Precise Measurements in the Infrared Spectrum of Carbon Monoxide

Earle K. Plyler, W. S. Benedict, and Shirleigh Silverman

Citation: The Journal of Chemical Physics 20, 175 (1952); doi: 10.1063/1.1700164
View online: http://dx.doi.org/10.1063/1.1700164
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/20/1?ver=pdfcov
Published by the AIP Publishing




















This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
THE JOURNAL OF CHEMICAL PHYSICS VOLUME 20, NUMBER 1 JANUARY, 1952
Precise Measurements in the Infrared Spectrum of Carbon Monoxide*
EARLE K. PLYLER AND W. S. BENEDICT,
National Bureau of Standards, Washington, D. C.
AND
SHIRLEIGH SILVERMAN
Applied Physics Laboratory, The Johns Hopkins University, Silver Spring, Maryland
(Received August 23, 1951)
New measurements have been made in the CO infrared spec-
trum in both emission and absorption, using a 15,OOO-line/inch
grating spectrograph and a 7200-line/inch vacuum grating spec-
trograph. High precision in wavelength determinations was ob-
tained by superimposing on the spectra emission lines in higher
order of Hg, Kr, and A discharges. The measurements are believed
accurate to 0.02 cm-
I
The observed lines in absorption are those
of the 2-0 and 3-0 bands of C'20'6, between R 3G-P 29 and
R 9-P 16, respectively; 27 lines of C
13
016 and 10 lines of CI'O'8
have also been measured. In emission from oxy-acetylene and
CO-O, flames, over 300 lines in the 2-0, 3-1, 4-2, and 5-3
bands have been identified between 4360 and 4000 cm-
I
The
emission bands form heads in the R branch near J = 50, and the
2-0 band shows resolved lines out to J =46 in both P and R
branches.
The new measurements are in good agreement with earlier
results at low quantum numbers, but show discrepancies at high
numbers which result in slightly different values for the molecular
INTRODUCTION
"X 11TH the development of infrared detectors of in-
VV creased sensitivity, permitting the construction
of grating spectrometers of high resolving power, it has
become both desirable and possible to increase the pre-
cision by which the wavelengths of absorption or emis-
sion lines are known. The conventional method of
fixing wavelengths has been to make use of the grating
constant, which is empirically determined by measuring
the angle at which one or two standard lines appear at
the higher orders. This method, although satisfactory
when extreme precision is not required, demands an
instrument of great mechanical and optical perfection
if systematic or random errors are not to be introduced.
In the photographic region of the spectrum, the
optical principles are the same, but because of the
greater simplicity of recording, the conventional method
of fixing wavelength has been to superimpose on the
desired unknown spectrum a -comparison spectrum of
known standard lines, which ultimately go back to the
international primary standards. For this purpose iron-
arc or rare-gas discharges are commonly used. By this
means it has been possible to measure a number of
molecular spectra with high accuracy, and thereby to
obtain knowledge of the frequencies at which other
transitions in the same molecule must appear in the
infrared region. These calculated frequencies are, how-
ever, often not in agreement.
A case in point is the vibration-rotation spectrum of
* This work was supported by the U. S. Navy, Bureau of Ord-
nance and ONR.
constants, especially in y, and De. Our constants are in cm-
I
We
2169.81
'"
0.017543
XeWe
13.284
l'
4.5XI0-6
YeWe Be
0.010 1.93130
De {3
5.95XlO-6 0
The observed isotope shifts of the 2-0 band origin for CI2-C'3
and 0
16
_0
18
are, respectively, 93.270.02 cm-
I
and 101.030.05
cm-
I
, in good agreement with the calculated values.
The emission spectra permit the determination of rotational
"temperatures" with an accuracy of 5 percent and vibrational
"temperatures" with an accuracy of 1O percent. Thermal
equilibrium between the two modes is observed in the outer cone
of the flames, but in the region just above the reaction zone, the
rotational temperatures are several hundred degrees higher than
the vibrational values or the calculated equilibrium values.
CO. The fundamental and first overtone have been
measured several times in the infrared, most recently by
Lagemann, Nielsen, and Dickey.1 The third overtone
has been measured photographically by Herzberg and
R,.ao.
2
The ground state of the molecule appears in
several electronic transitions, and recently quite ac-
curate measurements in the Cameron system (3Il _11:),
including bands involving the five lowest vibrational
levels, have been reported by Rao.
3
He has also pointed
out
4
that a discrepancy exists between the photographic
and infrared data, and has suggested that the frequen-
cies computed from the photographic data be used as
standards for the infrared region.
In the course of a study of infrared flame spectra
under high resolution,S we have been able to observe
the vibration-rotation spectrum of CO in emission.
Because of the high temperatures involved (of the order
of 2500
0
K) lines of high rotational and vibrational
quantum numbers appear. Their wavelengths were in
agreement neither with the earlier infrared measure-
ments nor with values calculated by the constants sug-
gested by Rao.
4
Accordingly, we have endeavored to
remeasure as precisely as possible the spectrum, both
in absorption and in emission. For wavelength stand-
ards we have used higher order atomic lines, super-
imposed directly on the spectra. The measurements
include absorption measurements of the 2-0 and 3-0
I Lagemann, Nielsen, and Dickey, Phys. Rev. 72, 284 (1947).
, G. Herzberg and K. N. Rao, J. Chern. Phys. 17, 1099 (1949).
3 K. N. Rao, Astrophys. J. 110, 304 (1949).
K. N. Rao, J. Chern. Phys. 18, 213 (1950).
6 S. Silverman and W. S. Benedict, Phys. Rev. 82,337 (1951).
175
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
176 PLYLER, BENEDICT, AND SILVERMAN
0 bands of C
12
016; in the former region a number of lines

z ;5
of the isotope bands 0
3
0
16
and C
12
018 also have been
Q
recognized. The emission measurements were confined l-
n.
tr
to the first overtone region, where nearly 300 lines of the


2-0,3-1,4-2, and 5-3 bands have been measured
0
between 4360 and 4000 cm-
1
. The present measure-
w- 0
;;r <.:>
ments are in quite good agreement with the data cal-
-d
culated by Herzberg and Rao for the lower energy
'"
0
levels observed by them, but yield somewhat different
u
v
'" centrifugal-stretching and higher order anharmonic
0
v
<t
..<:1

....
constants .

