Perturbations in the pure rotational spectrum of CoCl (X3Φ): A submillimeter study

Article abstract of DOI:10.1063/1.1795691

The millimeter/submillimeter-wave spectrum of the CoCl radical (X ³Φ_i) has been recorded using direct absorption techniques in the frequency range 340-510 GHz. This work is the first pure rotational study of this molecule. The radical was created by the reaction of Cl₂ with cobalt vapor. Rotational transitions arising from the Ω=4, 3, and 2 spin-orbit components of Co³⁵Cl have been measured, all of which exhibit hyperfine splittings due to the ⁵⁹Co nucleus (I=7/2). Transitions arising from the Co³⁷Cl species were also recorded, as well as those originating in the v = 1, 2, 3, and 4 vibrational states of both isotopomers. The spin-orbit pattern exhibited by the molecule is unusual, with the Ω=3 component significantly shifted relative to the other spin components. In addition, the regular octet hyperfine splittings become distorted above a certain J value for the Ω=3 transitions only. These effects suggest that the molecule is highly perturbed in its ground state, most likely a result of second-order spin-orbit mixing with a nearby isoconfigurational ¹Φ₃ state. The complete data set for Co³⁵Cl and Co³⁷Cl were fit successfully with a case (a) Hamiltonian but required a large negative spin-spin constant of λ=-7196 GHz and higher order centrifugal distortion corrections to the rotational, spin-orbit, spin-spin, and hyperfine terms. The value of the spin-spin constant suggests that the Ω=3 component is shifted to higher energy and lies near the Ω=2 sublevel. The hyperfine parameters are consistent with a δ³π³ electron configuration and indicate that CoCl is more covalent than CoF.

Full text of DOI:10.1063/1.1795691

Perturbations in the pure rotational spectrum of CoCl (X 3 Φ i ): A submillimeter study
M. A. Flory, D. T. Halfen, and L. M. Ziurys
Citation: The Journal of Chemical Physics 121, 8385 (2004); doi: 10.1063/1.1795691
View online: http://dx.doi.org/10.1063/1.1795691
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/121/17?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Rotational spectra of rare isotopic species of fluoroiodomethane: Determination of the equilibrium structure from
rotational spectroscopy and quantum-chemical calculations
J. Chem. Phys. 137, 024310 (2012); 10.1063/1.4731284
Molecules in high spin states III: The millimeter/submillimeter-wave spectrum of the MnCl radical (X 7 Σ + )
J. Chem. Phys. 122, 054309 (2005); 10.1063/1.1824036
The pure rotational spectrum of TiF ( X 4 Φ r ):3d transition metal fluorides revisited
J. Chem. Phys. 119, 9496 (2003); 10.1063/1.1615753
Further studies of 3d transition metal cyanides: The pure rotational spectrum of NiCN (X 2 Δ i )
J. Chem. Phys. 118, 6370 (2003); 10.1063/1.1557471
The rotational spectrum of chloryl chloride, ClClO 2 , in its ground vibrational state
J. Chem. Phys. 110, 11865 (1999); 10.1063/1.479179
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07

JOURNAL OF CHEMICAL PHYSICS
VOLUME 121, NUMBER 17
1 NOVEMBER 2004
Perturbations in the pure rotational spectrum of CoCl „X 3⌽i…:
A submillimeter study
M. A. Flory, D. T. Halfen, and L. M. Ziurys
Department of Chemistry, Department of Astronomy, and Steward Observatory, University of Arizona,
Tucson, Arizona 85721
͑Received 13 May 2004; accepted 29 July 2004͒
The millimeter/submillimeter-wave spectrum of the CoCl radical (X 3⌽i) has been recorded using
direct absorption techniques in the frequency range 340–510 GHz. This work is the first pure
rotational study of this molecule. The radical was created by the reaction of Cl2 with cobalt vapor.
Rotational transitions arising from the ⍀ϭ4, 3, and 2 spin-orbit components of Co35Cl have been
measured, all of which exhibit hyperfine splittings due to the 59Co nucleus (Iϭ7/2). Transitions
arising from the Co37Cl species were also recorded, as well as those originating in the ϭ1, 2, 3,
v
and 4 vibrational states of both isotopomers. The spin-orbit pattern exhibited by the molecule is
unusual, with the ⍀ϭ3 component significantly shifted relative to the other spin components. In
addition, the regular octet hyperfine splittings become distorted above a certain J value for the ⍀ϭ3
transitions only. These effects suggest that the molecule is highly perturbed in its ground state, most
1
likely a result of second-order spin-orbit mixing with a nearby isoconfigurational ⌽3 state. The
complete data set for Co35Cl and Co37Cl were fit successfully with a case ͑a͒ Hamiltonian but
required a large negative spin-spin constant of ␭ϭϪ7196 GHz and higher order centrifugal
distortion corrections to the rotational, spin-orbit, spin-spin, and hyperfine terms. The value of the
spin-spin constant suggests that the ⍀ϭ3 component is shifted to higher energy and lies near the
3
⍀ϭ2 sublevel. The hyperfine parameters are consistent with a ␦3␲ electron configuration and
indicate that CoCl is more covalent than CoF. © 2004 American Institute of Physics.
͓DOI: 10.1063/1.1795691͔
I. INTRODUCTION
the range 415–725 nm using laser-ablation/molecular beam
laser-induced fluorescence. They concluded both bands arose
from a X 3⌽4 term. Consequently, in analogy to CoF ͑Ref.
11͒ and CoH,12 these authors assigned the ground state of
CoCl as X 3⌽i . Adam et al. also were able to resolve the
cobalt hyperfine structure and determine the h parameter for
several states. This work was followed by Fourier transform
infrared spectroscopy of Hirao, Pinchemel, and Bernath in
2003,13 who recorded the same bands as Adam et al., but
also observed additional subbands; they were assigned to the
⍀ϭ3 spin component of the ground state. In addition, the
10.3 3⌽ –X 3⌽ and 11.0 3⌽4 –X 3⌽4 electronic transi-
Although cobalt has its origin in the helium shell burn-
ing in asymptotic giant branch stars and in explosive nucleo-
synthesis in supernovae,1 it has many practical uses here on
earth. For example, this element is employed as a catalyst in
many industrial processes, such as hydroformylation,2 and is
used in olefin insertion reactions.3 Cobalt appears as the
metal center in various biological molecules, including vita-
min B12 , and in low symmetry sites in enzymes.4 It is used
in metal alloys and in supramolecular chemistry in the con-
struction of large metal complexes.5 Cobalt complexes have
additionally served as dyes in ceramics and glass for thou-
sands of years. Yet, despite the many applications of this
element, the exact nature of how it bonds to even simple
ligands is not well understood.6
͓
͔
͓
͔
4
4
tions were studied in the near infrared by Wong, Tam, and
Cheung,14 who determined the h parameter in several vi-
bronic levels.
