Journal of Research of the National Bureau of Standards Vol. 49, No. 2, August 1952 Research Paper 2345

Infrared Spectra of Noble Gases (12000 to 19000 A)1

Curtis J. Humphreys and Henry J. Kostkowski

The first spectra of helium, neon, argon, krypton, and xenon, excited by discharges inGeissler tubes, operated by direct connection to a transformer, have been explored in theinfrared (12000 to 19000 A). A high-resolution, automatically recording, infrared spectrom-eter, employing a 15000-lines-per-inch grating and lead-sulfide photoconducting detector,was used as the dispersing instrument. A new set of wavelength values is reported for allthese spectra. New data include 18 previously unreported lines of neon and 36 of krypton,all of which have been classified. The descriptions of the spectra of argon, krypton, andxenon represent essentially a repetition of the observations of Sittner and Peck. Severalpreviously missing classifications are supplied, also a few amended interpretations. Theanalysis-of these spectra may be regarded as complete. Use of selected lines as wavelengthstandards is suggested..

1. Introduction

The essentially complete character of both thedescription and interpretation of the photographedspectra of the noble atmospheric gases makes itapparent that any reopening of the subject can bejustified only on the basis of the availability of newsources of information, such as a new technique ofobservation permitting an extension of the observa-tions into a previously unexplored region, leading tosignificant additions to the experimental material.Such a technique is the utilization of lead-sulfidephotoconducting detectors in combination with high-resolution gratings for radiometric observation. Thelead-sulfide cell extends the range of such high-resolution observations beyond the photographiclimit to the limit of its sensitivity near 30000 A, andbecause it is also sensitive to the visible and ultra-violet as far as 3500 A, at least, use of higher orderstandard lines for comparison is possible.

Three of these spectra, argon, krypton, and xenon,have been observed by this radiometric technique bySittner and Peck, whose reported observations andanalysis [1]2 cover essentially the same region asthose herein presented. These excellent observa-tions appear to have been essentially complete, andthe overwhelming majority of the classifications arecorrect. In the intervening period, however, suf-ficient new information has been accumulated tomake it appear justifiable to prepare a new publica-tion, which should complete these analyses as far asany reasonable effort will permit. Discussion of thespecific points of difference between this analysis andthat of Sittner and Peck, together with extensions tothe analyses as reported in other earlier publications,will foe included i n the, separate .sections, dealing, with,the respective spectra. In brief, these consist ofinclusion of He i and Ne i, presentation of pre-viously unreported data consisting mainly of 18 newlines in Ne i and 36 in Kr i, interpretation of nearlyall reproducible previously unclassified lines, andamended classifications in a few instances.

Observations of Kr i and A i, obtained with a prismspectrometer, equipped with a glass prism and ther-

1 Presented, in part, at the meeting of the Optical Society of America, Buffalo,N. Y., Oct. 1949.2 Figures in brackets indicate the literature references at the end of this paper.

mocouple detector were reported by Humphreys andPlyler [2]. These observations covered the samespectral region in which the data herein reported wereobtained, but, because of well-known limitations af-fecting the precision of spectral data obtained byprism spectrometers with thermal detectors, may beconsidered as entirely superseded by the presentwork. The earlier paper may be referred to for afairly extensive list of references on the first spectraof the noble gases, which will not be repeated in fullhere. It also reported two new levels, designated 4Uand 4W, in Kr i, computed from photographic data byMeggers [3] but requiring confirmation by radiomet-ric observation of infrared lines, arising from combi-nations of the same levels. The first, or neutral-atom, spectra of all the noble gases are very rich ininfrared lines. Considerable portions of these infra-red spectra lie in the photographically accessibleregion, and have been the subject of exhaustive in-vestigations. Eeference is made to a few of the morerecent publications, which, in addition to those al-ready mentioned, will serve as a background andintroduction to the current work, and provide crossreferences to earlier work as required. A paper en-titled "The Infrared Spectra of Neon, Argon, andKrypton", published by Meggers and Humphreys [4]in 1933 brought the analysis of these spectra essen-tially to completion. A separate paper by Hum-phreys and Meggers [5], and of similar scope, appear-ing a few months earlier, brought Xe'i up to date.Shortly after this, plates incorporating new photo-sensitizing dyes, made available by the EastmanKodak Co. [6], permitted photographic observation asfar as 13000 A in favorable instances. With thesenew plates Meggers [3] reobseryed all the noble gasspectra to the photographic limit and interpretednearly all the new lines. In the intervening years nofurther extension of the range of photographic sensi-tivity has been accomplished, and no further observa-tions of noble-gas spectra were made up to the time ofthe radiometric investigations at the National Bureauof Standards [2] and at Northwestern University [1].

2, Energy Levels of the Noble GasesAlthough the analysis of the spectra of the noble

gases may be regarded as essentially complete, in thesense that nearly all the levels predictable from the

73

electron configurations have been found, and that,in most instances, long series have permitted highlyprecise calculation of absolute term values, thesespectra are somewhat unusual in their structure inthat they do not permit arrangement into regularmultiplets with any reasonable conformity to rulesregarding intervals or intensities. Paschen [7]adopted a special notation in reporting his analysisof Ne i. This notation has been retained in allsubsequent publications on these spectra; evidentlybecause it is not possible to describe the levels accord-ing to the currently accepted notation, which isactually meaningful only when vector coupling ofLS-type is present. Thisanability to identify mul-tiplets in rare-gas spectra points to the probableexistence of a different type of coupling. This hasgenerally been supposed to resemble the i/-case butevidence, principally from Zeeman effect, indicatesthat extreme ^'-coupling is not realized, and that anintermediate type prevails. Considerable light wasshed on the problem by a theoretical paper by Racah[8], who discussed an intermediate coupling scheme,designated jl, and discussed the conditions for itsexistence. It was pointed out that, whereas fe-coupling occurs when the spin-orbit interaction isweak compared to the electrostatic, and ^'-couplingoccurs when the electrostatic interaction is weaker, athird possibility, the ^7-case, may be realized,according to which the electrostatic interaction isweak compared to the spin-orbit interaction of theparent ion, but is strong compared to the spin cou-pling of the external electron. The vectorial repre-sentation of this case is to combine the total angularmoment j of the parent ion with the orbital momentI of the external electron to form a resultant K,known as the intermediate quantum number. Final-ly, K is combined with the spin of this electron toobtain the resultant J , which has the usual signifi-cance, namely, total angular moment of the resultantconfiguration.