0
C\I
0
CD-
"0
EXPERIMENTAL METHODS AND RESULTS
;;r
v
'"
0
0-
In the measurements of the CO spectrum two grating
s
.;::
spectrometers were used. The spectrometer which was v
0-
;::I
constructed at the National Bureau of Standards con-
'"
v
tains a 15,000 lines-per-inch grating ruled at Johns ....
0
0:
0_
D
Hopkins University. An off-axis paraboloidal mirror of
C\I
<t
;::I
1 meter focus is used as the collimator. The detector is a

u
....
v
PbS cell. The impulses from the photoconducting cell

s

are amplified by a Wilson-type ac amplifier of narrow
v .... band pass. The amplifier is fed into a Leeds and North-
;o
rup Speedomax recorder. In absorption measurements
0 S
C\I_ o u
a tungsten ribbon filament lamp is used as the source C\I

<t
S 1::
of radiation. When the lamp is operated at 40 amperes
V;::l
very narrow slits may be used. The spectral slit width


was about 0.15 cm-
1
in the regions of 1.6 and 2.4Jl .

The other spectrometer used in these measurements
..<:1
0
"0 ....
was a vacuum spectrograph. It was constructed at the 0
<J) l8

on
Z
,..,

Applied Physics Laboratory of The Johns Hopkins <t 0 C\l .no:
tr
Os <.:>
University and employs a 7200 line/inch grating at

I
NO
aperture f: 6. The optics are the conventional Pfund
3l:
VIO
.;J
perforated plane mirror type, using on-axis parabolic
:r: .S{l
mirrors of 72 cm focal length. Ahead of the spectrometer
0.- I- vbl)
..,

is a monochromator employing a 20 KBr Littrow fore-
0 z v
w_
w

C\I
-'
0_
prism. The working pressure is approximately 0.1 mm, <t 0 W

tr

o v
3:
and the total outside air path from flame source to
0
R
entrance window is 6 cm. The detector used in these
0
<t
-e
experiments was a PbS cell. The signals were amplified

0:
u
....
by a 1000 cycle high gain narrow-band pass phase-
0
0
g:-
s gating amplifier designed and built at the Applied
<t ;::I
'"
Physics Laboratory. ....
u
'" The emission spectrum of CO was obtained from a
0-
'"
variety of flames produced by burning C
2
H
2
- O
2
and CD
0
.0
CO-0
2
mixtures. For th!! oxy-acetylene flame a weld-
8<
0 ing torch was employed. Various fuel-oxygen ratios
0-
'"
0
0 wt\fe used; the strongest lines were obtained when the ,..,
!::!
'"
<t
'"
'"
ratio was about 2C
2
H
2
: 10
2
. A region of the flame about
R
...<
1 inch above the inner cone was focused on the slit .
!!!

A mirror was placed back of the flame so that an image
0
<t tr
was formed inverted and superimposed on the flame.
"'-

This resulted in very nearly doubling the intensity of
,..,
<t

the CO lines. For the best runs the effective slit width

was 0.3 cm-
I
. When comparison wavelengths were de-
o tr
sired the mirror was used to focus the light of a rare-
1:
<t W
gas lamp through the flame and onto the slit.
",CD <>
0
<.>
:r: I'- tr
<t
The CO-oxygen flame spectra were obtained by
0 00

<t
- 30NI1J.lII'jSNI1IU burning premixed gases in the ratio of 5 CO to 1 of
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
INFRARED SPECTRUM OF CO 177
GO ABSORPTION
'1 I 'f I -p R-
15 to
! I I I I
I I 5 10 15
I I I I I I I I j I I I I' I ; o,e
o
V' \J V V I V I
29
,
4120
I I I I
,
4140
,
4160
,
4180
,
4200
,
4220
,
4240
I
42130
FIG. 2. A part of the 2-0 band showing some lines of the C
13
016 and the C
12
018 molecules. The cell length was 105 cm and
the pressure 76 cm. A few absorption lines of H
2
0, marked W, appear.
oxygen. Commercial grade CO containing approxi-
mately 1.5 percent of H2 as principal impurity and ordi-
nary tank O
2
were used in these experiments. Iron
carbonyl was removed from the CO by passing the gas
over a train of porcelain chips heated at 275C. The
gases were carefully metered through calibrated flow-
meters, and burned from the end of a rectangular chan-
nel of 30XO.S mm cross section and 60 mm high. The 30-
mm dimension of the burner was parallel to the optic
axis of the instrument. To further increase the signal,
a mirror was placed behind the flame, thus forming an
image on the flame. The flames as described above with
a total flow of 6 liters per minute burned with a stability
such that the noise in the system was mainly electrical.
Slits of 60 microns were used, which at this dispersion
and frequency are equivalent to about 0.5 cm-
1
for the
vacuum spectrometer.
In the absorption measurements two cells were used.
One absorption cell was 60 cm in length, and the pres-
sure of the carbon monoxide was 20 cm. The other cell
was of the multiple reflection type and had a total path
length of LOS meters. The pressure of the gas was
76 cm.
The 2-0 band, as obtained with the smaller amount
of gas, is shown in Fig. 1. A portion of the same band,
with the larger quantity of gas, is shown in Fig. 2; a
number of lines due to the isotopic molecules C13016 and
C12018 may be observed. At the longest wavelengths, a
few absorption lines due to the H
2
0 also appear, and are
identified by the symbol w. With the same quantity of
gas the 3-0 band also appears weakly in the region of
1. 58}.' ; it is illustrated in Fig. 3.
In the emission spectra, the lines of the 2-0 band
extend to much higher ] values. A head is formed
near] = 50 in the R branch, and is a prominent feature
of the spectra. Additional heads due to the higher state
bands 3-1, 4-2, and 5-3 also appear. Figure 4 is a
reproduction of the region from 4360-4180 cm-
l
from
an oxy-acetylene flame. Figure 5 shows the same region
for a CO-0
2
flame from the slit burner. The region of
the flame represented was one-inch high starting just
above the inner cone. As may be seen by comparing the
figures, the heads and resolved lines of the 5-3 and
4- 2 bands are much more prominent in the oxy-
acetylene flames, because of the higher temperature. A
further discussion of the relative intensities of the lines
is given later. The emission lines of CO, corresponding
to the P branches of the bands, extend to longer wave-
lengths than illustrated. Emission lines due to H
2
0 are
a complicating factor beyond 4200 cm-
l
, particularly
in the oxy-acetylene flames. These may be recognized
by comparing the spectra taken under different condi-
tions with oxy-hydrogen flame spectra; a few such lines
are indicated in Fig. 4. Between 4050-4000 cm-
1
the
H
2
0 emission lines are very strong compared with the
lines of CO, even when relatively hydrogen-free CO is
burned, and beyond 4000 cm-I, few if any CO lines
can be recognized.
In order to measure the wavelengths of the absorption
lines to a high accuracy, higher orders of atomic spectra
were superimposed on the absorption trace. This was
accomplished in several ways. For the absorption
measurements there were available lamps with elec-
trodes which become hot when the enclosed gas is
excited. These provided a continuum for the absorp-
tion and calibration standards at the same position.
The S-1 mercury lamp is of this type and was used in
obtaining the record shown in Fig. 1. The figure is made
"'P"

