635.3

5634-3

5633-9

1144728.42.

SPECTRUM ANALYSIS,

Edge:

3-

Edge:

4.

Edge:

SPECTRA OF THE ELEMENTS.

1*5

SPECTRUM ANALYSIS.

IV.

SPECTRA OF THE ELEMENTS.'

\\'J

observed that the chlorine-spectrum is produced by passing powerful sparks through glass tubes containing The spectrum chlorine compounds under low pressure.

berg

obtained with a powerful continuous current differs from that 8 produced when a condenser is introduced into the circuit.

There are no recent measurements of the absorptionspectrum of chlorine. Spark-spectrum:

5457.8

5444-7

5425.0(4905.4

5393-44897.8)3

(5220.8

52172)4810.7

5103.2

5099.0

5078.4

4918.1

4820.8

4794.9

Absorption-spectrum:

Numerous absorption-bandsin the violet.

in the

green and blue, total extinction

CHROMIUM.of chromium between D and A = 3430 measured by Hasselberg. Huggins and has been accurately Thalen had previously investigated lines of the spark-spec-

The arc-spectrum6

4

5

trumlines

in

the visible region, LockyerA.

between worked on thetions of1

=

observed some additional and 3900, and Liveing and Dewar 8 4000Solucharacteristic absorp-

7

ultra-violet portion of the arc-spectrum.

chromium compounds produceAcad.St.

Bull.

Petersb. 28, 405.r.

106, 624.

Ditte, C.

73, 622.

Pliicker,

See also Van der Willigen, P. A. P. A. 107, 528. Pliicker and

Hittorf, P. T. 155, i. Angstrom. C. r. 73, 369. Thalen, K. Svenska Vetensk. Akad. Handl. 12, No. 4, p. 8. Salet, A. c. p. [4] 28, 24. Ciamician, Wien. Ber. 78, 872.2

Trowbridge and Richards, Amer. Jour.

Sci.

(1897) [4] 3, 117.

P.

M.

43, 135-

Morren, P. A. (1869) 137, 165. Sillim. Journ. [2] 47, 417. Svensk. Vetensk. Akad. Handl. (1894) 28, N. 5P. T. (1864) 154, 139.

N. A.

S.

U. [3]

6.

P. T. 1881.

P. R. S. (1881) 32, 402. See also

1878, p. 413.

Ber. 8, 1533.

ches sur

le

spectre solaire.

H. W. Vogel, Monatsber. Berl. Akad. Kirchhoff, A. B. A. 1861. Angstrom, Recher1868. Lecoq de Boisbaudran, Spectres lumi-

neux

(Paris, 1874).

1

101

SPECTRUM ANALYSIS.

tion-spectra that have been frequently studied; the more The spectroimportant are shown in Fig. 41, Chapter VIII. scopic

determination

of

potassium

chromate,

potassium

bichromate, and chrome alum is described in G. and H. Kruss* work on Colorimetry and Quantitative Spectrum Analysis.

Arc-spectrum5791-20

:

SPECTRA OF THE ELEMENTS.that of the arc; Lockyer portions of the spectrum.l

119

have also investigated has recently measured and A = 3450, and the lines of the arc-spectrum between 4 and Dewar, those of the arc and spark spectra in the Liveing8

and Cornu

*

Hasselberg

D

ultra-violet region.

The

lines of the

two spectra

differ

not

number, but also in intensity. The absorption-spectra of cobalt glass, and of solutions of cobalt compounds are very & characteristic; they have been examined by H. W. Vogel, Russell, Russell and Orsman, and by C. H. Wolff, and are shown in Fig. 41, Chapter VIII. The following test is stated by Wolff to be one of the most delicate known in chemistry: Ammonium thiocyanate is mixed with cobalt chloride solution, and shaken with amyl alcohol and ether; this dissolves the cobalt thiocyanate, and the solution gives a characteristic absorption-spectrum. The method was used for the spectroonlyin6 7 8

colorimetric determination of cobalt

when presentof cobaltJ.

in

smallin

quantity.