Racah suggested a special notation for representinglevels originating under the condition of jt-coupling.This notation has been selected for the appropriatesections, pertaining to noble-gas spectra, of "AtomicEnergy Levels" [10]. This innovation has been fol-lowed in tabulating the descriptions of spectra listedin this paper, except for He i, which is described inconventional notation. The Racah notation hasbeen abbreviated by omitting the description of theparent ion, mp5(2P°iyi) or (2P^), the omission orinsertion of the prime with the letter indicating therunning electron being sufficient to distinguishbetween the two respective possible cases. Thisusage is also borrowed from [10], which not onlyemployes the prime in the manner indicated, butalso supplies the complete parent ion description.Descriptions of the transitions, according to thePaschen notation, are also included along with theRacah notation in parallel columns in the tablespertaining to argon, krypton, and xenon. This isintended to permit ready comparison with earlierpublished analyses and to provide a basis for trans-lating the old notation into the new. It is to benoted that every symbol in the new notation is phys-

ically significant. The number within the bracket isRacah's KOY intermediate quantum number obtainedby vectorial addition of the j-value of the parent ionto the I-value of the external electron. In theinstance of the noble-gas configurations, where theparent ion has j-vahie=l% (unprimed case), therecan be a maximum of four K-values; and where theparent ion has j-value^OK (primed case), no morethan two values of K appear. For each K-value, byaddition or subtraction of the spin moment S of theexternal electron, always=0%, two possible j-valuesappear for the resultant vector sum. We thus obtaina pair of levels for each K. A maximum total oftwelve levels is possible for any rare-gas configurationwhere the external electron has /-value 2 or greater.As might be expected, this is the same total numberand the same set of ^'-values that would be obtainedwith Z£-coupling. Table 1 illustrates the develop-ment of the set of levels and appropriate quantumnumbers associated with the binding of the/-electronin the configuration mp5 nf.

TABLE 1. Development of notation for jl coupling, accordingto Racah, illustrated by f-type levels of noble gases

Ionconfiguration

p5(2PO)

p5(2P°)

U

oy2

+1 = 3

+1=3

K

WA]

[3V2]

m)

\\y2\

[3^]

[2V2]

+s=oy2

+s=oy2

J

{ i{ $( 3

I 2r 21 i

{ . J/ 31 2

The pair structure resulting from the ^7-couplin^scheme was pointed out by Shortley and Fried [9].This structure is quite apparent in the level schemesof the four noble-gas-atomic spectra, Ne i, A i, Kr i,and Xe i, up to and including the levels based on theconfigurations p5 nd. In the instance of the pairarrays from pl nj these features are much less ob-vious. They are developed in three out of fourpossible cases in Xe i, and show diminishing sepa-rations for the gases of smaller atomic number,until in Ne i these /-type pairs of levels merge intosingle levels even with the high resolving power nowemployed in presently available techniques. Lackof knowledge regarding these /-type levels has con-stituted the most conspicuous gap in the analysisof these spectra, and a considerable part of the in-formation supplied by this investigation is con-cerned with these structures. Details for therespective spectra are given under the appropriateheadings.

The numerical values of the levels used in obtain-ing the calculated wave numbers of emission linesthat appear in the tables that follow, pertaining toA i, Kr i, and Xe i are those that are currently

74

appearing in [10]. The levels reported in that pub-lication for Ne i, A i, Kr i and Xe i, are taken fromunpublished manuscript by Edlen. These valuesrepresent a revision, together with reassignmentsin a few instances, of the tables of levels listed intwo publications already quoted [4, 5], and are basedon the same data. A small number of missinglevels, required to complete the arrays and predictedfrom series calculations, have been confirmed byobserved transitions.

3. ExperimentsThe high-resolution grating spectrometer used for

these observations has been described briefly in anearlier publication [11]. It is of conventional designincorporating a 15000-lines-per-inch Johns Hopkinsgrating, 1-meter focus off-axis paraboloidal collimat-ing mirror, ground and figured in the Optical Instru-ment Shop of the National Bureau of Standards.The cone-bearing, grating-mounting, worm-gear as-sembly, and simultaneously movable bilateral slitswere constructed in the Instrument Shops of theDepartment of Physics, University of Michigan. Theslits conformed essentially to the design of Roemerand Oetjen [12]. Figure 1 shows the instrumentwith the cover and baffles removed. The amplifiershown on the table beneath the spectrometer wasconstructed by W. R. Wilson, and is of the samedesign as that used for similar purposes at North-western University [13]. The electronic componentsvisible in the picture also serve to establish the scaleof relative sizes of parts. It may be noted that themirrors are 7 inches in diameter, and that the grat-ing is ruled on a 9-inch blank, with segments cut offto adapt it to the ruling machine.

The data reported were obtained in most instancesby using Geissler tubes as radiation sources. In in-stances where higher-order comparison standardsfrom the same spectrum were employed, the sourcewas simply imaged on the entrance slit by means ofa quartz lens. Where lines of a different spectrumwere used as standards an arrangement similar totha t i l l u s t r a t e d in t h e p a p e r b y t h e s e n i o r a u t h o r o nCa I [14] was employed. In this sys tem the quartzlens was left in position and a concave mirror wasalso set up on the opt ic axis in such a position thatit formed an image of the source in the position ofthe conjugate focus of the lens with respect to the

FIGURE 1. National Bureau of Standards infrared gratingspectrometer with the cover and baffles removed.

slit. When the comparison source included a tubeof fairly large bore, it was set up with its axis coin-cident with the image of the first source, so that bothcould be imaged on the slit simultaneously. Inother cases, the comparison source was moved intoor out of the described position, but the comparisonlines were always introduced under the condition ofcontinuous scanning.

4. Characteristic Features ofNoble Gas Spectra

4.1. Helium

The observation of the helium spectrum was fothe purpose of improving the wavelength data, be-cause there was no reasonable prospect of findingnew level combinations of appreciable intensity inthis thoroughly analyzed spectrum. The results aregiven in (able 2 and comprise data on six lines. Thefirst two lines listed, 3 3 D-5 3 F° , and 'S1D-51F°were the lines of greatest wavelength reported byMeggers [3] and were evidently at the limit of photo-graphic sensitivity. The wavelengths of the othersare known only from the early radiometric measure-ments of Paschen [15]. Because of a radiometricapplication utilizing some of these lines, special can-was taken to evaluate the relative intensities pre-c i s e l y . It is b e l i e v e d t h a i t h e a p p a r e n t r e l a t i v e i n -

TABLB 2. Description of I To i in the infrared region

( MISCI ved

wavelengi hin air

127X1. 701271)0. 27I 700)]. 1 118685. 12L8697. 0020580. «)

()bservedintensity

101

207010

5000

Term combination

: w : t i ) r,f>v°:w >i) 5 / 'K°

a3p »P°-4d»D3d :tI) - 4 / 8 P °3d M) - 1 / 'F°2.s IS 2/;1I)O

W a v e number

Observed

cm l

78 H). 667816. 315879. 675350. 395346. 994857. 55

Calculated

cm '7819. 897S15. 095879. 735350. 715348. 024857. 15

Difference(calc—obs.)