6340
"':'" "':'"
",\00
,
I

t!
j
vnVlIl
'r

,

, , ,
4
, ,
,
9 1111, I , , ,
P R
co ABSORPTION
FIG. 3. The absorption of CO in the 3-0 band. The cell length
was 105 cm and the pressure 76 cm of Hg.
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
178 PLYLER, BENEDICT, AND SILVERMAN
GO EMISSION
5-3- R504035 30
1111111111111111
4-2- R 50 40 35 30 25 20 18 16 14 12 10 8
11111111111111111111 II I I I I I I I I I I I I I I !
3-1- 509540 35 30 25 20 18 16 14 12 10 8 4 ---R P------- 4
IlillI,11111111111111111 II I I 1 I I I I I I I I I I I I I I
2-0-504540 35 30 25 20 18 16 14 12 10 8 6 4
11111111111111111111111111 I I I I I I I I I I I I I I I : I
I

I
4340
I
4320
I
4300
I
4280
-R p-
I
4260
4 6
! ! I I
I
4240
10
I
I
4220
12
I
14
I
I
4200
16
I
18
I
I
4[80
FIG. 4. The emission spectrum of CO in the 2.4M region as observed in the outer cone of a near-stoichiometric oxyacetylene flame.
from an original record on unruled recorder paper.
Alternately, the low pressure H-ll mercury lamp could
be placed in the beam which originated at the tungsten
source so that the yellow mercury lines in the fourth
order were recorded on the absorption curve. No
difference in the wavelengths of the CO lines was found
by comparison with the two sources. Other rare-gas
discharge tubes were similarly used and gave consistent
results, both with the absorption and emission measure-
ments.
The higher order lines of Hg, A, and Kr obtained in
this way served as standards throughout the region from
2.30 to 2.50J,t. They were particularly useful when pairs
of atomic lines fell so closely together that a linear dis-
persion curve could be assumed without introducing
appreciable error. The yellow mercury doublet illus-
trated in Fig. 1 is one such pair; others may be recog-
nized in Table I, which lists the atomic lines used. These
are given to O.OlA, which is sufficient for our measure-
ments. They are, however, known to eight significant
figures. Other atomic lines appear on the traces; for
example, the third line shown in Fig. 1 is a first-order
4-2-
Hg line at 23,254A. Since it has not been previously
measured with high accuracy it was not used in the
CO reduction.
By making use of close pairs of atomic lines it was
possible to measure 23 CO lines with the highest ac-
curacy. These include lines near both ends of the 2-0
absorption band and near the center. Lines were de-
termined on nine different absorption curves, taken at
different recording speeds, and with different standards.
They were measured and reduced independently by
three different observers. In the measurement the dis-
tances between the centers of the standard and CO lines
were measured with a steel ruler, graduated to 0.01
inch. The agreement between the several reductions
was such that the greatest deviation of an individual
value from the mean was 0.55A, and the probable error
of the mean less than O.lA. As a further example of the
consistency of the measurements, Table II lists a com-
parison of the average wavelengths of eight lines of the
R branch, as obtained in absorption using fourth-order
Hg lines as standards, with the values obtained in emis-
sion using third-order Kr lines as standards.
CO EMISSION
R 5040 30 25 20
1 II 11111111111111
R 50 40 35 30 25 20 18 16 14 12 10 8
I 1111111111111111111111 I I I III I I I I I I
4 6 8 10
I I I I I I 1
2-0-40 35 30 25 20 18 16 14 12 10 8 6 4 -R p_
1111111111111 II I I II I 1 I I I II 1 I I I I I I I I ! 1
4 6 8 10 12 J4 :6 18
I I I 1 1 1 1 1 I I 1 I I I 1
20
1
I
436Qcm"
I
4340
I
4320
I
4300
I
4290
I
4260
I
4240
I
4220
I
4200
I
4180
FIG. S. Emission spectrum of CO in the 2.4M region as observed just above the inner cone of a fuel-rich CO-0
2
flame.
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
INFRARED SPECTRUM OF CO 179
The same method, using Kr standards in the second
order was used to obtain wavelengths of the 3-0 band
, .
in the region of 1.6J.1. Because of the weaker absorptIOn,
the lines could not be measured with quite such ac-
curacy, but the error is believed not to exceed O.lA.
The wavelengths, measured in air, were reduced to
wave numbers in vacuum by inverting Kayer's "Tabelle
der Schwingungszahlen" and applying Babcock's cor-
rection
6
which incorporates the extrapolated value for
the index of refraction in the infrared. An experimental
verification that the refractive-index correction is not
far in error was obtained by making measurements
with the vacuum grating spectrometer, both evacuated
and filled with air. No shift of the CO absorption lines
relative to the fourth-order Hg standards could be
detected. According to the Babcock correction, the
shift should amount to +0.018 cm-
I
; the sharpness of
the lines was such that a shift of 0.03 cm-
I
would have
been noted.
As was mentioned in the introduction, it was not
possible to obtain measurements of such accuracy on
lines distant from the standards. While the average of a
number of runs may yield fairly reliable values, indi-
vidual readings, if based on a dispersion curve fitted to
the standards, either in terms of linear distance on the
chart, or in terms of the fiducial marks impressed on
the chart at equal angular displacements, may vary by
1A or more. These effects are apparently random, and
occur to about the same extent in both spectrometers.
Accordingly it was believed that in order to estab-
lish the wavelength scale between the atomic standard
lines recourse should be made to the known regu-
, .
larities which the CO lines must follow. To a hIgh ac-
curacy, the equation
v= vo+ (B' + B")m+ (B'- B")m
2
-2(D'+D")m
3
+2(D'-D")m4, (1)
where m = (1"+ 1) for the R branch and m= - I" for
the P branch, must represent the line positions. The
lines listed as observed in Table III, which include a
number in both branches from 1=17 to 27, as well as a
few near the band center, were used to determine the
TABLE 1. List of atomic lines used as standards.
5769.60A
7948.18A
8006.16
7601.55A
7685.25
7694.54
7854.82
8059.51
8104.37
Mercury
Argon
Krypton
5790.66A
8014.79A
8112.90A
8190.06
8263.24
8281.06
8298.11
6 H. D. Babcock, Phys. Rev. 46,382 (1934).
TABLE II. Comparison of absorption and emission measurements.
Line Absorption Emission
R Aair Aair
17 23157.16A 23157.04A
18 23143.86 23143.80
19 23130.90 23130.69
20 23118.38 23118.32
21 23106.04 23106.20
22 23094.51 23094.67
23 23083.05 23083.10
24 23072.34 23072.23
constants of Eq. (1). This was done with a slight varia-
tion of the usual manner, by obtaining D.2F" / (I +!) and
D. 2F'/(1+t) and plotting against 1(1+1). The results
are displayed in Fig. 6. For the state v=O, our results
are in excellent agreement with the values, at lower I,
derived from the measurements on the 4-0 band.
2
Also
included in the figure is the highly accurate value, at
1(1+1)=0, derived from the microwave determination
of the 0-1Iine.
7
Our points, since they extend to higher