Here we present the first investigation of the pure rota-
tional spectrum of CoCl using millimeter/submillimeter di-
rect absorption spectroscopy. Fifteen and twelve rotational
transitions were recorded for Co35Cl and Co37Cl, respec-
tively. All three spin components were clearly identified in
the spectra, which repeated in both chlorine isotopomers and
in several vibrational states, providing conclusive evidence
The lack of understanding of the bonding in cobalt is
illustrated by the history of a simple species, CoCl. Although
cobalt ͑II͒ chloride is routinely used in general chemistry
labs, until recently the electronic ground state of the
monochloride CoCl was not known. Spectroscopic studies of
this diatomic began as early as the 1930s, where various
electronic bands were recorded in discharge experiments.7,8
However, the data were too complex to definitely assign an
electronic ground state. As recently as 1997, theorists were
still postulating what the term might be; using density func-
tional theory methods, for example, Bridgeman concluded
the ground state was 3⌺Ϫ.9 However, in 2002, Adam et al.10
recorded two electronic systems of cobalt monochloride in
3
that the electronic ground state is ⌽i . In addition, cobalt
hyperfine structure was resolved in all three sublevels. Both
the fine and hyperfine structure showed evidence of pertur-
bations from a nearby excited state. In this paper we present
these results and their analysis. We also discuss the origin of
perturbations and the nature of the bonding in CoCl.
0021-9606/2004/121(17)/8385/8/$22.00
8385
© 2004 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07

8386
J. Chem. Phys., Vol. 121, No. 17, 1 November 2004
Flory, Halfen, and Ziurys
II. EXPERIMENT
The spectrum of CoCl was recorded using one of the
quasioptical millimeter/submillimeter direct absorption sys-
tems of the Ziurys’ group. These instruments are described
elsewhere.15 Briefly, millimeter-wave radiation originating at
a Gunn oscillator/Schottky diode multiplier source is passed
through a polarizing grid and a series of mirrors into a reac-
tion chamber. This chamber is a double-pass system with a
roof-top reflector at the far end. After its second pass through
the cell, the radiation is directed back through the optics and
is reflected by the grid into a He-cooled InSb detector. The
source is modulated at a rate of 25 kHz and signals are re-
corded at 2f using a lock-in amplifier.
CoCl was produced by reacting cobalt vapor with pure
Cl2 . CCl4 was originally used as the chlorine donator with
some success, but switching to Cl2 increased the signals five-
fold. No dc discharge was necessary in either case. Cobalt
vapor was obtained by melting chips of cobalt ͑Aldrich
99.5%͒ in a Broida-type oven through resistive heating. Be-
cause of cobalt’s high melting point ͑ϳ1495 °C͒, the crucible
was wrapped in zirconia insulation. Chlorine was added over
the top of the crucible; between 0.5 and 1.0 mTorr of Cl2
resulted in the best signal. Both chlorine isotopes were ob-
served in their natural abundance of 35Cl:37Clϭ3:1. No car-
rier gas such as argon was used, as it did not increase the
signal-to-noise ratio. Use of a dc discharge increased the
intensity of CoCl lines somewhat but also added additional
noise, so in the end little net improvement was gained. The
reaction mixture exhibited no obvious fluorescence.
FIG. 1. A stick diagram illustrating the progression of the Jϭ43 42 rota-
tional transition of Co35Cl (X 3⌽i), observed in the range 460.5–470 GHz
͑black lines͒, interspersed with its vibrational progression from the Jϭ44
43 transition ͑dashed lines͒ and lines originating from the Jϭ45 44
transition of Co37Cl ͑gray scale͒. Each line actually consists of an octet, but
this detail is not shown. The unusual spin-orbit pattern of CoCl is evident,
with the ⍀ϭ3 component lying about 500 MHz lower in frequency relative
to that of ⍀ϭ4, while the ⍀ϭ2 line is located Ϸ8 GHz to higher frequency.
The vibrational pattern for the ⍀ϭ4 ladder of Co35Cl is also easily identi-
fied, but the presence of the Co37Cl isotopomer, its vibrational satellite lines,
and the vibrationally-excited states of the ⍀ϭ2 and 3 components of Co37Cl
complicate the spectrum.
tive to the ⍀ϭ4 component, while the possible ⍀ϭ2 line
was located ϳ8 GHz higher. Both lines had counterparts in
the 37Cl isotopomer, as well as their own vibrational progres-
sion. After this tentative assignment was made, the Hirao,
Pinchemel, and Bernath paper13 was published, which con-
firmed the ⍀ϭ3 identification; the remaining sets of lines
were assigned to the ⍀ϭ2 component by process of elimi-
nation.
Rotational constants from optical studies10,13 provided
the basis for the initial measurements. After scanning
through several transitions ͑ϳ45 GHz͒ such that harmonic
relationships and spectral patterns could be established, ob-
served lines were identified. Transition frequencies were
measured by averaging an even number of scans in both
increasing and decreasing frequencies. These scans were
typically 5 MHz in width, and two to four were necessary to
obtain adequate signal to noise. Center frequencies were de-
termined by fitting the observed lines to Gaussian profiles.
The instrumental accuracy is ϷϮ50 kHz. Typical linewidths
were 1.0–1.5 MHz over the range 343–508 GHz.
The spectral pattern is illustrated in Fig. 1, which is a
stick figure ranging over 2B at the frequency of the Jϭ43
42 transition of Co35Cl. The ⍀ϭ4 and 3 spin components
are clustered at the left ͑lower frequency͒ and the ⍀ϭ2 fea-
ture is at the far right ͑higher frequency͒, all indicated by
solid black lines. The ⍀ϭ2 and 3 components are consider-
ably weaker in intensity. Vibrational satellite features due to
Co35Cl are portrayed by dashed lines; the ϭ1, 2, 3, and 4
v
progression is clearly visible for the ⍀ϭ4 component, and a
few lines arising from the other spin levels are additionally
present. The Co37Cl lines for the ϭ0 spin components are
v
III. RESULTS
shown in gray scale; these appear at ‘‘odd’’ frequencies be-
cause they come from the Jϭ45 44 ͑⍀ϭ4 and 3͒ and J
ϭ44 43 ͑⍀ϭ2͒ transitions—a result of the isotopic shift.