0. 231. 220. 06

. 321. 030. 10

Triple level unresolved in these experiments.75

tensities are correctly evaluated, bearing in mindthat no correction has been made for spectral sen-sitivity of the detector, or effect of the properties ofthe grating, on apparent spectral distribution ofenergy. The calculated values of wave numbersare based on the values of the levels quoted in [10].The wavelengths of the various lines of the funda-mental series have been evaluated by bracketingthem between fairly close third-order neon lines.There is no possibility of an experimental errornearly as large as the difference between observedand calculated wave numbers indicated for thesinglets, particularly since the relative precision ofthe determination of the singlet and triplet transi-tions for a given series member is high. It is sug-gested, therefore, that reevaluation of the first twomembers of the /-series is in order. The followingvalues would bring the levels into agreement withthe reported wavelength measurements:

4/ T 0 , 191446.81

4/3F°, 191446.29

5 / ^ ° , 193915.53

5/3F°, 193915.56

The indicated extremely close proximity of thesinglet and triplet F-terms of given order number isin accord with their relative positions for the higherseries members. The difference between the ob-served and calculated wave numbers for 2s *S —2p T 0 , namely, 0.10 cm ~\ is close to the expectedexperimental error. The line at 17003 A arises froma transition between multiple levels, unresolved byavailable techniques. No revision of accepted levelvalues based on the measured wavelengths of thesetwo lines appears to be required.

4.2. Neon

The observations on Ne i have been confined tothe limited wavelength region, 18035 to 18625 A,in which the electron transitions of the type 3d—4/occur. The only previous data pertaining • to thisregion consist of a few lines reported by Hardy [16]that could not be resolved or measured by methodsthen available with sufficient precision to improvethe existing set of level values. The new data arelisted in table 3 and comprise 18 lines. A reproductionof a typical record is shown in figure 2. The wave-lengths have been evaluated by interpolation be-tween third-order neon lines included in the well-known group of red neon lines, which are acceptedinternational standards. The calculated wave num-bers are based on a set of adjusted values of/-levelsdetermined from these observations rather than uponthe values currently published [10]. These revised/-levels are compared with the current set in table 4.The bracketed values quoted are, of course, predic-tions, which are now confirmed by the present setof observations.

I82ZI29 (O-fi)

1•S

•8

00

©

^ - 2 9 0 W 9 (0-£)

8.

1

P?

76

Observed wave-length in air

A18035. 4918082. 7118220. 7618226. 5718276. 59

18282. 5818304. 0018359. 2118385. 1718390. 10

18403. 1618422. 4318458. 5818475. 7918591. 12

18597. 3018618. 6918624. 94

TABLE 3.

Observedintensity

22130

1511

260

200140

6160180

65110

103

25

1201522

Newly observed and

Term combination

3d3d3d3d3d

3d3d3d3d'3d'

3d3d'3d3d'3d

0^]§-4/

gi^jl — 4j3^]o__4j33^11-4/

3y2] §-4/

1-4/1-4/5-4/'

23^ j — 4/ '

lj^]l»21J43y2"

±y2

1>2

3)4

3,4

4,5

4^]4,5

13^]l>23J^]3,43^13,4

1J^]| —4/ [23^]2,31 y£\i 4/' [2 3^2 3

\X^\\ 4/'[23^2 32^12-4 / [33^]3,'4

3d [2J^]§-4/[33^]3,43d [2J^]2—4/ [23^2,33d[2y2]i-4f[2y2h,3

classified lines of Ne 1

Wave i

Observed

cmr1

5543. 115528. 635486. 745485. 005469. 98

5468. 195461. 795445. 375437. 685436. 22

5432. 365426. 685416. 055411.015377. 44

5375. 655369. 485367. 67

lumber

Calculated

cm~l

5543. 115528. 575486. 755484. 985469. 97

5468. 205461. 815445.485437. 715436. 19

5432. 455426. 685416. 125410. 985377. 44

5375. 655369. 415367. 62

Difference(calc —obs.)

0.00- . 0 6

.01- . 0 2- . 0 1

.01

.02

. 11

.03- . 0 3

.09

.00

.07- . 0 3

.00

.00- . 0 7- . 0 5

TABLE 4. Revision of f-type levels in Ne 1

% Proposedvalues

167054. 70167062. 28167071. 03167079. 06167848. 33167848. 62

AtomicEnergy

Levels [10]

167054. 59[167062. 5]167071. 08

[167079. 1]

167848. 67

Racah notation

2pH2PiH)4/4/4/

2/>5(2P§M)4/'4/'

\y2

23J3y23y2

1,24,52,33i43.42,3

It is to be noted that a total of only six of these/-levels have been found, two primed and four un-primed. This represents the complete developmentof the level system, but with the Racah pairs eithercoalesced or so close together that they cannot bedistinguished in cases where there is a possible com-bination of both members of a pair with a givend-level, producing a close doublet. Such close levelsmight also be distinguished on the basis of a smalldifference in numerical value based on determina-tions from strictly single transitions according toselection rules, but here again the attainable pre-cision appears inadequate for the /-levels of Ne 1.