I values, permit a more reliable determination of the
slope of the lines which yield the constants D. The points
of highest 1(1+1) are derived from emission lines in the
P branch which fall close to atomic standards. The cor-
responding R-branch lines require a longer interpolation,
and may be in error by O.l cm-I, but at 1=40 this
introduces very little uncertainty in the determination
of the constants. A further check on the constants is the
position of the band head, which should also be meas-
urable to 0.1 cm-
I
. Any slight uncertainty in the con-
stants D cannot affect the calculated position of the lines
of low 1. These are given in Table III. The average de-
viation from the twenty-three accurately observed lines
is 0.09A, or 0.016 cm-I, the greatest deviation being
0.29A. It may be noted that our values are about 0.07
cm-
I
lower than those calculated by Rao. It is believed
that the calculated wavelengths listed in Table III
should provide good wavelength standards in the region
23,000-24,300A.
Using the lines of the 2-0 band as standards, to-
gether with atomic comparison lines when they occur,
it is now possible to obtain measurements of the other
emission lines of CO, with an accuracy of O.OS cm-I,
except where blended. Values, averaged from a number
of measurements of both oxy-acetylene and CO-oxygen
flames, are listed in Table IV. They are compared with
values calculated from equations of the form (1) for
each band. The constants used are listed in Table V.
Combination differences for the various vibrational
states are included in Fig. 6. The observed lines in
Table IV that are italicized are those obtained directly
from the atomic standards, and are presumably the
most accurate. No "observed" values are given for a
number of 2-0 lines that appear since the calculated
values were used in establishing the wavelength scale.
The lines of the CO isotopes were measured on a
7 Gilliam, Johnson, and Gordy, Phys. Rev. 78, 140 (1950).
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
180 PLYLER, BENEDICT, AND SILVERMAN
TABLE III. A list of the observed and calculated wavelengths
in air and calculated wave numbers in vacuum for the 2-0 CO
absorption band.
Line
R33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
o
P 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
vv.c(eak)
4348.14 cm-
1
4346.76
4345.30
4343.77
4342.16
4340.48
4338.73
4336.90
4335.00
4333.03
4330.98
4328.85
4326.66
4324.39
4322.05
4319.63
4317.14
4314.59
4311.95
4309.24
4306.47
4303.62
4300.69
4297.70
4294.63
4291.50
4288.29
4285.01
4281.66
4278.23
4274.74
4271.18
4267.54
4263.84
4256.22
4252.30
4248.32
4244.26
4240.14
4235.95
4231.68
4227.35
4222.95
4218.48
4213.94
4209.33
4204.66
4199.92
4195.10
4190.23
4185.28
4180.26
4175.18
4170.03
4164.81
4159.53
4154.18
4148.76
4143.27
4137.72
4132.10
4126.42
4120.67
4114.85
4108.97
4103.02
Xair(eak)
22992.07A
22999.39
23007.09
23015.20
23023.71
23032.62
23041.93
23051.63
23061.74
23072.25
23083.17
23094.50
23106.21
23118.34
23130.89
23143.80
23157.14
23170.87
23185.03
23199.60
23214.56
23229.93
23245.72
23261.92
23278.52
23295.54
23312.93
23330.75
23349.02
23367.76
23386.85
23406.37
23426.30
23446.65
23488.43
23510.25
23532.31
23554.79
23577.70
23601.04
23624.81
23649.01
23673.67
23698.75
23724.27
23750.23
23776.65
23803.51
23830.81
23858.55
23886.76
23915.41
23944.54
23974.10
24004.13
24034.63
24065.59
24097.03
24128.94
24161.30
24194.15
24227.47
24261.29
24295.54
24330.36
24365.62
23051.53A
23061.64
23072.34
23083.05
23094.51
23106.04
23118.38
23130.90
23143.86
23157.16
23554.90
23577.68
23830.64
23858.40
23886.82
23915.38
23944.69
23973.88
24003.84
24034.66
24065.55
24161.20
24194.17
number of tracings similar to Fig. 2, using the C
I2
0I6
lines as standards. In the region beyond 4130 em-I,
where 0
2
0
16
lines become too weak to measure, the
pairs of Kr and A lines furnished standards, by which
the C
13
0I6 P 13 and P 14 lines could be measured with
the highest accuracy. The results are given in Table VI,
which lists the observed wave number and the wave
number difference from the corresponding line in C
I2
0I6.
The 3-0 band, as shown in Fig. 3, is relatively weak.
Second-order Kr lines served as standards to locate a
few lines in each branch with an accuracy of 0.05
em-I; the known 1::.2F" values, as obtained from the
constants of the 2-0 band, fixed the remainder of the
dispersion curve. The results are given in Table VII.
DISCUSSION
The band constants given in Table V are in essential
agreement with earlier accurate data. The sum of Po for
the 2-0 and 4-2 bands, 8414.44 em-I, may be com-
pared with Po for 4-0,8414.46 em-I, and with the sum
of the first four vibrational intervals for the Cameron
bands, 8414.40 em-I. If we fit our data for the 2-0,
3-1,4- 2, and 5-3 origins to the vibrational formula,
Gvib = We(V+!)-XeWe(V+!)2+YeWe(V+!)3, (2)
we obtain the constants and results listed in Table VIII.
The fit with our results is excellent. Agreement with t ~
1-0 difference obtained from the Cameron bands is less
good, but may be within the error of the measurements.
The constants we obtain differ from those of Rao prin-
cipally in WeYe, which is very sensitive to small changes
in the data; it may be noted that our constants give
good agreement with the levels up to V= 20 determined
(with less accuracy) from observations of the fourth
positive system.
s
The rotational constants we obtain are also in good
agreement with the other accurate results, as may be
seen from the plots of 1::.2F in Fig. 6. Our value of De,
5.95X 10-
6
em-I, is probably accurate to 3 percent;
it agrees within this limit with the theoretically calcu-
lated value, 4B.sjw.
2
=6.12X1O-
6
em-I. Even at the
high J values observed, the determination of D would
not be affected by the sixth-power rotational constant,
whose theoretical value is He=5.81O-I2 em-I. There is
no observable regular trend in D with vibrational state;
we have therefore arbitrarily taken (3=O; the theoreti-
cally calculated value is -1.8 X 10-
9
em-I. The recom-
mended value of B
o
, 1.92253 em-I, which fits our data
well, is based on the microwave value, using our D, and
c= 2.99790.10
10
em/sec. The Bv values have been
smoothed to fit an equation of the form Bv=B.- a(v+!)
+'Y(V+!)2, with Be= 1.93130 em-I, a= 0.017543 em-I,
and 'Y=4.5XlO-
6
em-I.
It may be noted that although the changes in the
constants Ye, D, a, (3, and 'Y that we propose are rela-
tively minor, so far as any effect on the lines of J:::: 20
is concerned, they are needed to account for the ob-
served position of the band heads. A comparison of the
heads, as calculated from our constants and from those
of Ra0
4
is given in Table IX.
8 D. N. Read, Phys. Rev. 46, 571 (1934).
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
INFRARED SPECTRUM OF CO 181
7.60
1 755
+
,
--,
g 750
,
!D
<t

