The pure rotational spectrum of CoO(X4Δi): Identifying the high-spin components

Article abstract of DOI:10.1016/j.cplett.2005.07.106

New measurements of the pure rotational spectrum of the CoO radical in its X⁴Δ_i state (v = 0 and v = 1) have been conducted using millimeter-wave direct absorption techniques. CoO was created by the reaction of cobalt vapor and N₂O. Spectra arising from the Ω = 3/2 and 1/2 spin levels, both which exhibit Λ-doubling, have been recorded, as well as from the Ω = 7/2 and 5/2 ladders, which have been studied in the past. Rotational, fine structure, Λ-doubling and hyperfine parameters for the v = 0 and v = 1 states have been established from a global analysis of all four spin components. The data suggest a near-by ⁴Σ state.

Full text of DOI:10.1016/j.cplett.2005.07.106

Chemical Physics Letters 414 (2005) 301–306
www.elsevier.com/locate/cplett
4
The pure rotational spectrum of CoO(X D ):
i
Identifying the high-spin components
1
*
S.K. McLamarrah, P.M. Sheridan , L.M. Ziurys
Department of Chemistry, Department of Astronomy and Steward Observatory, 933 North Cherry Avenue, University of Arizona,
Tucson, AZ 85721, USA
Received 28 March 2005; in final form 13 July 2005
Available online 9 September 2005
Abstract
4
New measurements of the pure rotational spectrum of the CoO radical in its X D
i
state (v = 0 and v = 1) have been conducted
O. Spectra arising from
using millimeter-wave direct absorption techniques. CoO was created by the reaction of cobalt vapor and N
2
the X = 3/2 and 1/2 spin levels, both which exhibit K-doubling, have been recorded, as well as from the X = 7/2 and 5/2 ladders,
which have been studied in the past. Rotational, fine structure, K-doubling and hyperfine parameters for the v = 0 and v = 1 states
4
have been established from a global analysis of all four spin components. The data suggest a near-by R state.
Ó 2005 Elsevier B.V. All rights reserved.
4
1
. Introduction
lated to be a D ground state. Subsequently, Clouthier
i
et al. [7] conducted an analysis of the hyperfine structure
of these two spin components using laser spectroscopy,
and the A P –X D transition in the near-infrared was
i i
recorded by Ram et al. [8]. The latter work included
rotational analyses of the 5/2–7/2 and 3/2–5/2 sub-
bands, but the higher energy spin–orbit transitions
could not be assigned. Recently, Barnes et al. [9]
Establishing the ground electronic states of molecules
4
4
containing transition metals is useful in understanding
their bonding characteristics [1]. Because of the presence
of d electrons, many electron configurations are possi-
ble, which in turn generate a variety of electronic terms.
Ab initio calculations are not necessarily conclusive in
determining the electronic ground states of these species
and experiments can be misleading. Cobalt oxide, CoO,
is a good example.
measured over 100 bands of CoO originating from the
4
X D
spin component using beam techniques, while
/2
7
Namiki and Saito [10] recorded the first pure-rotational
spectrum of CoO in its X = 7/2 and 5/2 sub-levels.
4
ꢀ
Earlier work on CoO suggested a
R
ground
electronic state (e.g. [2,3]), as was supported by several
ab initio studies [4,5]. The first gas-phase high resolution
investigation by Adam et al. in 1985 [6], identified
X = 5/2 and 7/2 spin components in what was postu-
Subsequent ab initio studies have suggested the ground
4
state assignment as D [11,12].
i
Here, new measurements of the pure rotational spec-
4
trum of CoO in its X D state are presented. Transitions
i
arising from all four spin components have been
recorded for the v = 0 and v = 1 levels – the first
4
complete rotational study of a molecule in a D state.
*
Corresponding author. Fax: +1 520 621 1532.
E-mail address: lziurys@as.arizona.edu (L.M. Ziurys).
Present address: Department of Chemistry, University of Water-
Improved spectroscopic constants have been derived
for CoO and an interpretation of the K-doubling and
hyperfine parameters is given.
1
loo, 200 University Avenue West, Waterloo, Ont., Canada N2L 3G1.
0
009-2614/$ - see front matter Ó 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2005.07.106

302
S.K. McLamarrah et al. / Chemical Physics Letters 414 (2005) 301–306
2
. Experimental
v = 1 state, shifted ꢁ4 GHz to lower frequency. In both
vibrational states, the intensity of each spin-component
diminishes with decreasing X value, consistent with an
inverted term.