4.3. Argon

The observations on A1 extend from approxi-mately 12000 to 17000 A, covering essentially theregion explored by Sittner and Peck [1], but slightlyless extended in the directions of both shorter andlonger wavelengths: There appeared to be no ad-vantage in overlapping the region observed photo-

graphically by Meggers [3] below about 12000 Awhere the photographic emulsions were sufficientlysensitive to permit essentially complete recordingof the spectrum. The principal reason for reobserv-ing A 1 was that a considerable number of the linesbeginning with 13273 A and lying in the region ofgreater wavelengths were either left unclassified bySittner and Peck or had been assigned to transitionsthat appeared improbable. Part of Sittner andPeck's observations were made with flash tubes.The differences in excitation are responsible for con-siderable variation in the intensities of some levelcombinations, when Geissler tube and flash tubeexcitations are compared. In general, the effect ofthe flash tube excitation is to increase the intensitiesof combinations of levels where one or both thelevels belong to the family converging to the higherlevel of the inverted doublet constituting the ionlimit, namely, 2P0>J. A few of the weak lines observedby Sittner and Peck do not appear on our records and,vice versa. We have observed a small number notincluded in their list. These few weak lines remainunclassified, and there may be some doubt of theirorigin in A 1. Table 5 contains the pertinent infor-mation regarding 66 lines included in the currentset of observations. All but three of these lines areclassified. The wave numbers observed by Sittnerand Peck and by Meggers for the overlapping regionare listed for comparison. A slight majority of theobserved wave numbers show better agreement withthe calculated values than those of Sittner and Peck,but the precision of observation is not regarded asbeing significantly improved. The observations onargon were not expected to give precise evaluationsof intensities. The same remark also applies to thesections on krypton and xenon. Because of the

77

jljlis00

HC

ON

00

00

HcO

WO

NO

O

OO

SC

OiO

iO»

O C

O C

O O

S O

S

C^

WN

OC

OC

OC

OO

O

OO

OO

0000000000

0

00

00

0X

T-H *O CO

CO CO

OS OS

CO rJH l

>O

S CO

rf< CO 0

0lO

OS

CO

r^C

OC

OH

OC

ON

OlO

H C

OO

1O

H O

SiO

CO

CO

CO

CO

1>I>

CO lO

CO

*O

X•H

OS

£*• ^f CO

cocTF C

rH

OC

OC

OO

SC

DC

O

OS iO

*O CO

*O

N »O

CO »O

(NiH

OO

NC

O

lOiO

MH

N

l> CO

CO *O

lO

iOid

iOi

OS O

S I>

iO CO

T-H

iO rH

X X

l>

O OS

X l>

CO

CO CO

OS r-I

^ "t 1 CO

CO r-• Tt<

Ol

MO

OO

00

CO iO

CO O

X

CO CO

X X

O

S i-i H

rf CO

CO

O O

00

CO T

Ht

^t

n^

tH

C

OC

OC

OC

OC

O

NN

(N

HH

O

O O

00

00

X X

X 1

> t>»

CO-

- - - -

i>-1

*» l>

co

co

co

co

co

co

co c

o

o103

CO CO

T

P CO

TJH CO

r^ CO

iO iO

CO

Th i

I II

II

II I

I I

MM

I I

I

OT

H

OC

O

OCO] O

<N CO

O

CO ^

OH

OCO

O

CN O

O

OCO) O

CM

OCM

tO C

O iO

CO

CO

C

O u

§ C

O C

O C

O

TJ?0O <rt?iO C

O

iO tO

CO

CO

CO

C

OlO

TJH'CO T*<" . .

II II I

M

il

l II II I II II I II II I II II I I I

I II

-- -

- -

--

- -

».

oe* o

«

^^

^

•^f

"^

CO

"^

CO

a? o

CO

T

H

CO

CO

I M II

II I I

I II I

IC

0<

N<

N C

O

MM

__

__

_ _

..

_.

..

_.

. _

O

DO

O'O

D^

OD

'TS

O3

l—|

h.t>

'^

S'-C

K^

OS

IK'

^J

K^

^C

K-

'C

CC

OC

OC

OC

OC

O

CO

CO

CO

CO

CO

T

^C

OT

tiCO

CO

C

OC

OC

OC

OC

O

CO

CO

^C

OT

tH

CO

CO

^T

^T

}H ^

T

H C

OT

^C

O

I II

II

Mill Mill

II

II I

Mill

Mill I I

II I

(N CO

T*

2

CO rt<

2?a, ?a, ja, sa, ja,

<N (M

(N <M

<Nsa, ^

^s s

a,t

CO

CO

CO

CO

CO

C

OC

OC

OC

OC

1C

0C

0C

O

CO

CO

CO

CO

CO

00

eC

ON

O®

00

T

-I CO CO

GO

MC

NO

iOiO

CO <N

OiO

i<

N<

N(N

OO

00

X0

0 00

00

O^

CO

NO

00 GO

GO 0

0 00

oi io

io<N

i X

I

s-s-aC

KN

WO

No

cq

cc^o

oTtH

iO(N

CO

XTtn

OiO

XC

ON

OO

IHX

CO

O<

NC

O T—

I O5 lO

TH

CS O

NC

ON

O1

00

H<

NO

iOi(N

Hrt

t<

N ^

CO

N^

ON

OO

C0

(N

CO

TJH

O5

X

CO

»O

C5

HO

5 N

©

OO

M

OO

TH

CO

OCO

rH X

I>C

O!>

!>• CO

CO iO

as as

co io

coos i>

to o

oO

f rH X

Xooco

ioc

oo

CO CO

CO CO

CO

CO iO

CO

OS

T-I

OS

rH

T—I TfH

CO

oooxx

!>!>

!> CO

CO

CO

CO

CO O

^C

O

O

X

CO T

tX

X

I>I>

CO

CO

CO

C

O C

O C

O

W

M

NH

CO

CO

00

CO

rHO

CO

CO

C

OX

T-H

CO X

rH TH CO

rH CO

CO

XC

Oi>

OS IO

coocoL

co to

i> os

og

lO

CO

CO

OS

OS

«

CO

CO

CO

OO

X X

X X

X

O

OS CO

CO iO

CO

CO O

OS CO

CO

^C

OC

OO

C

OC

OC

OO

OO

OO

OO

O

so

sO

sc

XX

XX

X '

'"

"

OT

-HC

O

XX

X

TjJ CO OS

rH iO

XO

SrH

IO

CO

I>C

O O

OiN

coN

OM

CO

XO

S!>

T-I

IOC

OC

OT

HC

OT}H

CO W

OO

Nl>

l>

O

ON

O

CO

XrH

CO

CO

O

SiO

^O

CO

iOT|H

CO rH

XI>

CO

i>

i>|>

CD CO

iO!>•

|>iO

C0

lOO

S^

OM

OO

SX

CO

UD

CO

C5M

OO

O^

iO T-i GO*

X0

0 CO iO

CO

O"^ CO

CO CO

CO

NO

OiO

OC

Oxcocoxx

CO

CO

C

O r

I

CO

»O

T-H O

S

T—I

OS

T

-H rH

TtH

CO

oooxx

I>N

I>C

OC

O

o.9

8S

2g§

§§§§§

«

«•«

Ǥ

§§

§CO

rH

rH rH

ooocoo

OC

O iO

O

CO ^

nl>

CO

JiO

iOi

DC

O

iO»