i 7.45
<]'"
7.40
7.35
0- 4-0 BAND
0- 2-0 "
.- 3-1 "
0- 4-2 "
,,- MICROWAVE

J (J+ I)
FIG_ 6_ Determination of the rotational constants of CO from combination differences, for the states v=O, 1,2,3, and 4.
The observed isotopic shifts are compared in
Table VI with values calculated from our molecular
constants and the accepted values of the nuclear
masses,9 C12= 12.00382, ClS= 13.00751, 0
16
= 16, and
0
18
= 18.0049. The constants for the isotopic molecules
were obtained by multiplying the corresponding con-
stants for C120l6 by the ratio of the reduced masses,
taken to the appropriate power. The agreement with
observation is quite good. There appears to be a con-
sistent discrepancy in the ClSO- C120 shift; the average
difference Cobs-calc) for 26 lines is +0.030.02 cm-
l
.
If our results are used to obtain the ClS_ 0
2
mass
difference, they yield 1.004040.0002 amu compared
with the mass spectroscopic value, 1.003690.OO002. Al-
though our results appear to be among the more precise
determinations of mass differences by spectroscopic
methods, they are not comparable in accuracy with
nuclear or mass-spectroscopic methods. It may be of
interest to point out that they are nearly as accurate,
however, as the microwave determination. Gilliam,
Johnson, and Gordy found a frequency shift of 5069.5
0.5 mc between the J = 0-1 lines in C120 and ClSO;
using our values for the constants D and a, this corre-
sponds to a mass difference of 1.003430.OOOl amu.
Our measurements of the C016- C018 isotope shift are
less accurate, since the lines are weaker; the agreement
9 Hornyak, Lauritsen, Morrison, and Fowler, Revs. Modern
Phys. 22, 291 (1950).
with the calculated positions is fortuitously nearly
perfect, the average deviation of the 8 lines that are not
overlapped being -0.020.03 cm-
l
.
The good general agreement of our frequency meas-
urements with other accurate data, and the internal
consistency of our results, lends support to our belief
that the use of comparison standards in higher orders
has resulted in an accuracy of 0.02 cm-
l
for the lines
listed in Table III.
INTENSITIES AND TEMPERATURE
DETERMINATIONS
The intensities of the CO absorption lines in the 2- J
band are in good agreement with a recent determination.
Penner and Weber
lo
have determined, from low resolu-
tion studies of the pressure-broadened band, a value of
the integrated absorption coefficient, namely 1.64 cm-
2
atmos-
l
. They have also computed from this datum the
integrated absorption coefficients of the individual
lines. In our experiments with the gas at 1 atmos pres-
sure, the spectral slit width of 0.15 cm-
l
appears to be
of the same order as the line width. The integrated ab-
sorption coefficient of a weak line, such as P 26, which
shows a peak absorption of 15 percent, would then be
about 0.055 cm-
l
. According to Penner and Weber's
tables, for a path of 105 cm atmos, the value should be
10 S. S. Penner and D. Weber, J. Chern. Phys. 19, 807, 817
(1951).
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
.....
TABLE IV. Observed and calculated frequencies of CO lines from emission measurements.
00
N
Band 2 -0 3-1 4-2 5-3
R(J) P(J) R(J) P(J) R(J) P(J) R(J) P(J)
J" vobs "calc vobs Peale .JIobs Peale vobs i'calc ).lobs "calc vobs "eale vobs "eale "obs Peale
0 a 4263.84
1 a 4267.54 a 4256.22
2 a 4271.18 a 4252.30
3 a 4274.74 a 4248.32
4 a 4278.23 4244.24 4244.26 4225.08 4225.17
5 a 4281.66 424().14 4240.14 4228.55 4228.56 4187.39 4188.42
6 a 4285.01 a 4235.95 (4231.75) 4231.87 4183.20 4183.27
7 a 4288.29 a 4231.68 4235.04 4235.11 4179.09 4179.04 4182.05 4182.05
'"0
8 a 4291.50 a 4227.35 4238.29 4238.29 (4174.68) 4174.71 (4185.28) 4185.20 4122.27 4122.18
t"'"
9 a 4294.63 a 4222.95 4241.43 4241.39 (4170.03) 4170.39 4188.27 4188.25 (4117.89) 4117.89 ><:
10 a 4297.70 a 4218.48 4244.44 4244.42 4165.90 4165.94 4191.34 4191.25 4113.50 4113.50 t"'"
11 a 4300.69 a 4213.94 4247.32 4247.39 4161.53 4161.44 4194.13 4194.18 (4108.97) 4109.03 tTl
12 a 4303.62 a 4209.32 4250.38 4250.28 4156.91 4156.85 4197.03 4197.04 4104:41 4104.48 4143.95 4143.95 (4052.04) 4052.28 ::0
13 4306.47 4306.47 a 4204.66 4253.13 4253.09 4152.13 4152.22 (4199.92) 4199.82 4099.83 4099.89 4146.69 4146.69 4047.71