The absorption-spectrumis

chloride

hydrochloric acid solutionparticularly sharp, butif

statedis

by W.

Russell to be

the acid

concentrated the broad

bands usually observed are resolved into smaller ones, almost coincident with those produced by ferric chloride under the

same conditions.

He

believes that the solvent causes a dis-

sociation of the dissolved substance.

Arc and spark spectra:6143.8*5483.571

6122.3*5477.13

5531-06

5525-275444-81

5523-565407-75

5489.905381.99

5484.22

5454-793.

5369.79

P. T. 1881.

Part

2

3

Spectre normal du soleil (Paris, 1881). Svenska, Vetensk. Akad. Forhandl. (1896) 28, No.

6.

From D

A 3450.

Astrophys. Jour. (1896) 3, 288; 4, 343; (1897) 5, 38. 4 From A. 3450-2244. P. T. (1888) 179, 231. 5 Ber. 11, 916. Monatsber. Berl. Akad. 1878, p. 415.67

J.8

Ber. 14, 503. Ber. 24, 619. See also Etard, C, Zeitschr. Anal. Chem. 18, 38.P. R. S. 31, 51.

Chem.

Soc. 1889, p. 14.

r.

(1895) 120, 1057.

* Visible only in the spark-spectrum.

I2O5362.97

SPECTRUM ANALYSIS.

SPECTRA OF THE ELEMENTS.

121

COPPER.

The spark-spectrum of copper, in the visible field, has been 2 measured by Kirchhoff, and Thalen, and, as far as wavelength 4275, by Lecoq de Boisbaudran; in the ultra-violet 4 Hartley and Adeney have measured the portion between that A 3999 and 2103, and Trowbridge and Sabine between A = 2369 and 1944. Liveing and Dewar" have 2294-2135, whilst, photographed the arc-spectrum from A1

3

&

Kayser and Runge have done the same for the between A. = 6000 and 1944; they measured 304 lines, region and obtained the spectrum by substituting, for the carbon poles, rods of copper 1-2 sq. cm. in section.

more

7

recently,

on one

Scarcely any of the copper lines are sharply defined even side, so that the spectrum has a peculiar appearance. In the Bunsen flame cupric chloride produces a band-spectrumfield,is

extending over the wholeviolet;

the same spectrum

with the exception of the obtained with the metal if the

The absorption-spectra of flame contains hydrogen chloride. salts are not characteristic, as the compounds produce coppertotal extinction both in the red

and the

violet.

Ewan found

B

that, in aqueous solution, the spectra of the chloride, sulphate, and nitrate change with progressive dilution tending to become identical; this observation is in agreement with the

C. H. Wolff has sugtheory of electrolytic dissociation. gested a spectro-colorimetric method for the determination of

9

A. B. A. 1861. N. A. S. U. (1868)

6.

Spectres lumineux (Paris, 1874). P. T. (1883) 175, 63. Proc. Amer. Academy, 1888. P. M.P. R. S. (1879) 24, 402.

[5] 26, 342. P. T. (1883) 174, 205.

A. B. A. 1892.89

P.

M.

1892, p. 317-

Ber. 25, 495c.18, 38.

P. R. S. (1895) 57, 128.

Zeitschr. anal.

Chem.

122

SPECTRUM ANAL YSIS.'

copper

in small quantity,

and P. Sabatier

has studied the

of cupric bromide. absorption-spectra of solutions

Arc and spark spectra:5782.30

4704.77 4480.593602.1 1

3247.65

2618.462303.18

2199.77

SPECTRA OF THE ELEMENTS.

123

substances have not been sufficiently studied to render their Kriiss recognition as elements absolutely free from doubt.

and Nilson, as thetion-spectra,

result of their investigation of the absorp-

that didymium, erbium, holmium, and thulium are composed of more than twenty samarium, elements; their conclusion is based on the assumption that each element has a characteristic maximum of absorption, butcious.

consider

Schottlander's extensive investigations show that this is fallaBailey has also raised objections to Kruss and Nil-

son's conclusions.