OiO

CrH

CO IO

CO

OO

CO

O

O1

>O

OO

O iO

CO

CO

OO

CO

CO

T-H

IO

CO

-O

TJ

HX

«O

J

>.

c 23

oost>coco

C0

IO

tX

xo

so

sio

co

XC

O

T-H OC

OT-I

COiOCO

CO

CO

C0

0S

l>l>

O

OS

l>»O

I>O

SC

OiO

tiCO

I>

^H

CO

OC

OC

OC

OX

IJ

OC

OC

O^

OC

OO

SlO

OX

XC

O

t^O

SO

SrH

CO

IC

OC

0O

N

OS

lt

HX

T-H CO CD l> rH

CO O

S'* CO rH

t. OS

OS CO

rH OS CO

rH CO T^

—I CO OS rHCOCO

N T-H C

O iC Th IO

T-H T-H

T-H CO CO

CO CO CO CO CO

CO

OS

OS

CD

t^ TIH

CO

T-HO

ST-H

CO C

O C

D CO

CO

WO

OO

OO

^ X

T-H

CO

CO

CO

O

N-C

D

OS

CO

C

0C

0O

5C

0X

(X

HO

CO

NO

O

T-H CO

tO X

iO

OS

CO CO

iO

OW

^O

H

CO tO

O CO

T-H

C0

C0

1>

OT

-H CO

CO O

OS O

rJH

|>

OS

CO l

>

T-H CO

CO O

T-HTJH T

H TJH T

ji rJH

-OIO

CO

CO

CO

NN

NO

0O

5

OS

OS O

O CO

C

OC

OC

OC

OC

O

CO CO

iO

»O tO

i__ C

D C

D

N X

X O

S O

SC

OC

OC

OC

OC

O

CO

CO

CO

CO

CO

C

OC

OC

OC

OC

O

CO

CO

CO

CO

CO

C

OC

OC

OC

OC

O

CO

CO

CO

CO

CO

C

OC

OC

OC

OC

O

CO

CO

CO

CO

CO

C0

05

_>

CO

T*

H

<J5

CO

OS

"*

tO O

S CO

"^

O

SO

CO

CO

tOc

D

CD

C

O

TABLE 5. Description of A i in the infrared region— Continued

Observedwavelength

in air,Humphreys

andKostkowski

A15172. 3315229. 9215302. 2615329. 5615349. 52

15353. 5115402. 5815899. 9315989. 3416436. 92

16520, 1416549. 8116739. 8416940. 39

Observedintensity

221

755

10

210202018

965

100

Wave number observed

Humphreysand

Kostkowski

cm~l

6589. 136564. 216533. 186521. 546513. 06

6511.376490. 636287. 606252. 446082. 18

6051. 556040. 705972. 125901. 42

SittnerandPeck

cm~l

6588. 82

6533. 626521. 736513. 24

6511.516490. 846288. 426252. 326082. 33

6052. 556041. 065971. 865901. 59

Meggers

cm~l

Term combination

Paschennotation

2p5 - 2 s 4

3d; - 4 U2p8 — 3d33d; - 4 Y

2pA —2si3d; - 4 V3s; - 4 Z2pi - 2 s 23d2 - 4 Y

2p7 — 3d3

3d2 - 4 X2p2 — 2s52p6 — 3d3

Racah notation

4p [ 0 ^ ] 0 - 5 s [1V2]1

3d [2H»-4 f [33^]M

4?> [2^ ] 2 -3d [iy2y2

3d [ 2 ^ - 4 / pjfla4p;[l^]i-5s [ljflf3d [2J4B-4/ UV2),Sd'[iy2]l-4f'[2y2}24p'[0^]0-5s'[0^]t3d [lJfll-4/ [2^]2

4?> [iy2]i-zd [iy2y2

3d [1HB-4/ [l^]i,24p'[o^]i-5s [iy2y2

4p[lH]2-3d[lHB

Wavenumbercalcu-lated,Moore

[10]

6588. 94

6533. 54'6521. 666513. 20

6511. 516491. 166287. 686252. 406082. 32

6051. 68/6040. 59\6040. 915972. 095901. 38

essentially linear character of the dependence ofdetector response upon incident energy, these recordsdo, nevertheless, provide an excellent means ofintensity evaluation. Because of the limited scalerange of the recorder, it is necessary to make a largenumber of records with different amounts of energyincident on the entrance slit or with different slitwidths, or else to introduce controlled attenuationinto the amplifier output in order to reveal the rangeof intensities between the strongest and weakestlines. Both devices have been used to a limitedextent. The relative intensities may be consideredreasonably good estimates over moderate ranges, butare subject to the limitations mentioned in the sectionon helium. The nonuniformity in spectral distribu-tion associated with use of a grating with a pro-nounced "blaze" angle is probably the most import-ant of these limitations. It seems fairly certainthat the most intense line included among thoseoriginating in transitions of the classes studied is2p9—3^1. The newly classified lines originate inmost instances in combinations involving four newlydiscovered levels which are the first members of theold V, U, and W series previously known only inhigher members. In the Racah or pair-couplingnotation, these levels are 4/[4^]5, 4/[4^]4, 4/[3^]3>4,and 4//[33^]34. It is to be noted that in the lasttwo instances the separation of the pairs has notbeen directly observed, or inferred, from existingdata. These newly interpreted levels were actuallyidentified from the data in the Sittner and Peckpublication, previous to the observations hereinreported, and communicated to Dr. Moore for inclu-sion in the first volume of [10], as noted in the sectionon argon. The following are the wave numbers oflines for which no classification has been published

previously: 7673.51, 7532.00, 7515.39, 7456.98,7381.45, 7011.94, 6824.07, 6533.18, and 6490.63cm"1. New classifications are proposed for 7499.64,and 6831.47 cm"1. Each of the lines of wave number7229.85 and 7187.34 cm"1, as listed by Sittner andPeck, has a close resolved companion. The com-panion line in each instance should have the classi-fication listed by those authors, and new interpreta-tions are proposed for the lines in the positions firstreported.

