Resonance Raman spectroscopy in the dissociative A band of nitrosyl chloride

Article abstract of DOI:10.1063/1.1355656

Resonance Raman spectra were obtained for CINO at three wavelengths in the ultraviolet A band. Twelve previously unobserved vibrational levels in the ground state were observed. These new levels, together with previously known vibrational levels, allowed the construction of a large amplitude vibrational model valid in and beyond the Franck-Condon region. The ground state surface showed a large degree of separability in Jacobi coordinates even in the triatomic molecular region.

Full text of DOI:10.1063/1.1355656

Resonance Raman spectroscopy in the dissociative A band of nitrosyl chloride
Jeffrey L. Mackey, Bruce R. Johnson, Carter Kittrell, Linh D. Le, and James L. Kinsey
Citation: The Journal of Chemical Physics 114, 6631 (2001); doi: 10.1063/1.1355656
View online: http://dx.doi.org/10.1063/1.1355656
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/114/15?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Photodissociation of N2O: Potential energy surfaces and absorption spectrum
J. Chem. Phys. 134, 064313 (2011); 10.1063/1.3553377
An ab initio study of the C H 3 I photodissociation. I. Potential energy surfaces
J. Chem. Phys. 126, 234102 (2007); 10.1063/1.2736695
Ab initio potential energy surfaces, total absorption cross sections, and product quantum state distributions for
the low-lying electronic states of N 2 O
J. Chem. Phys. 122, 054305 (2005); 10.1063/1.1830436
On the ultraviolet photodissociation of H 2 Te
J. Chem. Phys. 121, 9389 (2004); 10.1063/1.1799572
Emission spectroscopy of photodissociating N 2 O 4 excited near 200 nm to the π nb,O π NO 2 * /nσ N–N *
avoided crossing
J. Chem. Phys. 111, 8486 (1999); 10.1063/1.480189
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

JOURNAL OF CHEMICAL PHYSICS
VOLUME 114, NUMBER 15
15 APRIL 2001
Resonance Raman spectroscopy in the dissociative A band
of nitrosyl chloride
Jeffrey L. Mackey,a) Bruce R. Johnson, Carter Kittrell, Linh D. Le,b) and James L. Kinsey
Department of Chemistry and Rice Quantum Institute, Rice University, MS 600, Houston,
Texas 77251-1892
͑Received 30 November 2000; accepted 24 January 2001͒
Resonance Raman spectra measured for ClNO photoexcited at 212.5, 219, and 222 nm produce
evidence that the strong transition at the vacuum ultraviolet end of the merged A band system
dominates to wavelengths as long as 222 nm. The spectral resolution is sufficient to enable definitive
assignments of excited vibrational levels in the ground electronic state, several of which have not
been previously observed. A curvilinear coordinate model of the ground state potential surface
around the Franck–Condon region has been constructed. This model yields vibrational eigenvalues
in excellent agreement with all known transitions. © 2001 American Institute of Physics.
͓DOI: 10.1063/1.1355656͔
I. INTRODUCTION
results published to date come from a low resolution study at
an excitation wavelength of 266 nm by Bell and Frey.19 At
this far red edge of the A band the absorption cross section
has dropped to 1.4ϫ10Ϫ19 cm2, ϳ0.2% the value at band
peak. These experiments revealed a sequence of about 12
vibrational features out to a Raman shift of ϳ3400 cmϪ1, the
isolated ␯ peak at 332 cmϪ1 being the most intense. The ␯
Ultraviolet photoexcitation of nitrosyl chloride ͑ClNO͒
in the A band (␭ϳ180–250 nm) produces Cl and NO
photofragments.1 ͑The band was named before current nam-
ing conventions were established, and so the A band lies
above the B, C, etc., bands in energy.͒ The A band is intense,
with an absorption cross section of approximately 6.5
ϫ10Ϫ17 cm2 at its peak near 196 nm wavelength.2,3 The
breadth of the band ͓ϳ5000 cmϪ1 full width at half maxi-
mum ͑FWHM͔͒ and absence of structure indicate rapid and
direct dissociation following excitation. It has been sug-
gested that as many as three electronically excited states may
participate in the absorption,4–6 with the proportion from
each state varying with wavelength within the band. This
supposition is not unequivocally established, however, and
runs counter to the prediction of ab initio calculations on
ClNO.7,8
3
1
fundamental ͑1800 cmϪ1͒ was completely absent. Their in-
strumental resolution was insufficient to cleanly resolve the
␯2 (ϳ596 cmϪ1) and 2␯3 (662 cmϪ1) bands, and there were
similar merged features at higher Raman shifts. No detailed
analysis was possible at the time.
Our work examines Raman emission spectra of ClNO
using significantly higher resolution and shorter wavelengths
͑i.e., toward the center of the A band͒. At 212.5, 219, and
222 nm, the absorption cross sections are, respectively, about
2.1ϫ10Ϫ17, 1.0ϫ10Ϫ17, and 0.7ϫ10Ϫ17 cm2.3 There is no
striking evidence of electronic interference at these wave-
lengths, suggesting that a single electronic state is important
in the direct absorption in this wavelength range. This result
is placed in the context of recent assignments in a discussion
below.
The primary result of the current paper is the develop-
ment of a model capable of accurately reproducing all of the
vibrational levels that have been observed. Of these, a sig-
nificant fraction are reported for the first time. The large-
amplitude vibrational model is fully curvilinear, with nu-
merically converged three-dimensional ͑3D͒ eigenvalues all
giving transitions within a few cmϪ1 of the experimental
values. A unique aspect of this model construction turns out
to be the appropriateness of curvilinear Jacobi coordinates
for the vibrational motion in the vicinity of the ground elec-
tronic state equilibrium geometry. This model will serve as
the basis for a subsequent treatment of the dissociation dy-
Despite
a
large number of studies of the
photofragments,4–6,9–12 a complete picture of the photodisso-
ciation is still elusive. Raman spectroscopy, while it cannot
resolve all the mysteries, can in certain cases identify over-
lapped transitions through interference in fundamental tran-
sition intensities viewed as functions of excitation
wavelength.13,14 In addition, when the excitation is into a
dissociative continuum, Raman spectra produce long pro-
gressions of vibrational modes in the ground electronic state,
thereby yielding information about the ground state potential
energy surface ͑PES͒. At the same time, Raman spectros-
copy carried out with sufficient care is able to provide unique
insight into the early femtoseconds of direct dissociation in
the excited state.15–17
While such results have now been obtained for a variety
of directly dissociating molecules,18 the only ClNO Raman
¨
namics using solution of the time-dependent Schrodinger
equation in a wavelet basis.
a͒
Permanent address: Micron Technology, Inc., MS 306, 8000 Federal Way,
Boise, Idaho 83707-0006.
b͒
In Sec. II of this paper, the experimental apparatus is
described. Section III contains the experimental results and
Permanent address: The Dow Chemical Company, Bldg. 2504, P.O. Box
400, Plaquemine, Louisiana 70765.
0021-9606/2001/114(15)/6631/10/$18.00
6631
© 2001 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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