At

present the results of the spectroscopic

work on the rare earths is so uncertain, that the data given in " " this book usually refer to the old elements. Bahr and Bunsen found that didymium oxide, like erbium oxide, when heated in the Bunsen flame gives a continuous spectrum, andalso

characteristic bright lines

which are almost coincident

with the absorption-lines exhibited by solutions of the salts, or by glass which contains the metal; this is no exception to the rule that solids only yield continuous spectra, for Huggins

and Reynolds showed that the earths are volatile

in

the oxy-

hydrogen

flame.

didymium

Comparison of the absorption-spectra of chloride, sulphate, and acetate led Bunsen to the

conclusion that the bands tend to approach the red as the molecular weight increases. The absorption-spectra of therare earths in the ultraviolet has been investigated

by Soret.

Spark-spectrum:5486.0*5192.5

5372.05191. 5

5361.55130-341.09.8

5319-9

5293-5

5273-5

5249.5

4924.5 4060.7

4463.2

4452.3

4446.7

4328.1

4303.6

*

Due

possibly to samarium.

124

SPECTRUM ANALYSIS.DIDYMIUM CHLORIDE.Absorption-spectrum:'

b (7431.7* 5 886f5206)

7361.7* 5824*(5125.8*

7308.7*)

6895.6*

6793.3*5720)

[>]b(5963t[/?]

5789*5088*)

5748*4823!

b (5313*

4759

4692!

4441.7

'OLD" DIDYMIUM NITRATE. Absorption-spectrum: positions of minimum4

of brightness:575946335317

72915253

6906 52174289.6

679451474173-6

64075126

6235

61894771

57974695

4826

4443

4341

PRASEODYMIUM NITRATE.729167945916

5797

5759

531?

52172

5125

4826

4695

4443

ERBIUM.

The remarks on didymium applySpark-spectrum58274899.9:

also to this element.

i

5763487.2.4

5344-4

52574674-9

52i84606.3

51894501.3

4952

4820

Lecoq de Boisbaudran, Spectres lumineux (Paris, 1874). C. r. (1887) Bunsen, P. A. 155, 366. Ann. Chem. Pharm. 128, 100; 131, 255. Huggins and Reynolds, P. R. S. 18, 546. Lippich, Sillim. Journ. (1873) [3] 13, 304. Auer v. Welsbach, Sitzungsber. Wien. Akad. (1885) 92. Crookes, C. N. 54, 27. Schuster and Bailey, B. A. R. 1883. H. Becquerel, C. r. 104, 777, 1691; 106, 106. Haitinger, Monatsch. f. Chem. (1891) 12, Kruss and Nilson, Ber. 20, Soret, C. r. 86, 1062; 88, 422; 91, 378. 362.1

105, 276.

2143.569.2

Bailey, Ber. (1887) 20, 2769, 3325.

Schottlander,

Ber. (1892) 25,

Spectra Yttrium, Erbium, Didym och Lanthan (StockBunsen Ofversigt K. Vetensk. Akad. Fdrhandl. (1881) 40. and Bahr, Ann. Chem. Pharm. 137, i. Huggins, P. R. S. 1870. Bunsen, P. A. 155, 366.

Thalen, holm, 1874).

Om

*f

Neodymium.Praseodymium.

SPECTRA OF THE ELEMENTS.ERBIUM CHLORIDE.Absorption-spectrum6839 53646671[a] 5232:

12$

1

6535

6405 4875

5410

4922

4516

FLUORINE.no accurate measurements of the spectrum of fluorine. By the passage of induction-sparks through silicon a obtained a beautiful blue band-spectrum of fluoride Salet the compound; the incission of a Leyden jar produced the

There

are

spark-spectrum of fluorine.3

Commencing

at Salet's last lines

Liveing

measured the flame-spectrum. Spark-spectrum:

6922*

6862*:

6782*

6401

6231

Flame-spectrum6231

6091

6011

5571

5321

GALLIUM.

Lecoq de Boisbaudran, who discovered this element, measured its spark-spectrum, and Liveing and Dewar that of5

4

the arc.