The rotational spectrum of CoO was measured
using the high temperature millimeter-wave spectrome-
ter of the Ziurys group [13]. Briefly, the instrument is
composed of a radiation source, a reaction chamber,
and a detector. Phase-locked Gunn oscillators with
Schottky diode multipliers provide frequency coverage
over the range 65–650 GHz. The radiation is propa-
gated quasi-optically through the reaction cell, which
is a double-pass system, where it is subsequently de-
tected by an InSb bolometer. The source is FM-mod-
ulated and signals are processed by a lock-in-amplifier
at 2f.
In Fig. 2, laboratory spectra of the X = 3/2 and 1/2
spin–orbit components of the J = 12.5
11.5 transition
of CoO (v = 0) are shown. The X = 3/2 data (a) consist
of 16 resolved features. The larger splitting, indicated by
the F quantum number, arises from the cobalt nuclear
spin. Each hyperfine feature is further split by
ꢁ2 MHz as a result of K-doubling. In the case of the
X = 1/2 component (b), the doubling is much larger
(ꢁ300 MHz) and necessitates a frequency break in the
spectrum. The individual doublets, labeled by e and f,
are in turn split into hyperfine octets.
CoO was created by the reaction of N O with cobalt
2
vapor. The vapor was produced by heating solid cobalt
in a Broida-type oven. Approximately 15–20 mTorr of
The spectra of CoO were fit to the following effective
Hamiltonian [14–16]:
N O was then added to the cell from beneath the oven.
A dc discharge was not found necessary for the
reaction.
2
ð3Þ
H
eff ¼Hrot þHso þHss þHld þHhf þH þHeqQ
.
ð1Þ
so
Final frequency measurements were obtained by
averaging 5 MHz scans, one in increasing and one in
decreasing frequency. Averages of up to three scan pairs
were required for the weaker features. Center frequen-
cies were established by fitting each line to a Gaussian
profile. Typical line-widths varied from 1000 kHz at
This expression consists of rotational, spin–orbit, spin–
spin, K-doubling, magnetic hyperfine, and quadrupole
terms, as well as the third-order spin–orbit interaction.
The v = 0 and v = 1 data sets were fit separately.
Unlike the past rotational analysis, A, the spin–orbit
constant, k, the spin–spin parameter, and g, the third-
order spin–orbit correction, could all be fit indepen-
dently. The value of A was found to be nearly identical
3
75 GHz to 1400 kHz at 525 GHz.
ꢀ
1
to that derived from optical data (163 vs. 166 cm : [7]).
3
. Results and analysis
In principle, four constants characterize K-doubling in
4
D states: ~n , ~o , ~p , and ~q [16]. In the case of CoO,
D
D
D
D
Using predictions for CoO based on Namiki and
however, this splitting is sufficiently small such that only
Saito [10], a search was conducted for the X = 3/2 and
/2 spin–orbit components, which resulted in the
observation of several sets of harmonically related octets
Co: I = 7/2). The first group consisted of two sets of
~n and ~o
D
D
could be established. The parity assignment
1
was arbitrarily set; switching the parity did not change
the fit other than to reverse the signs of the K-doubling
constants. For the hyperfine structure, the Frosch and
Foley parameters a, b, and b + c were fit, as well as
two centrifugal distortion corrections bD and (b + c)D.
(
octets, separated by ꢁ5–10 MHz in frequency. These
features were assigned to the X = 3/2 spin-component
with K-doubling. In the second group, two distinct
octets were also observed with a much larger separation
of ꢁ300 MHz. These lines were attributed to the K-dou-
blets of the X = 1/2 sub-level. A further search of ꢁ6B
revealed no new spectral features, except for those
arising from the first-vibrationally excited state (v = 1).
Five complete rotational transitions that included all
four spin–orbit components were recorded for the v = 0
state of CoO – a total of 213 new features. Three rota-
tional transitions were measured for the v = 1 state; only
one of these included all four spin levels. A subset of
these data is presented in Table 1. The remainder can
be found at http://www.chem.arizona.edu/faculty/ziur/
ziur-group.html.