4.4. Krypton

Considerably more effort has been devoted to theobservation and interpretation of Kr i than to anyother part of this study. The principal reason forthis emphasis is that, of these noble-gas spectra, Kr ihas the greatest population of fairly uniformly dis-tributed lines in the region between 10000 and 20000A, and, as such, shows considerable promise as asource of standard wavelengths. The wavelengthshave been determined by use of second-order krypton,neon, and argon lines as comparison standards, and insome instances third-order lines of Fe i excited in anarc. A few first-order krypton lines originating intransitions of the type Is—2p, the wavelengths ofwhich can be computed with great accuracy fromknown levels, were also used in regions where avail-able. Actually, the relative scarcity of good stand-ards was the principal impediment to the mostsatisfactory wavelength determinations. One of thesources used was a Geissler tube, imported severalyears ago from the firm of Robert Gotze in Leipzig,and kindly loaned for this work by Wm. F. Meggers.This tube contained krypton of exceptional purity,and its operation under optimum conditions probably

79

TABLE 6. Description of Kr i in the infrared region

Observedwavelength

in air,Humphreys

andKostkowski

Observedinten-sity Humphreys

andKostkowski

Wave number observed

SittnerandPeck

Meggers

Term combination

Paschen nota-tion Racah notation

Wave num-ber calcu-

lated, Edle*nlevels

A11792.2511819. 4311996. 0011997. 0812077. 42

12117.81*12123. 4712156. 9712204. 3912229. 23

12240. 8112321. 4812598. 1912782. 3912825. 08

*12861. 89

12879. 0012934. 4812977. 9812985. 08

13022. 0513177. 3813210. 5613240. 5213304. 30

13337. 5213622. 2813634. 2213658. 3813711.23

*13738. 8613763. 7213800. 0313832. 5713882. 64

13924. 0013939. 1313974. 1514104. 2714156. 62

14341. 2514347. 82

14402. 58

14426. 9314469. 3314715. 5514734. 4614762. 83

14765. 6414961. 7614973. 7415005. 5715209. 52

1202000

25480115

100402

7004

2915100

55500

1212

1585010755

558001700360100

4006350

240

27085704015

9400

cmr18477. 828458. 338333. 828333. 078277. 64

8250. 05

"8223." 488191. 528174. 89

8167. 158113. 687935. 477821. 11

7795. 10

7762. 467729. 157703. 247699. 04

7677. 187586. 707567. 637550. 517514. 31

7495. 597338. 917332. 487319. 517291. 30

8277. 96

8250. 098245. 82

8191. 13

8113. 83

"8721." 127794. 70

7772. 417762. 56

7698. 32

7676. 477586. 667566. 357549. 90

7494. 87

7263. 497244. 377227. 337201. 26

7179. 877172. 087154. 107088. 107061. 89

6970. 976967. 78

7154. 11

6968. 53

6941. 8980

1100302

900250

2301108

2542

6941. 29

6929. 576909. 276973. 666784. 956771. 90

6770. 626681. 866676. 526662. 366573. 02

6929. 976909. 74

6772. 03

6770. 396681. 56

6661. 746573. 35

cmr1

8477. 678458. 33

8333. 038277. 79

8250. 068246. 15

8191. 43

3d;1*23d;3d'4

2p8

2pi2pi2s5

2p*2p8

2

ls22p2

2pz

2p3

2p2

-2s5- 4 Y- 4 T- 4 Z

- 4 W

-4U-3s4

Sd[ - 4

| 3 d 4

1*3

2p2 -

1534.60-1536.493-4W-4T-4Y

-3d ; '

-1536. 493-2s4-4d;'-2s2

-2s 3-Sd2-2s 5-2s,-3s ;

p6-4di'

-2s2

2pZff -

2p3

2p2

4W2s34Y

2p2 — 4

See footnote at end of table.

2pz

[2V213d;

2p7

Sd[2pi2

p62p73d52p72p9

80

- 4 W

- 4 Y-4T

-2s4- 4 U

-3d |-3d;

-Sd2

-2sl

bpbp4d

i-4d

[l^B-4/[1^2-4/

• - • - " | - 4 /

y2hiy2h

. . _J-4/5s' [0^]!-\

4d [3HB-4/5pr[l3^]i —7s

i -7p[ l^ ] 2

4d

3-4/'1§-4/§-4/§-4/

§-4/|

3^]i

5s' i

4d

5p5p'4d

5p ^]n-6s. i»-4/

OJ4Io-5d

2-6s[l^]5

cmr1

8477. 718458. 368333. 848333. 088277. 80

8250. 048246. 168223. 518191. 438174. 78

8167. 318113. 987935. 757821. 117794. 587795. 34

7772. 787762. 707729. 307703. 287698. 91

7676. 977586. 657567. 587550. 417514. 31

7495. 427338. 847332. 477319. 497291. 40

7276. 647263. 317244. 357227. 187201. 16

7179. 977172. 197154. 207088. 107062. 08

6971. 316968. 006968. 166942. 146942. 236941. 47

6929. 646909. 596973. 386784. 996772.01

6770. 696681. 836676. 456662. 486573. 02

TABLE 6. Description of Kr i in the infrared region—Continued

Observedwavelength

in air,Humphreys

andKostkowski

Observe(inten-sity

Wave number observed

Humphreyand

Kostkowsk

SittnerandPeck

Meggers

Term combination

Paschen nota-tion Racah notation

Wave num-ber calcu-

lated, Edtenlevels

A15239. 8515326. 8715335. 2915371. 8915433. 63

*15474. 0215634.9815680. 9415771. 4415820. 10

15823. 4015890. 5215925. 6416052. 3116109. 46

16315. 5816347. 3116465. 2916573. 10* 16726. 48

16784. 6516853. 4516890. 4016896. 5816935. 71

16994. 3617070. 0417098. 7617230. 2117367. 98

17404. 6717616. 5717630. 4417770. 2117842. 70

18001. 7118098. 4618167. 1218184. 4318418. 82

18581. 1918695. 9118785. 4518787. 7318797. 59

900358503504

657

751

35

225623

125151670

9504801000700800

1010

30010

360

323744

270

40010

1500154

3062371040

cm~16559. 946522. 696519. 116503. 586477. 57

6394. 156375. 416338. 836319. 33

6318. 016291. 326277. 456227. 916205. 82

6127. 426115. 536071. 706032. 21

5956. 185931. 865918. 895916. 725903. 05

5882. 685856. 605846. 765802. 155756. 13

5743. 995674. 905670. 445625. 845602. 98

5553. 495523. 805502. 925497. 695427. 72

5380. 295347. 285321. 795321. 155318. 36

cm~16560. 056522. 536519. 196503. 09

6460. 846394. 206375. 59

6319. 40

6290. 53

"6227." 31

6127. 23

6032. 595976. 46

5956. 055931. 865918. 955916. 755903. 06

5846. 91

~5756.~36

5744. 155674. 22

5602. 97

5553. 345523. 505502. 935497. 57

5380. 455346. 765321. 61

5317. 88

- 3 d 2- 3 d 3- 2 s 5- 4 d 6

2s5 - 4 T3d2 - 4 Y2s5 - 4 Z3d2 - 4 Z

3d2

3ds3d5

- 4 X— 2s2

2p9 — 3d4

2ps — 3d4

2pio — 3d52p7 - 3 d ; '