Spark and arc spectra:41714031

GERMANIUM.of germanium has been investigated and by Lecoq de Boisbaudran, who calculated its by Kobb,8 7

The spark-spectrum

Lecoq de Boisbaudran, Spectres lumineux (Paris, 1874). Bunsen, P. A. 155, 366. Bunsen and Bahr, Ann. Chem. Pharm. 137, i. 2 A. c. p. (1873) 28, 34. 3 Proc. Cambridge Phil. Soc. 3, Pt. 3. See also Mitscherlich, P. A. 121,1

476.46 61

Seguin, C.C.r.

r.

(1862) 54, 933.

82, 168.

P. R. S. 28, 482.

W.C.

A. (1886) 29, 670.r.

(1886) 102, 1291.

Ber.

l,

479.

*

Only approximate.

126

SPECTRUM ANALYSIS.*

atomic weight from his measurements. Rowland and Tatnall have recently photographed the arc-spectrum between A =

2300-4600. Arc and spark spectra:63375131

6021 4814

58934743

5256.5

5229.5

52104261

5178.5

5135

4685.3

4291-6

4226

41792709.734*

3269.628*

3124.945*

3039.198*2651. 219f)

2754.698*

2740.535* 2417.450*

2691.446* (2651.709*

2592.636*

GOLD.Liveing and Dewar measured three lines in the ultra-violet exhibited by the arc-spectrum, and these were the only ones*

known when Kayser and Runge commencedtion of the region

3

their investiga-

between wave-length 6600 and 2280. They

usually employed fine gold, but occasionally auric chloride and carbon poles. The visible portion of the spark-spectrum

has been measured by KirchhorT, Huggins, Thalen, 6 and G. Kriiss, who ascribes the line 5230.47 to platinum, and not7

4

5

to gold.

Arc and spark spectra:6278.374488.46 2676.055957-245837-64 3122.88

5656.00 3033-38

5230.47

5064.75

4792.792905.98

4065.22 2428.06

3029.32

2932.33

HELIUM.

The yellow /Vline

of the solar

until recently, to a hypothetical element,Astrophys. Jour. (1895) P. T. (1882) 174, 2219. A. B. A. 1892.Ibid.1

spectrum was ascribed, termed by Frank-

1, 149.

86 1.

P. T. 1864, p. 139.

N. A.Lieb.

S.

U.

[3]

6.

Ann.

(1887) 238, 30.1874).

See also Lecoq de Boisbaudran, Spectres

lumineux (Paris,*f

Arc-spectrum.Possibly a single reversed line.

SPECTRA OF THE ELEMENTS.1

I2/

land helium; this was isolated by Ramsay in 1895 from cleveite, in which it occurs together with argon; he also obtained it from certain meteorites from Augusta Co.,Virginia.

Clevein

2

showed thatin its

it

is

present,in

by argon,to D^.

cleveite

from Carlshuus

unaccompanied Norway, and he

observed the presence3

spectrum of

five lines in addition

Deslandres, the following lines:66784437.9 3819.7

using a very high dispersion, measured

5876.04388.43705.4

5048.4

5016.0

4922.24026.23187-9

4713.35

4143-93613-84

4120.93447-7

3964.02945.7

4471.75 3888.75

Runge and Paschen, in the course of an investigation of the gases from cleveite, showed that the line at 5876.0 is a double one, and, as the solar helium line had always been regarded as single, doubt was cast on the identity of solarThis point was speedily settled by and Hale, who showed that the solar Z> -line is also Huggins double. Kayser found that a Geissler tube containing what he supposed to be the purest atmospheric argon also showed

and

terrestrial helium.5

8

6

the Z> 3 -line, thus affording proof that helium7

is

present in the

helium lines many Lockyer atmosphere. coincide with some of hitherto unknown origin in the spectra of the chromosphere, and of the white stars of Orion. Understates that

of the

the influence of the silent discharge helium combines with mercury and benzene or carbon bisulphide to form a com-

pound resembling that mercury alone.1

of argon, but

it

does not combine with

C.

r.