14 4309.27 4309.24 a 4199.92 4255.84 4255.83 4147.45 4147.52 4202.53 4202.52 4095.21 4095.22 4149.38 4149.35 4043.07
I:d
15 4312.02 4311.95 4195.13 4195.10 4258.53 4258.51 4142.74 4142.74 (4204.9) 4205.17 (4090.6) 4090.48 (4151.8) 4151.94 4038.36
t'I:1
16 4314.59 4314.59 4190.25 4190.23 4261.05 4261.11 (4137.72) 4137.90 4207.79 4207.74 (4085.76) 4085.68 (4154.18) 4154.48 4033.58
Z
17 4317.15 4317.15 4185.26 4185.28 4263.60 4263.63 (4132.90) 4132.98 4210.28 4210.23 4080.69 4080.80 (4156.91) 4156.94 4028.74
18 4319.63 4319.63 4180.27 4180.26 4266.10 4266.09 4128.01 4128.01 4212.59 4212.65 4()75.87 4075.87 (4159.53) 4159.32 4023.82 4023.84
tTl
19 4322.04 4322.05 4175.15 4175.18 4268.51 4268.47 4122.91 4122.96 4215.03 4215.00 4()70.72 4070.83 (4161.53) 4161.63 t:I
20 4324.38 4324.39 4170.06 4170.03 4270.80 4270.77 4117.89 4117.84 4217.30 4217.26 (4065.71) 4065.76 4163.86 4163.88
21 4326.69 4326.66 4164.86 4164.81 4273.07 4273.01 4112.65 4112.66 4219.45 4219.46 4060.47 4060.61 (4165.90) 4166.05
(')
22 4328.84 4328.86 4159.52 4159.53 4275.27 4275.16 4107.39 4107.41 4221.60 4221.58 (4055.04) 4055.41 4168.00 4168.13 >-l
23 4330.98 4330.98 4154.18 4154.18 4277.33 4277.26 4102.05 4102.10 4223.61 4223.65 4049.85 4050.16 (4170.03) 4170.16
24 4333.00 4333.03 4148.78 4148.76 4279.34 4279.27 (4096.80) 4096.72 4225.65 4225.63 (4044.85) 4044.80 4171.99 4172.10
>-
25 4334.99 4335.00 4143.25 4143.27 (4281.37) 4281.21 (4091.06) 4091.26 (4227.35) 4227.54 (4039.24) 4039.37 4173.88 4173.99
Z
26 4336.90 4336.90 4137.74 4137.72 4283.21 4283.09 4085.76 4085.75 4229.37 4229.38 4033.86 4033.89 (4175.4) 4175.70
t:I
27 4338.69 4338.73 4132.09 4132.10 (4285.10) 4284.89 (4080.15) 4080.16 4231.11 4231.15 (4028.22) 4028.34 4177.64 4177.53
28 4340.43 4340.49 4126.34 4126.42 4286.74 4286.59 4074.43 4074.52 4232.78 4232.81 4()22.63 4022.73 (4179.09) 4179.16 (fJ
29 4342.17 4342.17 4120.75 4120.67 (4288.29) 4288.23 4068.75 4068.80 4234.32 4234.41 (4016.9) 4017.04 (4181.08) 4180.73
30 4343.76 4343.77 4114.87 4141.85 4290.00 4289.81 4062.87 4063.03 (4235.95) 4235.95 4011.14 4011.30 (4182.12) 4182.23 t"'"
31 4345.28 4345.30 4109.05 4108.97 (4291.50) 4291.31 4057.00 4057.18 4237.49 4237.41 (4005.6) 4005.49 4183.82 4183.65 <:
32 4346.69 4346.76 4102.16 4103.02 4292.91 4292.73 4051.22 4051.27 4238.85 4238.81 3999.6 3999.61 (4185.28) 4184.97 tTl
33 4348.08 4348.14 (4096.82) 4097.01 4294.23 4294.08 4045.19 4045.29 (4240.14) 4240.13 3993.8 3994.0 4186.35 4186.32
::0
34 4349.40 4349.44 4091.06 4090.95 4295.39 4295.35 4039.24 4039.24 4241.35 4241.36 4187.49 4187.51
:::::
35 4350.67 4350.67 4084.88 4084.79 4296.56 4296.55 (4033.36) 4033.14 4242.44 4242.51 (4188.27) 4188.64
36 4351.86 4351.83 4078.53 4078.58 (4297.70) 4297.57 (4026.95) 4036.97 4243.58 4243.59 4189.70 4189.68
>-
37 4352.92 4352.91 4()72.35 4072.30 4298.72 4298.72 4020.86 4020.73 (4244.54) 4244.60 4190.71 4190.65
Z
38 4353.93 4353.91 (4065.71) 4065.96 4299.70 4299.68 (4014.7) 4014.42 4245.61 4245.55 4191.57 4191.57
39 4354.86 4354.84 4059.56 4059.56 (4300.69) 4300.58 4008.10 4008.05 4246.38 4246.41 4192.45 4192.39
40 4355.75 4355.70 4053.11 4053.09 4301.36 4301.39 4000.41 4001.62 (4247.26) 4247.19
41 4356.54 4356.47 4046.51 4046.56 4302.14 4302.12 3994.8 3995.12 4248.00 4247.89
42 4357.28 4353.17 (4040.18) 4039.96 4248.61 4248.54
43 4357.90 4357.80 4033.30 4033.30 4249.16 4249.10
44 4358.42 4358.34 (4026.72) 4026.58
45 4358.99 4358.82 4019.72 4019.79
46 4359.21 4012.80 4012.93
50-52 4360.01 4360.02 4305.20 4305.39 4250.80 4250.86 4196.54 4196.49
NOTE a: Line observed (in absorption or emission). Wavelength not reported since calculated value w ~ used to establish the wavelength scale. See text.
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
INFRARED SPECTRUM OF CO 183
TABLE V. Molecular constants of CO emission-absorption bands.
Band va B'+B" B'-B" 2(D'+D") D'-D"
2-0 4260.063 3.81000 -0.035060 2.38X 10--
5
0
3-1 4207.17 3.77495 -0.035042 2.38 0
4-2 4154.38 3.73992 -0.035024 2.38 0
5-3 4101.73 3.70491 -0.035006 2.38 0
3-0 6350.47 3.79249 -0.052576 2.38 0
0.00061 X 105= 0.064 em-I. The agreement is as close as
can be expected from our semi-quantitative measure-
ments. It should also be pointed out that the relative