The remainder of the features arise from vibrationally ex-
cited states of the 37Cl isotopomer. Not shown in this dia-
gram are the hyperfine interactions, which split each depicted
line into octets.
Representative data recorded for the main isotopomer of
CoCl are shown in Fig. 2. Here a composite spectrum of the
three spin components of the Jϭ35 34 transition has been
created using two frequency breaks. The insert shows an
enlarged version of the ⍀ϭ2 data. The regular octet pattern,
arising from the 59Co spin of Iϭ7/2, is apparent in each
component. ͑Splittings due to the chlorine nuclear spin of I
It was clear from the beginning that CoCl was not a
good Hund’s case ͑a͒ species, and therefore did not have the
regular intervals between the spin components exhibited by
TiF ͑Ref. 16͒ or FeCl,17 for example. No obvious spin pat-
tern was apparent in the spectra. Therefore, features were
sought that repeated themselves in the 37Cl isotopomer. This
procedure enabled the vibrational progression in both isoto-
pomers for the ⍀ϭ4 spin component to be located, ϭ1
v
through 4 for Co35Cl and ϭ1 through 3 for Co37Cl. The
v
remaining sets of lines had to be due to the other ⍀ ladders
and their respective vibrational satellite features. The second
strongest of the remaining features, a likely candidate for the
⍀ϭ3 ladder, however, was shifted to lower frequency rela-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07

J. Chem. Phys., Vol. 121, No. 17, 1 November 2004
Rotational spectrum of CoCl
8387
FIG. 2. Spectra of the Jϭ35 34 rotational transition of Co35Cl (X 3⌽i)
FIG. 3. Spectra of the hyperfine octet recorded for the Jϭ42 41 ͑left͒ and
Jϭ44 43 transitions of the ⍀ϭ3 component of Co35Cl ( ϭ0). Unlike the
showing all three spin-orbit components ( ϭ0) on the same intensity scale.
v
v
This spectrum is a composite of three separate frequency ranges that are
actually widely spaced; hence, there are two breaks in the frequency scale
͑see Fig. 1͒. The octet hyperfine pattern arising from the cobalt nuclear spin
of Iϭ7/2 is clearly resolved in each sublevel. The ⍀ϭ3 and 2 features are
considerably weaker than those of ⍀ϭ4, suggesting that they lie close in
energy. However, these two components were still recorded with remarkable
signal to noise, as the enlarged spectrum of the ⍀ϭ2 octet demonstrates.
These three data sets are 80 MHz wide and were acquired in single, 60 s
scans.
octet in Fig. 2, these hyperfine patterns appear distorted. The Jϭ42 41
features are broadened and appear to be additionally split, while the Jϭ44
43 lines have evolved into nine components as opposed to eight. These
nonets appeared in successive transitions to higher J. Each spectrum is 40
MHz wide and was acquired in a single 30 s scan.
sured as well. All lie in the range 373–493 GHz (J
Х34–47). For each of the ⍀ϭ2 and 3 spin components,
ϭ3/2 were not resolved in the data—not surprising since
cobalt has a nuclear moment about six times larger than that
of Cl.18͒ As these data show, the overall splitting is smallest
in the ⍀ϭ3 lines and largest in the ⍀ϭ2 hyperfine set. Also
evident in this figure are the relative intensities of the three
spin components, which are ⍀ϭ4:3:2ϭ25:2:1. This ratio
suggests a rotational temperature of Trotϳ270–430 K based
on the overall spin-orbit splitting of 6A or ϳ1440 K. The
⍀ϭ3 and 2 components appear quite weak in comparison
with the ground state ⍀ϭ4 lines; however, these data were
actually recorded with high signal to noise with little scan
averaging, as the insert for the ⍀ϭ2 octet shows. Significant
signal averaging was only necessary for highly-excited
four to five transitions were measured in the ϭ1 level for
v
both chlorine isotopomers, from 373 to 493 GHz (J
Х34–46). In virtually every transition, frequencies for all
eight hyperfine components were recorded, a total of 279
separate measurements. These data sets also are also avail-
able electronically on EPAPS.19
As may be evident from Table I many of the ⍀ϭ3 fre-
quencies were not measured, in particular for those above J
Ͼ38, although data has been recorded for the other two spin
components in this range. These transitions were certainly
located, but the hyperfine pattern exhibited unusual behavior
that could not be satisfactorily analyzed. At the Jϭ39 38
transition in the ⍀ϭ3 state of Co35Cl, the individual lines of
the hyperfine octet were found to broaden into multiple, un-
resolved features. By the Jϭ44 43 transition, the lines be-
come narrow again but there are nine such features, not
eight. The extra lines likely arise from ⌬FϭϮ2 transitions,
perhaps a result of mixing with a nearby vibrational state.
These ‘‘nonets’’ remained in the ⍀ϭ3 data up to the highest
J level studied. A similar trend was observed in the Co37Cl
states, such as the ϭ1 level of the ⍀ϭ3 component of
v
Co37Cl. ͑The ϭ0–1 spacing is 370 K.͒ The similar inten-
v
sities of the ⍀ϭ2 and 3 states suggests that they may be
closer in energy than predicted by the spin-orbit splitting
alone.
Selected transition frequencies for Co35Cl in the ϭ0
v
state are presented in Table I. As the table shows, frequencies
of the eight hyperfine components were typically measured
per spin level per transition. Fifteen transitions were studied
for the main CoCl isotopomer; several include all three spin
components. In the case of Co37Cl, twelve transitions were
measured. The complete data set spans the range 343–507
GHz. The complete list of frequencies for both isotopomers
is available electronically on EPAPS.19
isotopomer as well for the ⍀ϭ3 levels. In contrast, the
v
ϭ1 transitions for ⍀ϭ3 did not show any such perturba-
tions. Wong, Tam, and Cheung also saw some perturbations
in the cobalt hyperfine structure, but it was for the ϭ7 level
v
in the 10.3 3⌽i excited state.14
͓
͔
Figure 3 shows examples of these perturbations for
Co35Cl. On the left, the Jϭ42 41 transition is displayed.