120, 660, 1049. 120, 834.

Ber. 28, 318, 448.

N. 52, 224.

Ramsay,

Collie,

and Travers, Jour. Chem. Socy.2

(1895) 67, 648.

C. C.

r.r.

Ber. 28, 373.

3

45 61

W.

(1895) 120, nio, 1331. A. Beibl. 19, 634.

C. N. 72, 26.Ibid. 72, 99.

See also Brauner, C. N. 71, 271. P. R. S. 58, 67. C. r. 120, 1103. Palmieri, Acad. di Napoli Rendic. (1882) 20, 233.

128

SPECTRUM ANALYSIS.Berthelot*

and Runge and Paschen

a

have observed that

the spectrum of cleveite gas consists of six series, two pairs of which are characterized as subseries, whilst two series are

Two spectra are thus differentiated which principal series. are ascribed to two constituents of the gas, and which bear a 3 Rydberg striking similarity to the spectra of the alkalies.has confirmed these conclusions, and termed the second constituent parhelium.thisin the separation of

Some

confirmation was also afforded

to-

view by Ramsay and Collie's researches, which resulted helium into a lighter and a heavier por-

tion; but

Ames

difference in

were unable to detect any and Humphreys their spectra, although they used a spectroscope

4

of high dispersive power. When further separated the heavier portion was found to consist chiefly of argon.

HYDROGEN.

Two spectra, termed the elementary and compound linespectra, are exhibited by hydrogen in a Geissler tube; theirproduction depends on the conditions of temperature and The former has been measured by Angstrom, 5 pressure.

H.

W.

8

7

Vogel,first

latter

was

Huggins, Cornu and Ames; the 10 but was investigated by Plucker and Hittorf,Lockyer,9

8

11 ascribed to acetylene by Angstrom, Berthelot and Richard, 12 and Salet. The incorrectness of this view was

proved by

(1897) 124, 3 2 Sitzber. Berl. Akad. (1895) 639, 759. Astrophys. Jour. (1896) 3, 4..

1

C.

r.

n

W. A.

Beibl. (1895) 19, 884, 885.

3

W. A.

(1896) 58, 674. Astrophys. Jour. (1896) 4, Astrophys. Jour. (1897) 5, 97. P. A. (1864) 91, I4I; 123, 489; (1872) 144. 300. Berl. Monatsber. (1879) 586; (1880) 190. Ber.P. R. S. 28, 157 P. T. 171, 669. P. M. (1890)[5]1011;

91.

(1880) 13, 274.

30, 31.

30, 33.

P. T. (1865) 155, 21. C. R. 68, 810. 1035,

Plucker, P. A. 107, 4Q 7.1107, 1546.

12

A.

c.

p. [4]

28,

17.

SPECTRA OF'

J^HE

ELEMENTS.

129

In the spectrum of exact measurements. Hasselberg's 2 in addition to dark hydrogen C Puppis Pickering found, lines and K, two broad lines at A 4633 and A 4688, and

=

a peculiar series of dark lines whose wave-lengths are rhythmiThese were A 4544, 4201, 4027, 3925, 3859, cally related. It was first thought that they represented some 3816, 3783.

new element not yet found on

the earth or in the stars, but

they are very probably due to hydrogen, produced under conditions of luminosity hitherto unknown. By applying

Balmer's formula, Pickering found that the new lines form a 3 harmonic series. This conclusion was confirmed by Kayser, who pointed out that hydrogen had been the only element,

having harmonically related

lines,

which had possessed only

a single series of such lines. Kayser and Runge had previously found that two of the series of lines of an element endat

nearly

the

same

place.lines,

On examining

the oscillation

frequencies of thethis characteristic,

new

Kayser concluded that they have

and constitute a new hydrogen series. If in laboratory experiments, imthese lines can be produced portant information as to stellar temperatures and pressures is likely to be obtained. At low temperatures ivater vapor gives an absorptionspectrum rich in lines which are chiefly confined to the redthese constitute a large number of the terrestrial lines, and are referred to under nitrogen they are strongest when the sun is low in the horizon, as its raysregion;

Fraunhofer

;

have then to traverse a considerable layer of the atmosphere. " " When the latter is saturated withmoisture, a

rain-band

is

visible with the1

help of a spectroscope of

low dispersive

Bull. Acad. St. Petersb. (1880) 11, 307;

St. Petersb. (1882)

30, No.