Both b , the third-order spin–orbit distortion term to
s
the Fermi contact interaction [17], and d , used to ac-
count for differences in hf splittings between lambda-
doublets, did not appear to influence the fit and hence
D
were not used. It is interesting to note that b was no
s
longer determinable using the full data set. The v = 1
data were analyzed in an identical manner except A
and eqQ were held fixed to the values established from
the v = 0 fit.
The results of these analyses are given in Table 2. For
the v = 0 data, one set of constants was established by
fitting the data obtained in this work only, and another
set was determined from a combined fit with the spectra
of Namiki and Saito [10]. Also given in the table are the
previous constants from Namiki and Saito, which in-
cluded spectra from the X = 7/2 and 5/2 ladders only.
The constants are in reasonable agreement, given the
difference in data sets.
A stick spectrum of the J = 15.5
14.5 rotational
transition of CoO is presented in Fig. 1. The four
spin-components of the ground state are shown with
approximate experimental intensities, as well as the

S.K. McLamarrah et al. / Chemical Physics Letters 414 (2005) 301–306
303
Table 1
Selected transition frequencies of CoO(X D
4
a
i
)
J + 1
J
F + 1
F
X = 3.5
X = 2.5
Par
X = 1.5
X = 0.5
m
obs
m
o ꢀc
m
obs
m
o ꢀc
m
obs
m
oꢀc
m
obs
m
o ꢀc
v = 0
12.5 11.5
9
8
375120.075
375108.829
375094.523
375077.141
375056.700
375033.238
375006.634
374977.042
0.218
0.155
0.108
0.056
0.014
0.020
376835.209
376825.068
376811.965
376795.823
376776.701
376754.580
376729.525
376701.433
ꢀ0.018
0.004
e
f
e
f
e
f
e
f
e
f
e
f
e
f
e
f
378030.779
378032.965
378023.640
378025.913
378014.229
378016.482
378002.606
378004.880
377988.696
377990.900
377972.570
377974.756
377954.242
377956.457
377933.770
377936.015
ꢀ0.164
ꢀ0.101
ꢀ0.052
0.096
0.011
0.138
0.074
0.221
0.052
0.127
379046.408
379350.973
379044.509
379348.961
379041.916
379346.332
379038.394
379342.804
379034.047
379338.567
379029.154
379333.623
379023.414
379328.014
379017.108
379321.724
ꢀ0.156
ꢀ0.100
ꢀ0.054
ꢀ0.141
0.137
10
11
12
13
14
15
16
9
10
0.048
ꢀ0.016
11
0.033
0.154
ꢀ0.030
0.073
b
12
13
14
15
0.015
ꢀ0.017
ꢀ0.023
ꢀ0.015
ꢀ0.059
0.006
0.061
0.141
0.001
0.024
0.043
ꢀ0.043
ꢀ0.018
ꢀ0.058
0.024
ꢀ0.089
0.021
ꢀ0.030
0.068
16.5 15.5
13
12
13
14
15
16
17
18
19
494945.902
494937.861
494928.492
494917.800
494905.802
494892.468
494877.816
494861.829
ꢀ0.139
ꢀ0.110
ꢀ0.091
ꢀ0.071
ꢀ0.025
0.025
497207.128
497200.108
497191.866
497182.256
497171.403
497159.306
497145.832
497131.084
ꢀ0.049
ꢀ0.034
0.031
0.006
0.024
0.091
0.085
0.121
f
e
f
e
f
e
f
e
f
e
f
e
f
e
f
498800.214
498796.030
498796.030
498791.953
498789.743
498786.169
498783.305
498779.629
498775.798
498772.001
498767.240
498763.464
498757.754
498754.022
498747.278
498743.570
ꢀ0.153
ꢀ0.645
0.420
c
c
14
15
16
17
18
19
20
0.037
ꢀ0.173
ꢀ0.051
0.022
500511.008
500208.043
500509.012
500206.033
500506.897
500203.910
500504.586
500201.556
500502.049
0.424
0.440
0.098
ꢀ0.044
0.092
ꢀ0.005
0.060
0.104
ꢀ0.041
ꢀ0.048
ꢀ0.067
ꢀ0.136
ꢀ0.006
ꢀ0.013
0.106
0.057
0.031
0.029
0.029
0.214
e
v = 1
5.5 14.5
1
12
11
12
13
14
15
16
17
18
461320.356
461311.702
461301.304
461289.402
461275.975
461260.893
461244.219
461225.920
0.156
0.144
463424.609
463417.050
463407.780
463397.009
463384.636
463370.844
463355.181
463338.340
ꢀ0.007
0.099
f
e
f
e
f
e
f
e
f
e
f
e
f
e
f
464898.770
464895.528
464893.226
464890.124
464887.213
464883.757
464879.477
464876.156
464870.845
464867.605
464861.052
464857.603
464849.980
464846.797
464837.812
464834.548
0.028
0.041
ꢀ0.241
ꢀ0.086
0.159
ꢀ0.038
ꢀ0.025
ꢀ0.085
0.036
0.060
0.081
ꢀ0.101
ꢀ0.001
0.085
13
14
15
16
17
18
19
ꢀ0.011
ꢀ0.064
ꢀ0.031
ꢀ0.033
0.002
0.039
0.025
ꢀ0.036
0.045
466497.155
466180.333
466494.805
466177.832
466491.755
466174.849
466488.979
466171.639
ꢀ0.069
0.145
0.035
0.092
ꢀ0.173
0.014
ꢀ0.192
ꢀ0.089
0.226
ꢀ0.148
0.052
ꢀ0.021
e
ꢀ0.014
a
In MHz; residuals from combined fit.