The crisp octet pattern of Fig. 2 has clearly degenerated into
broadened features for most of the hyperfine components. On
the right, the Jϭ44 43 transition is shown. Here the lines
have narrowed again, but there are nine components. These
In addition to the ϭ0 lines, four transitions originating
v
in each of the ϭ1, 2, 3, and 4 levels of Co35Cl ͑⍀ϭ4 only͒
v
were recorded. Four to five transitions originating in each of
the ϭ1, 2, and 3 states of Co37Cl ͑⍀ϭ4 only͒ were mea-
v
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07

8388
J. Chem. Phys., Vol. 121, No. 17, 1 November 2004
Flory, Halfen, and Ziurys
TABLE I. Selected transition frequencies for Co35Cl (X 3⌽ : ϭ0) ͑in MHz͒.
v
i
F
F
⍀
␯
␯
F
F
Љ
⍀
␯
␯
obsϪcalc
J
J
Ј
Љ
J
J
Љ
Ј
Ј
Љ
Ј
obsϪcalc
32
31
28.5 27.5
29.5 28.5
30.5 29.5
31.5 30.5
32.5 31.5
33.5 32.5
34.5 33.5
35.5 34.5
3
343 577.266
343 573.101
343 568.622
343 563.901
343 558.348
343 553.534
343 547.873
343 541.913
0.798
0.799
0.748
0.714
0.104
0.487
0.272
0.007
39
38
39
44
46
35.5 34.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
40.5 39.5
41.5 40.5
42.5 41.5
35.5 34.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
40.5 39.5
41.5 40.5
42.5 41.5
4
418 949.486
418 946.485
418 943.351
418 940.039
418 936.591
418 932.956
418 929.154
418 925.199
426 266.479
426 261.953
426 257.470
426 253.059
426 248.671
426 244.357
426 240.113
426 235.905
0.432
0.395
0.399
0.399
0.437
0.460
0.487
0.531
0.851
0.370
0.035
2
36
35
32.5 31.5
33.5 32.5
34.5 33.5
35.5 34.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
32.5 31.5
33.5 32.5
34.5 33.5
35.5 34.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
32.5 31.5
33.5 32.5
34.5 33.5
35.5 34.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
4
386 846.516
386 843.072
0.057
0.048
Ϫ0.126
Ϫ0.165
Ϫ0.033
0.267
386 831.518
386 827.245
386 822.773
386 818.099
386 369.743
386 366.488
386 362.939
386 359.259
386 355.318
386 351.238
386 346.885
386 342.280
393 809.112
393 804.384
393 799.670
393 794.953
393 790.238
393 785.701
393 781.104
393 776.625
0.132
0.178
0.243
0.321
Ϫ0.694
Ϫ0.491
Ϫ0.399
Ϫ0.258
Ϫ0.198
Ϫ0.100
Ϫ0.099
Ϫ0.177
0.150
Ϫ0.253
Ϫ0.512
Ϫ0.645
Ϫ0.650
Ϫ0.353
0.005
0.697
3
40
45
47
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
40.5 39.5
41.5 40.5
42.5 41.5
43.5 42.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
40.5 39.5
41.5 40.5
42.5 41.5
43.5 42.5
4
429 643.737
429 640.871
429 637.847
429 634.695
429 631.415
429 627.971
429 624.387
429 620.603
437 069.490
437 065.025
437 060.638
437 056.365
437 052.057
437 047.781
437 043.646
437 039.585
0.124
0.085
0.050
0.049
0.080
0.106
0.152
0.155
0.600
2
2
0.102
Ϫ0.224
Ϫ0.341
Ϫ0.401
Ϫ0.338
Ϫ0.045
0.411
0.602
37
36
33.5 32.5
34.5 33.5
35.5 34.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
40.5 39.5
33.5 32.5
34.5 33.5
35.5 34.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
40.5 39.5
33.5 32.5
34.5 33.5
35.5 34.5
36.5 35.5
37.5 36.5
38.5 37.5
39.5 38.5
40.5 39.5
4
3
2
397 550.793
397 547.501
397 544.039
397 540.364
397 536.535
397 532.509
397 528.265
397 523.860
397 060.610
397 057.088
397 053.684
397 050.157
397 046.484
397 042.732
397 038.581
397 034.269
404 636.294
404 631.556
404 626.970
404 622.277
404 617.726
404 613.186
404 608.812
404 604.263
0.564
0.538
0.548
0.548
0.598
0.652
0.689
0.763
0.546
0.334
0.409
0.527
0.663
0.884
0.868
0.850
0.889
0.376
0.134
Ϫ0.097
Ϫ0.069
0.084
0.516
0.883
41.5 40.5
42.5 41.5
43.5 42.5
44.5 43.5
45.5 44.5
46.5 45.5
47.5 46.5
48.5 47.5
41.5 40.5
42.5 41.5
43.5 42.5
44.5 43.5
45.5 44.5
46.5 45.5
47.5 46.5
48.5 47.5
4
483 060.020
483 057.630
483 055.209
483 052.618
483 050.037
483 047.371
483 044.533
483 041.608
490 959.967
490 956.020
490 952.087
490 947.968
490 944.403
490 940.582
490 936.904
490 933.355
0.525
0.401
0.360
0.262
0.286
0.338
0.329
0.344
0.485
0.192
Ϫ0.020
Ϫ0.351
Ϫ0.063
0.035
0.340
0.836
2
44.5 43.5
45.5 44.5
46.5 45.5
47.5 46.5
48.5 47.5
49.5 48.5
50.5 49.5
4
504 397.066
504 394.712
504 392.419
504 389.987
504 387.516
504 384.899
504 382.221
Ϫ0.292
Ϫ0.460
Ϫ0.467
Ϫ0.514
Ϫ0.503
Ϫ0.539
Ϫ0.539
anomalous splittings cannot be due to the chlorine nucleus,
as the expected pattern would be quartets of octets. Also, the
chlorine hyperfine interactions would be expected to play a
role at lower, not higher, J. Hyperfine interactions arising
from fluorine have been resolved in CoF,20 producing dou-
blets of octets I(F)ϭ1/2 , but not the unusual pattern ob-
͓
͔
served here. ͑The fluorine magnetic moment is about a factor
of 3 larger than that of chlorine.18͒
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07

J. Chem. Phys., Vol. 121, No. 17, 1 November 2004
Rotational spectrum of CoCl
8389
sublevel is shifted relative to the case ͑a͒ spin-orbit pattern.23
An excited ⌬2 state is thought to be the perturber in this
case. FeC has a large value of ␭ ͑1 285 GHz͒, except it is
positive, and the shift is in the opposite direction in fre-
quency.