2Vz - 4 d 5

3d5 —3pio2p2 - 3 s ; "

2p 6 — 3 d /3d3 — 3p 6

2p 4 —3s; '3ds — 3^72^io — 3d6

2p 5 — 3d2

P3 - 3 s ; " '2p 9 — 3d4

o ^ 3 s " "

3d3 — 3pg

2ps — 3d3

Is2 —2pio2??2 — 3 s "

bpbphp5p I5p'

5s6s4d6s [14d [1

4d

- 4 / [2y2}2

11-4/ [1^]2

; -4 / [ i ^ ] 2

2s43d6

2s42Pl

- 4 Y— 3pio— 4Z—3s;

6s4d6s5p

- 4 /

I!-4/

'[l^]2-4

: [03^]! -

p4d5p'4d|

4d

cm""16560. 046522. 886519. 276503. 536477. 56

6460. 686393. 996375. 406338. 716319. 36

6318. 246291. 266277. 426228. 286205. 88

6127. 596115. 696071. 556032. 245976. 90

5956. 055931. 885918. 905916. 695903. 03

5882. 655856. 635846. 745802. 005756. 27

5744. 085674. 845670. 405625. 705603. 00

5553. 365523. 515502. 955497. 495427. 87

5380. 405347. 175321. 815321. 155318. 30

*Calculated wavelength used as a comparison standard.

revealed all of the krypton spectrum that can be ex-cited in this type of source. The description andinterpretation of the observed krypton spectrum inthe region investigated is displayed in table 6, whichpresents the data on 98 krypton lines, all of which areclassified. Included in this number are 36 lines notreported by Sittner and Peck. About three-fourthsof these are weak lines scattered throughout the re-

gion, but nine lines between 13600 and 14000 A areamong the most intense in the infrared krypton spec-trum and may have been omitted inadvertently fromthe description by Sittner and Peck. A portion of arecord showing this group of lines is displayed in fig-ure 3. This was a survey record run with relativelywide slits in order to reveal as many of the faint linesas possible. The peaks of many of the intense lines

81

are cut off, because the deflections are off the scale ofthe recorder.

Edlen's revisions of the table of levels by Meggersand Humphreys [4] and listing of additional levels,are the basis of the entries on Kr i in [10]. Thesechanges are relatively few in number, emphasizingthe fact that this analysis has long been, regarded asessentially complete. The levels to which the desig-nations 2s2 and Ss[ were formerly assigned have •been interchanged. A similar switch has been madewith 3s" and 3s"". These changes should be takeninto account in comparing Sittner and Peck's classi-fications with ours. The Edlen manuscript also sup-plied a complete set of values of the levels represent-ing the first members of the various series originatingin the 4/ configuration, except that the doubletstructure of the level designated 4W or 4/[3 3^3,4 re-mained unresolved; a computed value only was pro-posed for one component of the 4U doublet, namely4/[43^]4; and only three of the four possible 4/-leveisconverging to the higher ion limit p5(2Po^) wereproposed. The experimental confirmation of the 4Uand 4W by Humphreys and Plyler [2], has beenmentioned. The new compilation of levels of Kr ialso includes the previously missing 2s3. Edlenlisted two choices, but it was noted that one of Sittnerand Peck's intense unclassified lines, v=7494.87cm"1, could be interpreted as 2p4—2s3, if the levelof absolute value 7823.56 were adopted as 2sz. Fur-ther confirmation is found in the classification ofthe moderately intense line, ^=7172.08 cm"1, as2p3-2s3. Two of the remaining intense lines reportedwithout classification by Sittner and Peck are as-signed to transitions involving 4/-levels of the familyconverging to the hig;her limit. These are v, 8113.83cm"1, classified in pair-coupling notation 4d'[lJ^]2—4/'[23^]3, and v, 7676.47 cm"1, classified 4d'[2%]°3-4/'[3 3^3,4. Sittner and Peck listed two other intensekrypton lines without classification, v, 7578.37 cm"1,which we have been unable to reproduce with oursources, and v, 5982.33 cm"1, which seems almostcertainly to be the third order of the intense visibleline, X, 5570.29 A. Of the two values of the pair4U, or 4/[4j^]4>5 proposed by Edlen, the one withe/-value 4 was bracketed indicating a predicted value.The pair separation is given as 0.20 cm"1. The onlypossibility of a combination of both levels of thepair with a common level is that involving the level3 4 or 4d[3}4]l yielding the line, v, 8191.52 cm"1.No structure has been observed, but a resolvingpower of over 40000 would be required to split thisdoublet, something beyond what we have obtainedin any clearly demonstrable case. It is probablethat most of the energy in this line is accounted forby the transition 4d[3 J^]4—4/[43^]5. All other com-binations of this 4U pair involve the level J = 4 ,uniquely owing to the A«7 selection rule.

Evidence, principally from the measured wave-length of the intense line representing the combina-tion, 4d[33^]3—4/[4H]4, for which we give XI2879.00,indicates that Edlen's estimated absolute value for4/[4H]4, namely, 6925.92 cm"1, is too small. Theobserved value, obtained by averaging our measuredresult with that of Sittner and Peck, is 6926.10,

82

making the level indistinguishable from / [ ^ ] 5As represented in [10], with the ground level=0,this level would become 105989.60 cm"1. TheHacah pair of closest separation that we have actuallyobserved in combinations with a common level is4d[l 3^2—4/[23^]2,3, represented by the wave numbers8333.82, and 8333.07 cm"1. The observed separa-tion is 0.75 cm"1 as compared with the calculated,0.768. The actual observation of this pair isextremely difficult on account of the great disparityin intensities of the components. The pair ofwidest separation in this 4/ group, 4X and 4Z or4/[l3^]i,2 occurs in easily observable combinationswith a common level. The best example is thepair of wave numbers 8879.23 and 8880.35 reportedby Meggers and Humphreys [3,4] and also demon-strable by radiometric techniques. So far there isno clue to the magnitude of the splitting of 4W, andthe missing level 4//[33^4 has not been found unlessit is unresolved with respect to 4//[33^]s. Otherwise,the level structure of Kr i may be considered•complete.

4.5. Xenon

The observations on xenon were less extensivethan on any of the other four noble gases becauseboth preliminary recordings and study of the levelscheme had revealed that there was little possibilityof adding significantly to the existing observationalmaterial. A small number of records were madein order that some material on all the noble gasspectra might be included in the current survey.