7;

Mem. Acad, (1884} 12, 203. W. A. 15, 45. See also (1883) 31, No. 14.Wiillner,P.

H. C. Vogel,144, 481.2

P. A. (1872) 146, 569.

A.

135, 497;

137, 337;

W.

A. 14, 355.

Seabroke, P. M.

[4]

43, 155.

Balmer,

W. A.

(1885) 25, 80.8

Astrophys. Astrophys.

J. J.

(1897) 5, 92.

(1897) 5, 95.

Science (1897) Science (1897)

5, 726. 5, 726.

I

30

SPK C'7'A' UM A A 'A L YSIS.

power: it consists of bands composed of water- vapor lines, The and situated between the red end and the ZMine. of the rain-band has been used by Piazzi-Smyth, presence1

Capron, Grace and others as a means of prognosticating rain. s Janssen investigated the absorption-spectrum of steam contained in long tubes under considerable pressure, and

2

Schonn

Anit

states that pure water exhibits an absorption-band. emission-spectrum consisting of numerous lines in theis

4

ultra-violet

through the electric arc;also

obtained by burning hydrogen in air, or passing it has been measured by Huggins,5

by Liveing and Dewar, who distinguished five series The first, between wave-length 3268.2 and 3063.7, of lines. 16 lines; the second, from 3057 to 2812, contains aboutandI

comprises 180 lines; the third, from 2807 to 2609, contains 141 lines; the fourth series, from 2606 to 2450, has 88 lines;

and the

fifth

series includes

79

lines

between wave-length

2449 and 2268. Elementary line-spectrum:[C or Ha] 6563.04 [F or H/3] 4861.49 [G or [H] 3970.25 [a] 3889. 15[ 2 ] 5890. 19*4983.53

5682.86

4979-30

3303.07

3302.47

2680.46

STRONTIUM.Mitscherlich obtained the line-spectrum by the use of the oxyhydrogen flame it is also produced by the passage of;

sparks through a solution of the chloride, but the best effects are given by the volatilization of the chloride in the electricarc.

This was the method employed by Kayser and Runge. 3 In the Bunsen flame the strontium haloid compounds chiefly

exhibit their individual spectra, together with the band-spectrum of the oxide, and the blue line, A. 4607.5, of the metal.

=

1

T. R. S. E. (1857) 21.

167.

Bunsen and Kirchhoff, P. A. 110 1890. Kirchhoff, A. B. A. 1861. Rutherfurd] Attfield, P. T. 1862, 221. Silliman's Journ. [2] 35, 407. Huggins, P. T. 1864, p. 139. Wolf and Diacon, C. r. 55, 334. Mtiller, P. A. 118, 641. Thalen, N. A. S. U. (i868>Kayser and Runge, A. B. A.Lecoq de Boisbaudran, Spectres lumineux (Paris, 1874). Lockyer 29, 140. Cornu, Spectre normal du soleil (Paris, 1881). Bunsen, P. A. 155, 366. Liveing and Dewar, P. R. S. 28, 367, 471 (1879) De Gramont, C. r. E. Becquerel, C. r. 94, 1218 97, 72. 29, 398, 402.[3] 6.

2

P. R. S. (1879)

;

;

(1896) 122, 1411, 1443. 8 A. B. A. 1891. See

also

Kirchhoff, A. B. A. 1861.p. 139.

Mtiller,

Bunsen and Kirchhoff, P. A. 110, P. A. 118, 641. Huggins, P. T.;

167.