IV. ANALYSIS
1
The observed rotational transitions were analyzed using
3
an effective ⌽i Hamiltonian of the form
HeffϭHrotϩHsoϩHssϩHhf .
͑1͒
This Hamiltonian describes molecular frame rotation (Hrot),
spin-orbit coupling (Hso), spin-spin coupling (Hss), and the
magnetic hyperfine interactions (Hhf), including their cen-
trifugal distortion corrections. The coupling scheme em-
ployed in this analysis is Hund’s case (a␤). Therefore,
JϭLϩS and FϭJϩI. This scheme is appropriate because
the spin splittings are far larger than the hyperfine interac-
tions.
To fit the data, the spin-orbit constant A was initially set
͑165 cmce1͒ to the value for CoF ͑Ϫ233 cmϪ1͒.21 It was later
determined that this value did not provide a good fit, so A
was varied until an optimal value was obtained. In the final
fit, it was then fixed to this value ͑166 cmϪ1͒. Higher order
corrections to the spin-orbit and spin-spin constants were
found to be necessary for a satisfactory analysis (AH , AL ,
and ␭H). The hyperfine splittings were fit to the a, b, and
(bϩc) parameters only, as no improvement was obtained by
using eQq. An additional term was also added to the hyper-
fine Hamiltonian exclusively for the ⍀ϭ3 levels of the form
A. Fine structure splittings
The spin-orbit and spin-spin interactions compete in a
3⌽ state to establish the fine structure pattern. When the
spin-orbit constant A is larger than ␭, the pattern is fairly
regular for a ‘‘good case ͑a͒’’ molecule. Examples are TiF in
its X 4⌽r state, where Aϭ1000 GHz and ␭ϭ4 GHz ͑Ref. 16͒
and FeCl (X 6⌬i), for which AϭϪ2274 GHz and ␭ϭ16.4
GHz.17 In the case of FeC (X 3⌬i), where some evidence of
perturbations is observed, AϷ␭.23 For CoCl, however, ␭ is
greater than A ͑Ϫ7196 GHz versus Ϫ5000 GHz͒. The large
value of ␭ is needed to compensate for the large frequency
shift of the ⍀ϭ3 spin component, which may be perturbed
1
by a ⌽3 term. Second-order spin-orbit interactions connect
1
3
ˆ
ˆ
the ⌽3 and ⌽3 states via the li"si operator. Although the
1⌽ state has not yet been observed in the spectrum of CoCl,
theoretical calculations predict such a term 2000 cmϪ1 higher
in energy for CoH, which has an identical electron
configuration.12 For CoCl, a substantially heavier molecule, a
lower energy would be expected for this state.
h3Dϭ3 F Fϩ1͒ϪJ Jϩ1͒ϪI Iϩ1͒ /2.
͓ ͑
͑2͒
͑
͑
͔
It is possible to estimate the energy separation between
the 3⌽3 and 1⌽3 states, ⌬E, by ‘‘undiagonalizing’’ the spin-
orbit energy matrix, as described in Refs. 24 and 25. The
off-diagonal matrix element for this interaction is
3
This term was originally created for the analysis of the ⌽3
sublevel of CoH.22
The data for both chlorine isotopomers and their ϭ1
v
3
1
ˆ
ˆ
ˆ
ˆ
states were fit in an identical manner. For Co37Cl, however,
the range of J was smaller and hence the higher order cen-
trifugal distortion corrections were not needed. The resulting
spectroscopic parameters from these global fits are provided
in Table II. Also given are the Be and ␣e constants derived
⌽ ͉Hso͉ ⌽ , where H ϭ͚ a l "s . Assuming that both
͘
3 3 so i i i i
͗
3
these states arise from a single ␦3␲ electron configuration,
then
a
␲
3
1
ˆ
ˆ
⌽
a l •s
⌽
ϭa␦Ϫ
,
͑3͒
͑4͒
3
ͯ
͚
i i
i
ͯ
3
ͳ
ʹ
2
i
from the ϭ0 and ϭ1 data only for both species. For
v
v
comparison, the constants for Co35Cl ( ϭ0 and ϭ1) de-
v
v
while the diagonal elements are
rived from the infrared measurements of Hirao, Pinchemel,
and Bernath13 are listed as well. These constants were estab-
lished from the ⍀ϭ4 and 3 components only, however, so
there is some ͑expected͒ disagreement between the two pa-
3
3
3
3
ˆ
ˆ
ˆ
ˆ
⌽
a l •s
⌽
ϭϪ
ʹ ͳ
⌽
a l •s
⌽
i 2
͚
ͯ
ͯ
͚
ͯ
4
ͯ
i i
i
4
2
i i
ͳ
ʹ
i
i
rameter sets. The rms of the fits for Co35Cl are 434 ( ϭ0)
a
␲
v
v
ϭa␦ϩ
.
and 373 kHz ( ϭ1), and 212 ( ϭ0) and 277 kHz ( ϭ1)
2
v
v
for the chlorine-37 isotopomer.
Considering 2␭ as the displacement due to the perturbation,
the difference in energy between the 1⌽3 and 3⌽3 states can
be expressed as
V. DISCUSSION
2
This study has resulted in the identification of the third
spin component ͑⍀ϭ2͒ for CoCl, which confirms the ⌽i
1
a
␲
2
3
⌬Eϭ
a Ϫ
␦
ϩ4␭
.
͑5͒
ͫ
ͩ
ͪ
ͬ
2␭
2
ground state assignment for CoCl. The lack of ␭ doubling in
any ⍀ ladder, which might have been present in a ⌸ or ⌬
state, is additional evidence for the ⌽ term. Also of note is
the large shift of the ⍀ϭ3 spin component relative to the
⍀ϭ4 and 2 sublevels, which manifests itself in a large ͑nega-
tive͒ value of ␭. Furthermore, the hyperfine structure of this
sublevel appears perturbed. Both effects, which will be dis-
cussed in the following sections, probably arise from inter-
Here it is assumed that the effective ␭ is ␭so.
The ␦ orbitals in CoCl must be centered on the cobalt
atom, as chlorine has no d orbitals. Hence, to a reasonable
24,25
approximation, 2a␦ϭ␨CoϭϪ536 cmϪ1
.