The results are presented in table 7, which is organ-ized in the same fashion as the corresponding tables5 and 6 for argon and krypton. These observationscover the same region as those of Sittner and Peck,and their classifications are repeated without revisionfor lines common to both lists.

The principal reason for the relative paucity ofxenon lines in the 1.2— to 1.8—micron region is thatthe 2p and 3d levels fall in much the same range ofnumerical values, with the result that the 2p—3dcombinations, that are responsible for many of theintense analogous lines in argon and krypton,represent wavelengths much deeper in the infraredand out of range of observation with lead-sulfidedetectors. For instance, the position of the expectedmost intense combination in this group, 2ps—3d'4,is predicted at 1793.6 cm"1. Most of the possible3d—4/ combinations have been observed photograph-cally. Of the/-type levels, only those of the familyconverging to the 2Pf ion limit have been found.Owing to the extremely wide separation of the serieslimits, the first members of the /-series convergingto 2P§i are above the first ionization limit.

The only noteworthy changes in the Xe i levelscheme made by Edlen since the.last publication ofthis array [5] have been to change the designationof the level at 4215.65 from 2s2 to 3s[, to supply levelvalues for 2s2 and 2s3, and to separate the seriesformerly designated W, into two according to theRacah scheme. The first pair of this series 4/[3 3^3 4has a separation of 0.10 cm"1 on the basis of pre-viously existing photographic data. Discrimination

TABLE 7. Description of Xe I in the infrared region

Observedwave length

in air,Humphreys

andKostkowski

A11742. 0111793. 0411857. 0011911.4411952. 57

12084. 8212235. 1412258. 1012451. 2112590. 00

12623. 3213543. 1613656. 4814142. 0914241. 39

14364. 9014659. 8414732. 3815418. 0116052. 02

16727. 52

Observ-ed

intensity

9040303

10

2080

62

26

3005

1508040

205

20011050

50

Wave number observed

Humph-reys and

Kostkowski

8514. 098477. 258431. 528392. 998364. 10

8272. 588170. 938155. 638029. 147940. 63

7919. 677381. 777320. 527069. 157019. 85

6959. 506819. 426785. 906484. 136228. 03

5976. 52

Sittnerand Peck

cni-i

8514. 778476. 848431. 478392. 508363. 26

8272. 598170. 888155. 858028. 327940. 70

7919. 30

7320. 197068. 957020. 18

6959. 466819. 026785. 706484. 166227. 44

5976. 05

Meggers

C7Yl\-8513. 928476. 898431. 318392. 538363. 81

8272. 608170. 888155. 83

7919. 63

Term combination

Paschennotation

3d[ - 4 W3d[ - 4 U3d[ - 4 Z3d 2

3d'4 -3p8

22

p i o 2 53d4 — 3p92p9 — 2s42p9 - 2 s 53d2 - 4 V

3d2 - 4 Y3d['-3p7

2ps - 2 s 5

Racah notation

-4/ [3y2

-4/ [2K-4/ [±y6

5d

6p6

li-7«

-7s- 4 /

1J4B-4/

\V2

OH.

1M2y2IK'XK2K

-7p2^]3-7s"1H1I-7«lHh-78

-7« [IX*

Wave num-ber calcu-

lated, Edtenlevels

8513. 838476. 898431. 318392. 498363. 74

8272. 578170. 888155. 858028. 927940. 49

7919. 667381. 277320. 237069. 017020. 11

6959. 506819. 046785. 756483. 996227. 56

5976. 34

83

between levels this close is beyond the precision of thecurrent, observations.

The complete status of the level scheme of Xe imakes it unnecessary to consider further observationsin order to improve the analysis. Future availabil-ity of photoconductors sensitive to energy of greaterwavelengths, possibly as far as 7 microns would stillmake Xe i an attractive subject for study owing tothe existence of several predicted lines of probablehigh intensity that might be useful as wavelengthstandards.

5. ConclusionObservation and interpretation of the noble gas

spectra over a period of over thirty years have led tothe essentially complete analysis of the spectra ofthese gases. One of the most interesting features ofthis work is that it has been made possible largelyby the development of techniques for observing inthe infrared region by either photographic or radio-metric methods. The noble gases have providedsome of the best sources of standard wavelengths.The possibility of extending the range of highlyprecise observations farther into the infrared regionby combining interferometric with radiometric meth-ods still constitutes an attractive challenge.

Several members of the staff of the Bureau's Radi-ometry Laboratory, have given valuable assistance inbringing this work to completion. In particular,Arthur J. Cussen and Joseph J. Ball contributed to

the operation and maintenance of the electronic in-struments used with the photoconductive detector.Marshall E. Anderson assisted in obtaining andreducing the records. It is a pleasure to express toeach of them appreciation for his contribution tothe work.

6. References[1] W. R. Sittner and E. R. Peck, J. Opt. Soc. Am. 39, 474

(1949).[2] C. J. Humphreys and E. K. Plyler, J. Research NBS 38,

499 (1947) RP1790.[3] W. F. Meggers, Zeeman Verhandelingen, p. 190 6200

(Martinus Nijhoff, the Hague, 1935): J. Research NBS14, 487 (1935) RP781.

[4] W. F. Meggers and C. J. Humphreys, BS J. Research10, 427 (1933) RP540.

[5] C. J. Humphreys and W. F. Meggers, B$ J. Research10, 139 (1933) RP521.

[6] C. E. K. Mees, J. Opt. Soc. Am. 25, 80 (1935).[7] F. Paschen, Ann. Physik [4] 60, 405 (1919).[8] G. Racah, Phys. Rev. 61, 537 (L) (1942).[9] G. H. Shortley and B. Fried, Phys. Rev. 54, 749 (1938).

[10] Charlotte E. Moore, Atomic Energy Levels, NBS Cir-cular 467, I (1949).

[11] E. K. Plyler and M. A. Lamb, J. Research NBS 45, 204(1950) RP2125.

[12] H. Roemer and R. A. Oetjen, J. Opt. Soc. Am. 36, 47(1946).

[13] R. C. Nelson and W. R. Wilson, Proc. Nat. ElectronicsConference 3,, 579 (1947).

[14] C. J. Humphreys, J. Research NBS 47,, 262 (1951)RP2252.

[15] F. Paschen and R. Gotze, Seriengesetze der Linienspek-tren (Springer, Berlin, 1922).

[16] J. D. Hardy, Phys. Rev. 38,, 2.162 (1931).

CORONA, CALIF., February 15, 1952.

84