intensity of the isotopic absorption lines of low J and
the CI20I6 lines of high J are in good agreement with
Penner's tables, given the natural isotopic abundance
ratio 1: 92. Since the spectral slit width used in measur-
ing the 3-0 band was nearly the same as that for the
2-0 band, we may also compare the maximum absorp-
tion of the weak lines in 3-0 (about 10 percent in the R
branch) with those of 2-0 to obtain a value of the
relative intensities of the two bands. The result is
3-0/2-0=0.005710 percent, which gives an in-
tegrated absorption coefficient for 3-0 of 1.64XO.0057
= (9.31)XlO-
3
cm-
2
atmos-
I
.
The relative intensity of the emission lines may be
used to test whether the various upper-state energy
levels are in thermodynamic equilibrium, and to de-
termine the "temperature" corresponding to the modu-
lus of Boltzmann's distribution of states in various
modes. For the rotational modes the transition prob-
abilities are proportional to (1'+J"+1). For the
closely-spaced lines of the R branches the instrumental
sensitivity and the factor lJ4 are nearly constant within
each band; hence if we plot the height of each un-
blended line, from J = 13 to 34, as InI/ (2J + 1), vs the
energy E' of the upper state, we should obtain straight
lines of slope 1/kT if there is thermal equilibrium. This
has been done in Fig. 7 for the spectrum of the CO-0
2
J"
o
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
TABLE VII. 3-0 absorption band of CO.
R(J) P(J)
"obs Jl ca1c "obs Veale
6354.21 cm-
1
6357.84 cm-
1
57.85 6346.62 cm-
1
61.43 61.37 6342.62 cm-
1
42.67
64.80 64.80 38.64 38.62
68.12 68.13 34.47 34.46
71.41 71.32 30.23 30.20
74.44 74.43 25.75 25.83
77.44 77.43 21.26 21.35
80.40 80.32 16.70 16.78
83.10 83.11 11.98 12.10
85.73 85.79 07.17 07.31
88.31 88.37 302.37 02.42
90.83 90.83 297.55 97.43
93.20 93.19 92.44 92.34
95.42 95.45 87.25 87.14
97.51 97.59 81.92 81.83
399.54 99.63
401.39 01.56
403.24 03.38
flame previously illustrated (Fig. 5). This has also been
done for a number of other spectra. The accuracy of
measurement, limited by flame stability, resolving
power, etc., is about 5 percent. In Fig. 7 it is seen
that each of the vibrational states V= 2, 3, and 4 are
sensibly in rotational equilibrium with about the same
"temperature" of 2800
o
K. Similar straight lines for the
rotational distribution have been obtained from other
flames studied, with temperatures ranging from as low
as 1900
0
K for the outer fringes of the outer cone of
CO-0
2
flames to as high as 3600
0
K for the inner cone
of the C
2
H
2
-0
2
flame.
A "vibrational temperature" may also be obtained
from the relative intensity of the heads or of correspond-
ing resolved lines in the various bands. This requires
a knowledge of the relative vibrational transition
probabilities. These have been calculated
U
-
13
but
with slightly discordant results, depending on the
TABLE VI. Isotopic lines in the 2-0 absorption band of CO.
J"
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Cl3()l6
R(J) P(J)
"'obs .6.
lI
obs a"ealc Vobs Llvobs .6.v
ca
1c Vabs
4173.93 93.61 93.56 4163.02? 93.20 93.07
77.38 93.80 93.71 92.89
93.86 55.57 92.75 92.72 4172.92
84.16 94.07 94.01 51.67 92.59 92.53 76.43
87.48 94.18 94.15 47.76 92.38 92.34
94.29 43.77 92.18 92.18 (82.84)
93.88 94.41 94.42 39.67 92.01 91.96
196.96 94.54 94.55 35.55 91.80 91.75
94.67 31.34 91.61 91.55 91.93
202.90 94.80 94.79 27.12 91.36 91.34
05.70 94.99 94.90 22.82 91.12 91.12 97.87
(08.63) 94.99 95.01 18.44 90.89 90.91
11.31 95.16 95.11 13.99 90.67 90.68
95.23 09.50 90.42 90.46
16.66 95.29 95.32 (04.81) 90.29 90.22
95.41 4100.11 90.12 89.98
21.61 95.53 95.48
1l B. L. Crawford, Jr., and H. L. Dinsmore, J. Chem. Phys. 18, 983 (1950).
12 K. Scholz, Z. Physik 78, 751 (1932).
13 R. C. Herman (private communication).
C
12
0
18
R(J) P(J)
Llvobs Llvcalc Vobs Llvobs AVcalc
101.82 101.73
101.80 101.88
102.04
102.17 102.18
102.33
102.46
102.57 102.60 4123.77 99.18 99.21
102.73 19.54 98.94 98.99
102.82 102.85 15.19 98.75 98.76
10.86 98.47 98.52
06.37 98.29 98.28
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36
184 PLYLER, BENEDICT, AND SILVERMAN
TABLE VIII. Vibrational constants and levels for 0
2
0
16

