4458
J. Chem. Phys., Vol. 114, No. 10, 8 March 2001
Ram, Bernath, and Davis
ing hydrides. For example, the structure of TiF23 and TiCl24
closely resembles that of TiH25 and the electronic structure
of HfCl9 is similar to that of HfH.26 The VF and VCl mol-
19
5
ecules probably have a ⌬ ground state like VH and we
assign the observed VCl bands to a ⌬–5⌬ transition. The
5
0–0 band of this transition is located near 7004 cmϪ1, and
we have, therefore, decided to label this transition as
5
7.0 ⌬–X 5⌬.
͓
͔
5
The ⌬ states are expected to display Hund’s case ͑a͒
5
5
coupling leading to five subbands, ⌬0Ϯ –5⌬0Ϯ , ⌬1 –5⌬1 ,
5⌬2 –5⌬2 , ⌬3 –5⌬3 , and ⌬4 –5⌬4 . Normally ⌳ doubling
is expected to be small for ⌬ states and have J dependence27
given by J(Jϩ1) ⍀. However, mixing of the ground X 5⌬
5
5
FIG. 1. A compressed portion of the 7.0 5⌬–X 5⌬ transition of VC1.
͓
͔
͓
͔
5
perature of about 2200 °C. The emission from the furnace
was focused on to the entrance aperture of the 1 m Fourier
transform spectrometer of the National Solar Observatory at
Kitt Peak and the spectra were recorded using an InSb de-
tector and a RG 645 filter at a resolution of 0.05 cmϪ1. The
spectrum was recorded with only a single scan in about 3
min. In spite of this short integration time, the spectrum was
observed with sufficient signal-to-noise ratio for high resolu-
tion analysis.
A number of new bands observed in the 6000–8000
cmϪ1 region were readily attributed to VCl based on their
vibrational intervals and rotational line spacing. The spectral
line positions were measured using a data reduction program
called PC-DECOMP developed by Brault. The peak positions
were determined by fitting a Voigt line shape function to
each experimental feature. The observed spectra also con-
tained the vibration–rotation bands of HCl and HF as impu-
rities in addition to the VCl bands and V atomic lines. We
have used the HF22 line positions to calibrate our spectrum.
The molecular lines appear with a width of 0.070 cmϪ1 and
maximum signal-to noise ratio of about 8:1; the line wave
numbers are expected to be accurate to Ϯ0.007 cmϪ1. How-
ever, there is considerable overlapping and blending due to
the rotational structure of different subbands in the same re-
gion, so the uncertainty is somewhat higher for blended and
weaker lines.
state with the nearby ⌸ state ͓i.e., some Hund’s case ͑c͒
behavior͔ could lead to observable ⌳ doubling ͑or more cor-
rectly ⍀ doubling͒. The effects of ⌳ doubling should then be
apparent in the 5⌬1 –5⌬1 and 5⌬2 –5⌬2 subbands. In the
5⌬1 –5⌬1 subband the ⌳ doubling should increase as J (JH)
J, similar to that found in a normal ⌸–1⌸ transition. The
1
observed rotational structure of this subband is consistent
5
with this expectation. In the ⌬2 –5⌬2 subband the ⌳ dou-
2
bling, if present, should increase as J(Jϩ1) similar to a
͓
͔
3⌸2 –3⌸2 transition. The observed rotational structure of this
subband is also consistent with this expectation and a very
small doubling is seen at very high J values (JϾ63).
The observed pattern of ⌳ doubling in these two sub-
bands has been very helpful in the ⍀ assignment of different
subbands. Surprisingly some small ⍀ doubling has also been
5
observed in the ⌬3 –5⌬3 subband for JϾ82. This observa-
tion is consistent with some Hund’s case ͑c͒ behavior in the
lower state as discussed in the following. We have analyzed
5
the rotational structure of only the 5⌬1 –5⌬1 , ⌬2 –5⌬2 , and
5⌬3 –5⌬3 subbands in the 0–0 and 0–1 bands. The rotational
analysis of 0–2, 1–0, and 2–0 bands could not be achieved
because they were weak. The rotational structure of each
subband consists of R and P branches ͑no Q branch͒ appear-
ing with similar intensity indicating a ⌬⍀ϭ0 assignment.
The rotational constants for the different ⍀ states have been
obtained by fitting the observed lines to the following energy
level expression:
RESULTS AND DISCUSSION
2
Fv J͒ϭT ϩB J Jϩ1͒ϪD J Jϩ1͒
͓ ͑ ͔
͑
͑
v
v
v
A compressed portion of the observed spectrum is pre-
sented in Fig. 1. The bandheads observed near 6176, 6589,
7004, 7358, and 7710 cmϪ1 can readily be identified as the
0–2, 0–1, 0–0, 1–0, and 2–0 bands of a new electronic
transition. This transition has been assigned as a 5⌬–5⌬
transition. As seen in Fig. 1, the spectrum of each band is
very complex because of the crowded overlapping of the
rotational structure of different subbands. In fact, the over-
lapping is so severe that the rotational structure of all five
subbands could not be unambiguously identified. The recent
theoretical predictions for VH by Bruna19 were very helpful
in assigning our electronic transition. The ground state of
VH is predicted to be a 5⌬ state and a strong 5⌬–X 5⌬
transition has been predicted near 10 650 cmϪ1. Indeed, we
find a strong complex transition (5⌬–5⌬?) of VH20 near
7400 cmϪ1. In our previous studies of transition metal ha-
lides we have noted the similarity between the electronic
energy levels of transition metal halides and the correspond-
ϩHv J Jϩ1͒ 3Ϯ1/2 q J Jϩ1͒
͓ ͑
͔
͑
͕
v
3
ϩqD J Jϩ1͒ 2ϩqH J Jϩ1͒
͓ ͑ ͓ ͑
.
͔
͖
͑1͒
͔
v
v
The rotational lines were weighted according to resolu-
tion and the extent of blending. The badly blended lines were
heavily deweighted. The observed lines positions for the dif-
ferent subbands are available from EPAPS28 or from the au-
thors upon request. The molecular constants for the different
bands are provided in Table I. The e/f parity assignment in
the 5⌬1 –5⌬1 subband was made arbitrarily to provide a
positive ⌳-doubling constant qv in the X 5⌬1 spin compo-
5
5
nent. The parity assignments in the ⌬2 –5⌬2 and ⌬3 –5⌬3
subbands were also chosen arbitrarily. The higher order
⍀-doubling constant pD was also determined for the ϭ0
v
v
and 1 vibrational levels of the X 5⌬1 spin component. Like-
wise pD and pH were determined for the ϭ0 vibrational
v
v
v
On: Sun, 21 Dec 2014 18:09:34