Because a␦
ϩa /2ϭ3A, it follows that a /2ϭϪ232 cmϪ1. The energy
␲
␲
1
3
separation between the ⌽3 and ⌽3 states is then estimated
to be ⌬EϷϪ480 cmϪ1. This calculation implies that the 1⌽
state lies within the ground state spin-orbit manifold. Be-
cause of the very large value of ␭, a close-lying perturbing
state is probably expected. Furthermore, the negative value
1
actions with an isoconfigurational excited ⌽ state, which
has a single spin ladder with ⍀ϭ3. Evidence of similar per-
turbations occur in the X 3⌬i state of FeC, where the ⍀ϭ2
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07

8390
TABLE II. Spectroscopic constants for CoCl (X 3⌽i) in MHz.a
ϭ0
J. Chem. Phys., Vol. 121, No. 17, 1 November 2004
Flory, Halfen, and Ziurys
ϭ1
v
v
Submillimeter
Co35Cl
Submillimeter
Co37Cl
Optical
CoClc,d
Submillimeter
Co35Cl
Submillimeter
Co37Cl
Optical
CoClc
B
D
5422.94͑85͒
0.010 73͑88͒
1.89͑36͒
5232.870͑21͒
0.005 407 9͑67͒
5375.79͑72͒
0.003 783͑51͒
5391.41͑47͒
0.007 88͑38͒
0.50͑11͒
5207.61͑41͒
0.007 15͑30͒
0.407͑82͒
5346.22͑75͒
0.003 847͑54͒
H (106)
L (1017)
A
3.56͑73͒
Ϫ5 000 000b
Ϫ3.68͑13͒
0.000 766͑25͒
1.89͑38͒
Ϫ4 070 000b
Ϫ3.413͑64͒
0.000 673͑11͒
Ϫ6 985 000
3.27͑21͒
Ϫ5 000 000b
Ϫ9.390͑46͒
0.000 656 0͑42͒
0.102͑37͒
Ϫ6 985 300͑2300͒
Ϫ6.47͑15͒
Ϫ5 000 000b
Ϫ8.047͑81͒
0.000 574 2͑27͒
0.066͑25͒
Ϫ7 009 700͑5000͒
Ϫ8.14͑33͒
Ϫ6 955 080͑120͒
Ϫ3.63͑22͒
AD
AH
A
␭
␭
␭
a
b
L (1013)
Ϫ7 195 700͑3300͒
Ϫ34.25͑63͒
Ϫ0.000 13͑10͒
512͑45͒
Ϫ5 791 400͑1900͒
Ϫ30.19͑32͒
Ϫ0.000 158͑42͒
530͑41͒
D
H
569͑56͒
256͑40͒
636͑72͒
219͑40͒
135͑10͒
148.6͑7.5͒
(bϩc)
h3D
h⍀ϭ4
Ϫ170͑140͒
Ϫ0.88͑20͒
1339.2͑3.2͒
Ϫ220͑120͒
Ϫ0.76͑15͒
Ϫ310͑180͒
Ϫ0.63͑23͒
Ϫ530͑220͒
Ϫ0.40͑25͒
1318͑16͒
rms
Be
0.434
5438.7
31.5
0.212
5245.5
25.3
0.373
0.277
␣
e
aNumbers in parentheses are errors to three standard deviations for the last significant figure.
bHeld fixed to value established from initial fitting ͑see text͒.
cFrom Hirao, Pinchemel, and Bernath13 based on ⍀ϭ4 and 3 data only.
dFrom Adam et al.,10 based on ⍀ϭ4 data only.
of the effective ␭ parameter suggests that the ⍀ϭ3 substate
is shifted to higher energy in this case. Hence, the proposed
1⌽ state lies lower in energy relative to the ⍀ϭ3 substate. In
Fig. 4, this energy ordering is shown. Note that the ⍀ϭ3
sublevel is shifted closer to the ⍀ϭ2 substate in energy.
Without this perturbation, it would lie at 3A in energy, or
near 500 cmϪ1. This proposed shift is supported by the simi-
lar intensities of spectra originating from these two levels.
h
⍀ϭ3ϭ1755 MHz. Moreover, the global fit required the in-
clusion of a h3D centrifugal distortion correction for the
⍀ϭ3 hyperfine interactions. This term likely accounts for
more than centrifugal distortion because h3D /h⍀ϭ3ϾD/B.
As discussed by Azuma et al.,24 the Fermi contact and spin
3
Another possible perturber is a ⌬3 state. The study of
Hirao, Pinchemel, and Bernath13 suggests that there is a ⌬i
3
state in CoCl lying about ϳ20 000 cmϪ1 above ground state.
Because it is inverted, the ⍀ϭ3 level would lie lowest in
3
energy. This ⌬i state would therefore interact significantly
3
with the ⌽3 sublevel, although the ⍀ϭ2 level would addi-
tionally be affected.
B. Hyperfine splitting
As discussed, a series of rotational levels in the ⍀ϭ3
ladder displays highly irregular hyperfine patterns from prob-
able mixing with a vibrational state. Evidence for second-
order spin-orbit perturbations is also found in what appears
to be the more regular hyperfine interactions. The hyperfine
splitting per ⍀ is described by the equation
h ϭa⌳ϩ bϩc͒⌺,
͑6͒
͑
⍀
where h⍀ is the hyperfine parameter for each sublevel. Be-
cause ⌺ varies from 1, 0 to Ϫ1 for ⍀ϭ4, 3, and 2, the overall
hyperfine splittings should uniformly increase or decrease
with ⍀ value; this pattern does not occur for CoCl. Further-
more, h⍀ϭ3 should be the average of h⍀ϭ4 and h⍀ϭ2 . How-
ever, the average for the ⍀ϭ4 and 2 ladders is 2513 MHz,
which is significantly different from that of ⍀ϭ3, which has
FIG. 4. A qualitative energy level diagram showing the relative positions of
the three spin components arising from the ground state, X 3⌽i , indicated
by ⍀. Also depicted are the first three excited vibrational levels of the ⍀ϭ4
1
substate, shown in gray scale. A low-lying excited electronic
⌽ state,
3
located at EϷ400–500 cmϪ1, may be responsible for perturbing the ⍀ϭ3
sublevel of the X 3⌽i state and shifting it to higher energy.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07

J. Chem. Phys., Vol. 121, No. 17, 1 November 2004
Rotational spectrum of CoCl
8391
dipolar hyperfine interactions have cross terms with the spin-
orbit matrix elements that involve both the 3⌽ and 1⌽ states.