From [10].
b
c
Unresolved lines, not included in the least squares fit.
4
. Discussion
Because all four X ladders have been measured in
in Table 2. The bond distance r was calculated to be
e
˚
1
.627862(1) A. This value varies slightly from that of
˚
Ram et al. [8], who found r = 1.628626(32) A. Our
e
both the v = 0 and v = 1 levels, equilibrium constants
can be obtained for CoO. These parameters are listed
bond length agrees well with the equilibrium value
derived from computational methods. Dolg et al. [11],

304
S.K. McLamarrah et al. / Chemical Physics Letters 414 (2005) 301–306
4
CoO (X ∆ ): J = 12.5 11.5
a
i
1
6
15
Ω = 3/2
F =
5
1
4
13 13 12
1
14
e
11
11
10
10
1
2
f
9
9
8
377.945
377.965
377.985
378.005
378.025
b
f
Ω = 1/2
Fig. 1. A stick figure of the J = 15.5 14.5 rotational transition of
4
CoO(X D
i
), showing line positions and approximate relative intensities
e
of the four spin components in the v = 0 and v = 1 states. The pattern
is fairly regular, with K-doubling present in the X = 3/2 and 1/2 sub-
levels.
˚
for example, calculated r = l.637 and 1.623 A for
e
CI(SD) and CI(SD) + Q levels of theory, respectively.
˚
In contrast, a shorter bond length of 1.599 A was
4
ꢀ
obtained for the proposed R state. Bauschlicher and
˚
Maitre [12] determined r = 1.6208 A using CASSCF
e
methods.
The K-doubling constant ~n
is in the diagonal matrix
is off diago-
379.00
379.05
379.10
379.30
379.35
D
element for the X = 1/2 sub-level, while ~o
Frequency (GHz)
D
^
nal for the h3=2jH j1=2ielement [16]. In contrast, the
LD
4
Fig. 2. Spectra recorded for CoO(X D
i
) in its X = 3/2 and 1/2 spin
~
p term is in the diagonal element for the X = 3/2 spin
D
ladders. In (a) (X = 3/2), the eight hyperfine components arising from
the cobalt nuclear spin, labeled by quantum number F, are split again
into doublets indicated by e and f, as a result of K-doubling. In (b)
component. Based on the values of ~n and ~o , however,
p is estimated to be <0.001 MHz, and hence could not
be confidently fit. (The error on ~o
Consequently, it is odd that Ram et al. [8] were able
D
D
~
D
(
X = 1/2 data), the K-doubling is much larger such that two distinct
sets of octets are present, separated by over 300 MHz. (a) is a single,
20-MHz wide scan of about 1 min in duration, while (b) is a
composite of two such scans, with a frequency break of ꢁ150 MHz.
D
is 0.0082 MHz.)
1
to establish ~q ¼q ¼0.0181 MHz and ~p ¼p þ4q
D
D
D
D
D
¼
3.646 MHz especially since they only analyzed
X = 5/2 and 7/2 data of the ground state.