1864,

Mascart, Annales de 1'Ecole normale (1866) 4. Thalen, N. A. S. E. Becquerel, C. r. U. (1868) [3] 6. Lockyer, P. T. 163, 639 164, 311. 96, 1218 (1883) 97, 72. Liveing and Dewar, P. T. 174, 217. Rydberg, W;

A. (1894) 52, 119.Visible also in the flame-spectrum.

SPECTRA OF THE ELEMENTS.Arc and spark6550.53

I&3

spectra:

164

SPECTRUM ANALYSIS.

The band-spectrum is desired. spectrum when it is not of feeble sparks through a Geissler obtained by the passage Salet produced it by volatube containing sulphur vapor. or one of its compounds, in a hydrogen tilizing sulphur, it to impinge on a plate of metal or cooled'

flame,

by allowing of water was marble, on to the other side of which a stream

This spectrum was mapped by Salet, and by Pliicker and Hittorf, but the observations are limited to the and are too inaccurate to show more than the visibledirected.

existence

region, of the

bands.

The

flame

of

burning sulphur

exhibits a continuous spectrum which extends far into theviolet.

Line-spectrum:5660.7 5342.6

5640.35320.1

5604.9 5215.44902.8

5562.4 5201.1

5508.3

547*-4 5103.7

5451.9 5033.3

5439-

543O.7

5M3-34816.6

5013.5

4926.0 4464.7

4919.4

4885.4

4715.8

4552.3

4525.5

4994-7 4485.9

2

Band-spectrum. violet, but shading536652215191

Bands sharply bordered towards thetowards the red (Salet):49915041

off

5089

4946

4841

4796

4656

4616

4471

TANTALUM. Thelines of this

element were too feeble for Thalen to3

measure, but Lockyer observed eighteen of them in the arc4000 and 3900. spectrum between A

=

TELLURIUM.obtained by passing sparks between electrodes of the element, and has been measline-spectrum oftelluriumis

The

A. c. p. [4] 28, 37. C. r. (1869) 68, 404. See also Hasselberg, Bull. Acad. imp. St. Petersb. (1880) 11, 307. Astronomy and Astrophysics (1893), 12, 347. Mulder, J. pr. Chem. (1864) 91, 112. Ditte, C. r. 73, 559. Lock1

yer, P. R. S. 22, 374.300.

C.

r.

73, 368.p. 272.r.

Gernez, C. r. (1872) 74, 803. Ciamician, Wien. Ber. 77, 839

Angstrom,;

82, 425.

P. A. 137, Schuster, B.

A. R. 1880,

Ames, Astronomy and Astrophysics28, 37.C.r.

(1893),

12.

De

Gramont, C.9 8

(1896) 122, 1326.

Salet, A.

c. p. [4]

(1874) 79, 1231.

P. T. 173, 561.

SPECTRA OF THE ELEMENTS.1

I6 5

ured by Huggins, and Thalen in the visible region, and by 3 Salet * produced a Hartley and Adeney in the ultra-violet.

8

band-spectrum by passing a discharge through a Geissler tube of hard glass containing tellurium; to facilitate heating, the tube was covered with metal. The spectrum consists ofthe red, and channelled spaces in the green and blue; they are sharply bordered towards the violet, and shade

bandsoff

in

towards the red.

The same spectrum

is

produced by

Gernez 5 investivolatilizing tellurium in a hydrogen flame. gated the absorption-spectra of tellurium chloride and bromide; they consist of channelled spaces, the former in the green and orange, the latter chiefly in the red and yellow.Spark-spectrum6438.2:

1

661

SPECTRUM ANALYSIS.

2 Huggins, and Thalen, the ultra-violet region by Hartley and 8 4 The arc-spectrum has been measured Adeney, and Cornu. and Dewar, 6 and more recently by Kayser and by Liveing 6 Runge, who usually obtained it from the metal, but occa-

sionally used the chloride; they The limits 63O/f/u and 2io,/u/w.

photographed it between the numerous lines in the spark-

spectrum between 650/4^ and 3OOyw/* are almost all absent from the arc-spectrum. With the exception of the green line at 535W> an d a faint line at 553A