The hyperfine parameters can also be used to establish
bonding characteristics of CoCl. For example, the a param-
eter, which has a value of 512 MHz, results from the I"L
interaction. Because the two unpaired electrons in CoCl are
predicted to be in ␲ and ␦ orbitals, they both contribute to
this term, and can account for its relatively large value. For
comparison, a(Co)ϭ604 MHz in CoF ͑Ref. 20͒ and a(Co)
ϭ621 MHz in CoH.22 Because a is inversely proportional to
the average distance, cubed, of the unpaired electrons from
the nucleus with the spin, or 1/r3 , a smaller value of this
of atomic cobalt. However, if chlorine does contribute to the
5␲ orbital, then the participation of cobalt is a factor less
than one. If cobalt’s contribution is expressed as the coeffi-
Co
cient c , then the above equations can be recast as20
5␲
Co
5␲
2ϩc
a Co͒Ϸ
a01
,
͑
ͫ
ͬ
3d
3
͑8͒
3 2ϪcCo
͒
͑
5␲
c Co͒Ϸ Ϫ
͑
a12
,
3d
ͫ
ͬ
14
͗
͘
parameter implies that the unpaired electrons are located fur-
␬k
where a3d is the hyperfine parameter for atomic cobalt. For
a031dϭ617.9 MHz and a312dϭ857.1 MHz,20 c5C␲o is estimated to
ther from the nucleus. In fact, 1/r3 ϭ2.7ϫ1031 mϪ3 for
͗
͘
CoCl, while this quantity is 3.3ϫ1031 mϪ3 in CoH ͑Ref. 22͒
be 0.486 and 0.339, from the a and c parameters, respec-
and 3.2ϫ1031 mϪ3 in CoF.21 The same quantity for the co-
Co
4␲
tively. In contrast, c for CoF was found to be 0.934. Ob-
22
balt atom is 4.4ϫ1031 mϪ3
.
Thus, the unpaired electrons
viously these calculations are only estimates, but they do
suggest that CoCl has more covalent character to its bonding
than CoF. Based on electronegativity arguments, such a re-
sult might be expected, although previous studies have indi-
cated that CoCl is as ionic as its fluorine analog.10,13
are located in more diffuse orbitals in CoCl than the other
species; the larger size of the electronegative chorine atom
relative to hydrogen and fluorine may be pushing the cobalt
electron density further from the Co nucleus.
The Fermi contact term bF primarily arises from elec-
trons in ␴ orbitals which, assuming a single electron configu-
ration, do not exist for CoCl. A nonzero value for bF can
arise from spin polarization, but then it is usually negative.26
Because (bϩc) is not well defined in CoCl, in turn neither is
bF . The value determined here for bF is 33Ϯ48 MHz—
which within the errors could be negative.
C. Bonding trends in 3d transition metal chlorides
Trends within the 3d transition metal oxides and sulfides
have been discussed in depth by various authors.29–31 The
so-called ‘‘double-hump’’ structure in bond length is appar-
ent in both these species, namely, there is an increase in the
bond distance occurring from vanadium to manganese, a de-
crease at iron, and then a second increase towards copper.
The increase from vanadium to manganese is due to the ad-
The dipolar constant c is calculated to be Ϫ305Ϯ140
MHz—roughly comparable to the c parameters for CoF
͑Ϫ196 MHz͒ and CoH ͑Ϫ456 MHz͒.20,22 This constant de-
pends on the expectation value of the operator ͚i (3 cos2 ␪
i
Ϫ1)/r3i , where i is the summation over the unpaired electrons.
From c, the angular factor 3 cos2 ␪Ϫ1 can be calculated for
*
dition of electrons to antibonding orbitals. The highest ␴
͗
͘
orbital does not receive an electron until manganese. After
this, bond lengths decrease due to core contraction of the
metal as no new antibonding orbitals are being filled; the
trend repeats in the second half of the third row.
CoCl, using 1/r3 derived from a; its value is Ϫ0.3901. For
͗
͘
pure d and d␦ orbitals, these angular factors are ϩ2/7 and
␲
Ϫ4/7, respectively,27 which add to Ϫ0.286. Hence, the sign
of c appears to be correct. The slightly larger angular value
may result from polarization of the ␲ and ␦ orbitals by the
nearby chlorine atom.
A somewhat different trend is found in the 3d halides.
As discussed by Sheridan, McLamarrah, and Ziurys,16 the
fluorides exhibit a noticeable increase in bond length at MnF,
but only a minor one from cobalt to copper. In addition,
another large increase in bond distance appears at titanium.
This trend is shown in Fig. 5. The difference in bond length
trends between the oxides/sulfides versus the fluorides arises
from the presence of an extra electron in the latter species, as
well as the difference in the energies of the F and O atomic
orbitals. As a result, the orbital energies vary, and the elec-
tron configurations differ significantly from the oxides to the
fluorides ͑see Ref. 17͒.
The hyperfine constants can be used as well to calculate
the percent contribution of the two atoms in CoCl to the
molecular orbitals of the unpaired electrons. The valence
electron configuration of CoCl is (11␴)2(1␦)3(5␲)3.10 The
␦ orbital must exclusively be created from cobalt 3d orbitals,
but the 5␲ orbital could also arise in part from Cl 3p. Be-
cause both ␲ and ␦ electrons contribute to the a and c, these
parameters can be expressed as28
1
N 3
a Co͒ϭ2␮ g ␮ 2 1/r3 ϩ 1/r3
,
͔
5␲
͑
͓
͗
͘
͗
͘
B
N
1␦
The bond lengths of the chlorides are also plotted in Fig.
5. At first glance, the chlorides and fluorides exhibit a similar
trend, as might be expected from their identical electron con-
figurations. The bond distances of the chlorides are naturally
longer than the fluorides by about 0.4 Å due to the difference
in atomic or ionic radii ͑0.35 or 0.45 Å͒.32 However, the
bond distances shorten slightly at CrCl and CuCl, while they
increase in the fluoride species. The decrease in bond dis-
tance from MnCl to CoCl is more dramatic than in the fluo-
rides, as well. These variations may result from the fact that
͑7͒
3
1
2
3 cos2 ␪Ϫ1
c Co͒ϭ gs␮BgN␮
͑
N
ͳ
ʹ
ͫ
r3
2
1␦
3 cos2 ␪Ϫ1
ϩ
.
ͳ
ʹ
ͬ
r3
5␲
If only cobalt contributed to a and c, then the expectation
values of 1/r3 and 3 cos2 ␪Ϫ1 could be replaced by those
͗
͘
͗
͘
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07

8392
J. Chem. Phys., Vol. 121, No. 17, 1 November 2004
Flory, Halfen, and Ziurys
1 D. Arnett, Supernovae and Nucleosynthesis ͑Princeton University Press,
Princeton, NJ, 1996͒.