4
4
Both excited P and R states contribute to the
4
4
K-doubling in a D state. Assuming the A P state is
the main P contributor in CoO, the energy of the
ꢀ
1
4
ꢀ1
(
ꢀ163 cm ) and A P states (ꢀ152 cm ) in CoO. None
+
4
of these values agree with the Co spin–orbit constant of
principal perturbing R state can be estimated from
ꢀ
1
ꢀ
536 cm [18], suggesting that cobalt oxide is not en-
tirely represented by a Co O configuration.
the K-doubling constants using the expressions [16],
+
ꢀ
pffiffi
3
ꢀ1
ꢀ
24 5a B
Using a ꢂ163 cm
,
and assuming E ꢀE =
P
D
~
n ffi
;
ð2Þ
ð3Þ
D
2
4
ꢀ1
ðE ꢀE ÞðE ꢀE Þ
E(A P) ꢂ5500 cm [8], the K-doubling constants im-
4 ꢀ ꢀ1
D
P
D
R
2
2
ply that a R state lies ꢁ2200–3100 cm above ground
ꢀ
ꢀE
72a B
~
o ffi
.
D
state. This energy is remarkably close to a low-lying
2
ðE
D
P
ÞðE
D
ꢀE
R
Þ
ꢀ1
state at 3038 cm identified by Barnes et al. [9]. These
authors postulate that this state is the B R term arising
from a r d p electron configuration. In order for it to
contribute to K-doubling, however, the state must have
a different electron configuration to connect with
4
ꢀ
Here, a is the spin–orbit constant associated with the
P
1
4 2
^
one electron operator
a l ^s , and B is the rotational
i
i
i
i
constant. Equating the denominators of these
expressions and using the values of ~n
and ~o
deter-
D
D
ꢀ
1
2 2
3
4
1 2 4
mined here yields a ꢂꢀ226 cm . This agrees well with
the r d p A P state. An alternative is a r d p
configuration.
4
the spin–orbit constants established for the X D state

S.K. McLamarrah et al. / Chemical Physics Letters 414 (2005) 301–306
305
Table 2
Spectroscopic parameters for CoO(X D
4
a
i
)
b
Parameter
This work (v = 0)
Combined fit (v = 0)
This work (v = 1)
Previous work (v = 0: X = 7/2 and 5/2)
B
D
A
15096.3673(38)
0.0209598(68)
ꢀ4881000(78000)
ꢀ4.13(37)
15096.3664(29)
0.0209573(59)
ꢀ4884000(62000)
ꢀ4.14(29)
14976.798(10)
0.021051(18)
15096.8969 (14)
0.0209695(36)
c
c
ꢀ4884000
ꢀ4982900
A
k
D
ꢀ4.1825(12)
201963(67)
49799(33)
ꢀ53.370(36)
0.1506(63)
686.4(6.3)
ꢀ314(34)
ꢀ4.73853(43)
c
199800(6000)
48700(1400)
ꢀ51.103(16)
0.1524(82)
200000(4700)
48800(1100)
ꢀ51.103(17)
0.1527(76)
219400
35624
c
g
~n D
~
oD
a
b
c
685.7(2.1)
ꢀ292(12)
685.9(1.2)
ꢀ297(10)
649.1(1.3)
ꢀ135.9(2.2)
ꢀ274.2(2.3)
(
b + c)
ꢀ323.9(3.0)
ꢀ0.0770(98)
0.200(14)
ꢀ328.9(1.5)
ꢀ0.0751(74)
0.1802(89)
ꢀ327(10)
ꢀ0.078(25)
0.217(39)
b
D
(
b + c)
D
b
s
ꢀ9.41(45)
ꢀ37.9(1.5)
0.021
c
eqQ
rms
ꢀ60(18)
ꢀ52.4(7.4)
ꢀ52.4
0.095
0.104
0.096
B
e
15156.151(10)
119.569(10)
a
e
˚
r
e
1.627862(1) A
a
b
c
In MHz; errors are 3r and apply to the last quoted decimal places.
From [10].
Held fixed.
The proposed electron configuration for CoO in its
Acknowledgments
2
3
2
ground state is 9r 1d 4p . The a hyperfine constant,
therefore, must originate solely from the single unpaired
d electron. Because this orbital arises exclusively from
the cobalt atom, the a parameter should reflect the
This research was supported by NSF Grant 04-11551.
The authors thank Professor J.M. Brown for use of his
Hamiltonian code.
01
atomic hyperfine constant of a3
¼617.9 MHz [19];
dðCoÞ
a = 685.7 (2.1) MHz for CoO, which is in relatively
good agreement.