<0>, =2169.81 em-
i
, = 13.284 em -1, ".Y. =0.010 cm-
1
G,-G. vo, overtone bands
calc obs calc ob.
2143.38
8
v=l 2143.27 2143.16& 3-1 4207.17 4207.17"
2 4260.07 4260.06
0
4-2 4154.38 4154.38
3 6350.44 6350.47
0
5-3 4101.74 4101.73"
4 8414.44 8414.46
d
5 10452.18
a Reference 3; b Reference 1; (l This work; d Reference 2.
TABLE IX. Band heads in the CO emission spectrum.
Band
Pobs, cm-1
'eale. PBS veale, Rao
2-0 4360.01 4360.02 4359.46
3-1 4305.20 4305.29 4304.85
4-2 4250.80 4250.86 4250.09
5-3 4196.54 4196.49 4195.78
approximations made. In the simplest theoryll in
which both electrical and mechanical anharmonicities
are introduced as perturbations on harmonic oscillator
wave functions, the probability 1M 12 for a transition
is proportional to (v+n)!lv!. For the 2-0,
3-1, 4- 2, and 5-3 transitions this gives intensity
ratios 1:3:6:10. In the more detailed theories, using
anharmonic oscillator wave functions, the probabilities
deviate slightly from these integers; according to Her-
man/a the electrical anharmonicity portion is as above,
whereas the mechanical-anharmonicity portion is as
1 :3+28x:6+112x: 10+ 280x. Since for CO, x=0.00612,
there is an uncertainty of some 10 percent in the transi-
tion probabilities. This is the same order as the experi-
mental uncertainty in the determination of vibrational
temperature, for fewer points are available to determine
a slope, and the instrumental sensitivity may vary
over the wider frequency range involved. In Fig. 7 we
also present the results of a determination of vibrational
temperature, plotting the logarithm of the intensity of
the R 19 line of three bands, divided by the Herman
transition probabilities Xv4, against l. Again a fairly
good straight line is obtained, but with a lower effective
temperature, 2360
o
K, for the vibrational mode than
the 2800
0
K that was found for rotation. This would
indicate that complete thermal equilibrium has not
been established. This finding is not unexpected since
the observations were made in a region of the flame

0
l!eOO eeoc 7500
8600 I .. J 9500
,_0
11500
FIG. 7. Determination of "Vibrational Temperatures" Tv and
"Rotational Temperatures" TR for CO from the emission of the
CO-0
2
flame (Fig. 5).
just above the reaction zone, where conditions definitely
are not in equilibrium. In the outer cone of similar
flames the vibrational temperatures are nearly identical
with the rotational temperatures. A more detailed dis-
cussion of the intensity distributions will be reported
elsewhere.
The uncertainty in the vibrational transition prob-
abilities does not affect the above conclusions. Some-
what better linearity of the plots was found when the x
term was included, which might be taken to indicate
that the mechanical anharmonicity term predominates
in the expression for the first overtone intensity. It
may be noted that in one of the alternative derivations
of these constants,lO the electrical anharmonicity, Pe', is
nearly zero.
Since the absolute transition probabilities of the
overtone bands have been determined, it is possible to
calculate the expected emission intensity corresponding
to thermal equilibrium at the indicated rotational or
vibrational "temperatures." This leads to emissivities of
the order of 0.01 for the strongest lines whereas the ob-
served emissivity in the flames, which has not been
determined quantitatively, appears to be of the same
order. Further work is needed in this respect, to de-
termine if deviations from thermal equilibrium exist
between the upper and lower states involved, and if the
emitted intensity may be used to determine the CO
concentration and distribution in flames.
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.164.20.19 On: Thu, 21 Nov 2013 04:14:36