2 D. Shriver and P. Atkins, Inorganic Chemistry, 3rd ed. ͑Freeman, New
York, 1999͒.
3 N. Koga and K. Morokuma, Chem. Rev. ͑Washington, D.C.͒ 91, 823
͑1991͒.
4 J. E. Huheey, E. A. Keiter, and R. L. Keiter, Inorganic Chemistry: Prin-
ciples of Structure and Reactivity, 4th ed. ͑Harper Collins, New York,
1993͒.
5 J. C. Noveron, M. S. Lah, R. E. Del Sesto, A. M. Arif, J. S. Miller, and P.
J. Stang, J. Am. Chem. Soc. 124, 6613 ͑2002͒.
6 R. S. Ram, P. F. Bernath, and S. P. Davis, J. Chem. Phys. 104, 6949
͑1996͒.
7 P. Mesnage, C. R. Acad. Bulg. Sci. 201, 389 ͑1935͒.
8 K. R. More, Phys. Rev. 54, 122 ͑1938͒.
9 A. J. Bridgeman, J. Chem. Soc. Dalton Trans. 1997, 4765.
10 A. G. Adam, J. R. D. Peers, Y. Teng, and C. Linton, J. Mol. Spectrosc.
212, 111 ͑2002͒.
FIG. 5. A graph showing the periodic trend in bond lengths for the 3d
transition metal chlorides vs the fluorides. The two series show very similar
characteristics. Slight deviations occur at Cr and Cu, however, where the
bond lengths decrease in the chlorides but increase in the fluorides. These
differences suggest subtle variations in bonding between the two sets of
halides.
11 A. G. Adam, L. P. Fraser, W. D. Hamilton, and M. C. Steeves, Chem.
Phys. Lett. 230, 82 ͑1994͒.
12 M. Freindorf, C. M. Marian, and B. A. Hess, J. Chem. Phys. 99, 1215
͑1993͒.
13 T. Hirao, B. Pinchemel, and P. F. Bernath, J. Mol. Spectrosc. 219, 119
͑2003͒.
14 A. L. Wong, W. S. Tam, and A. S.-C. Cheung, J. Chem. Phys. 119, 3234
͑2003͒.
15 L. M. Ziurys, W. L. Barclay, Jr., M. A. Anderson, D. A. Fletcher, and J. W.
Lamb, Rev. Sci. Instrum. 65, 1517 ͑1994͒.
the chlorine bonds from 3p orbitals, which are considerably
closer in energy to the 3d/4s transition metal orbitals than
the 2p of the F atom. Greater molecular orbital overlap in
the chlorides produces these subtle bonding differences.
16 P. M. Sheridan, S. K. McLamarrah, and L. M. Ziurys, J. Chem. Phys. 119,
9496 ͑2003͒.
17 M. D. Allen, B. Z. Li, and L. M. Ziurys, Chem. Phys. Lett. 270, 517
͑1997͒.
18 C. H. Townes and A. L. Schawlow, Microwave Spectroscopy ͑Dover, New
York, 1975͒.
VI. CONCLUSIONS
19 See EPAPS Document No. E-JCPSA6-121-021439 for a complete list of
transitions arising from both Co35Cl and Co37Cl. A direct link to this
document may be found in the online article’s HTML reference section.
The document may also be reached via the EPAPS homepage ͑http://
www.aip.org/pubservs/epaps.html͒ or from ftp.aip.org in the directory
/epaps/. See the EPAPS homepage for more information.
20 T. Okabayashi and M. Tanimoto, J. Mol. Spectrosc. 221, 149 ͑2003͒.
21 A. G. Adam and W. D. Hamilton, J. Mol. Spectrosc. 206, 139 ͑2001͒.
22 S. P. Beaton, K. M. Evenson, and J. M. Brown, J. Mol. Spectrosc. 164,
395 ͑1994͒.
Although the three spin-orbit components have been
identified in CoCl in its X 3⌽i state, the irregular pattern,
coupled with unusual hyperfine splittings, suggests that the
ground state is perturbed by a close-lying excited state or
states. Because the ⍀ϭ3 sublevels appear to exhibit the most
irregularity, the perturbing state may be the isoconfigura-
1
tional ⌽3 term. These states can interact via second-order
23 P. M. Sheridan, L. M. Ziurys, and T. Hirano, Astrophys. J. 593, L141
spin-orbit coupling, which enters in as an off-diagonal term
in both the spin and hyperfine Hamiltonians. Interpretation of
the hyperfine parameters suggests that the bonding in CoCl is
more covalent than in CoF. An examination of the bond dis-
tances of the 3d chlorides versus the fluorides shows that
these species are similar, although subtle differences in their
bonding appear to be present.
͑2003͒.
24
¨
Y. Azuma, J. A. Barry, M. P. J. Lyne, A. J. Merer, J. O. Schroder, and J.-L.
´
´
Femenias, J. Chem. Phys. 91, 1 ͑1989͒.
25 W. J. Balfour, A. J. Merer, H. Niki, B. Simard, and P. A. Hackett, J. Chem.
Phys. 99, 3288 ͑1993͒.
26 A. Carrington and A. D. McLachlan, Introduction to Magnetic Resonance
͑Harper and Row, New York, 1967͒.
27 D. J. Clouthier, G. Huang, A. J. Merer, and E. J. Friedman-Hill, J. Chem.
Phys. 99, 6336 ͑1993͒.
28 J. M. Brown and A. Carrington, Rotational Spectroscopy of Diatomic
Molecules ͑Cambridge University Press, Cambridge, 2003͒.
29 A. J. Merer, Annu. Rev. Phys. Chem. 40, 407 ͑1989͒.
ACKNOWLEDGMENTS
30 A. J. Bridgeman and J. Rothery, J. Chem. Soc. Dalton Trans. 2000, 211.
This research was supported by NSF Grant No. CHE-98-
17707. The authors would like to thank Dr. J. M. Brown for
the use of his Hamiltonian fitting program and Dr. Phillip M.
Sheridan for his comments.
31
¨
I. Kretzschmar, D. Schroder, H. Schwartz, and P. B. Armentrout, Advances
in Metal and Semiconductor Clusters ͑Elsevier, New York, 2001͒.
32 G. Bodner and H. Pardue, Chemistry: An Experimental Science ͑Wiley,
New York, 1995͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.18.123.11 On: Thu, 18 Dec 2014 01:13:07