References
Both p and the d electrons contribute to the c hyperfine
constant. Because
c
is directly proportional to
[1] A.J. Bridgeman, J. Rothery, J. Chem. Soc., Dalton Trans. (2000)
211.
2
3
h3cos h ꢀ1=r i, angular factors play a major role in
i
i
[
[
2] T.C. DeVore, T.N. Gallaher, J. Chem. Phys. 71 (1979) 474.
3] D.W. Green, G.T. Reedy, J.G. Kay, J. Mol. Spectrosc. 78 (1979)
determining its value. For the 3dp and 3dd orbitals,
2
h3cos h ꢀ1i¼2=7 and ꢀ4=7, respectively [7], and thus
i
2
57.
they should effectively cancel each other. The c constant
should therefore be small; we measure c = ꢀ32 (12) MHz.
[4] M. Krauss, W.J. Stevens, J. Chem. Phys. 82 (1985)
5584.
The Fermi contact term is b = ꢀ303 (12) MHz. No s
[5] J. Piechota, M. Suffczynski, Phys. Rev. A 48 (1993) 2679.
F
[
6] A.G. Adam, Y. Azuma, J.A. Barry, G. Huang, M.P.J. Lyne, A.J.
Merer, J.O. Schr o¨ der, J. Chem. Phys. 86 (1987) 5231.
7] D.J. Clouthier, G. Huang, A.J. Merer, E.J. Friedman-Hill, J.
Chem. Phys. 99 (1993) 6336.
electrons are present in unfilled r orbitals in CoO, and
therefore spin polarization must be responsible for this
interaction. Usually spin polarization results in a small,
[
negative value of b [20]; while the term established here
[8] R.S. Ram, C.N. Jarman, P.F. Bernath, J. Mol. Spectrosc. 160
(1993) 574.
F
is negative, it is not particularly small. However, three
unpaired electrons are available in CoO to exchange
with the r bonding electrons. As demonstrated in other
molecules, multiple unpaired electrons can make a sig-
nificant contribution to spin polarization. In CrH, for
example, there is one unpaired electron in a r orbital.
[
9] M. Barnes, D.J. Clouthier, P.G. Hajigeorgiou, G. Huang, C.T.
Kingston, A.J. Merer, G.F. Metha, J.R.D. Peers, S.J. Rixon, J.
Mol. Spectrosc. 186 (1997) 374.
[
10] K.C. Namiki, S. Saito, J. Chem. Phys. 114 (2001) 9390.
[11] M. Dolg, U. Wedig, H. Stoll, H. Preuss, J. Chem. Phys. 86 (1987)
2123.
[
12] C.W. Bauschlicher Jr., P. Maitre, Theor. Chim. Acta 90 (1995)
89.
Its positive contribution to b is negated by the effects
F
1
of spin polarization due to four unpaired p and d elec-
[
13] L.M. Ziurys, W.L. Barclay Jr., M.A. Anderson, D.A. Fletcher,
J.W. Lamb, Rev. Sci. Instrum. 65 (1994) 1517.
trons [21]. The net result is a negative value of b . The
F
large negative value of b in CoO probably results from
the same phenomenon.
[14] J.M. Brown, E.A. Colbourn, J.K.G. Watson, F.D. Wayne, J.
Mol. Spectrosc. 74 (1979) 294.
F

306
S.K. McLamarrah et al. / Chemical Physics Letters 414 (2005) 301–306
[
[
[
15] J.M. Brown, D.J. Milton, J.K.G. Watson, R.N. Zare, D.L.
Albritton, M. Horani, J. Rostas, J. Mol. Spectrosc. 90 (1981) 139.
16] J.M. Brown, A.S.-C. Cheung, A.J. Merer, J. Mol. Spectrosc. 124
[18] H. Lefebvre-Brion, R.W. Field, Perturbations in the Spectra of
Diatomic Molecules, Academic Press, Orlando, 1986.
[19] G.H. Guth o¨ hrlein, H.P. Keller, Z. Phys. 17 (1990) 181.
[20] J.R. Morton, Chem. Rev. 64 (1964) 453.
(1987) 464.
17] A.S-C. Cheung, A.J. Merer, Mol Phys. 46 (1982) 111.
[21] D.T. Halfen, L.M. Ziurys, Astrophys. J. 611 (2004) L65.