6632
J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Mackey et al.
compares the Raman spectra at three different excitation
wavelengths. The significance of the results with respect to
recent interpretations of the A band is discussed. Section IV
presents the vibrational analysis, including both a Dunham
expansion and a potential expansion in Jacobi coordinates
believed to be accurate in the Franck–Condon region and
somewhat beyond. Section V gives the conclusions.
direction in the system. The gas flow and the incident laser
light are both capable of being positioned to match this axis.
The light passing through the monochromator impinges
onto a solar blind photomultiplier tube ͑PMT͒, Hamamatsu
model number R166HA, operated at ϳ750 volts to ϳ1200
volts depending on signal strength. The output of the PMT
passes through a SRS-250 gated integrator, then through a
coax cable delay line to allow the prior arrival of a trigger
signal from a fast photodiode ͑FPD͒ at the data acquisition
equipment. The data acquisition system and a stepper motor
governing the angle of the monochromator diffraction grat-
ing are both controlled by a LabView program ͑National
Instruments͒.
The Raman emission from the sample is collected at an
angle of approximately 170° from the incident laser propa-
gation direction. In an optically thick sample, such a back-
scatter geometry is preferred since the emission comes pri-
marily from target molecules near the ‘‘surface’’ of the
sample, where there is a greatly reduced chance of reabsorp-
tion. The coaxial gas flow provides a rather sharp boundary
region ϳ1 mm thick between the sample and purge gas
flows.
Nitrosyl chloride is easily synthesized by placing an ap-
proximately stoichiometric mixture of chlorine and NO in a
reaction vessel in a dry ice–acetone bath.20 Liquid ClNO
condenses on the vessel walls in a few minutes. When the
pressure drops significantly below one atmosphere, more Cl2
and NO are added until the desired amount is obtained, typi-
cally about 50 mL of liquid ClNO for our experiments. The
total pressure is kept low to maintain safety. ClNO is purified
by slow warming while some of the vapor phase is pumped
off. The remaining ClNO is distilled into a new vessel, leav-
ing insignificant amounts of less volatile components. ClNO
has a boiling point ͑Ϫ5.5 °C͒ which is much above that of
either Cl2 (Ϫ35 °C) or NO ͑Ϫ101 °C͒. Both reactants are
therefore easily removed. The resulting ClNO has a narrow
melting range, indicating the absence of impurities, and no
significant unidentifiable interferences occur in the Raman
spectra. Temperatures are always kept sufficiently high to
avoid condensation of NO, as violent reactions can occur in
the liquid state.
Nitrosyl chloride is extremely corrosive, as are the reac-
tant gases used to produce it, especially in the presence of
water. Materials used in production and storage of ClNO
must be kept free from moisture and should be resistant to
attack by acid and chlorine. Our gas regulators are made
from special corrosion-resistant alloys. Other materials used
in various components of the apparatus ͑glass, viton 3/6,
stainless steel, Teflon͒ are all resistant to short-term chemical
attack.21 The primary manifold and reaction vessels are
glass. Teflon stopcocks with viton O-ring seals control the
gas flow in the vessels. The main gas lines supplying reac-
tants from their storage tanks are made of stainless steel,
with a few brass components. Brass O-ring adapters of our
design are used with Cajon Ultra-torr seals to connect the
delivery lines with the reaction vessel manifold. The ClNO-
saturated helium vapor flows through Teflon tubing from the
storage vessel to the sample cell and from the cell to the
exhaust hood. Prior to production, all gas lines and reaction
II. EXPERIMENT
Light generation for these experiments begins with a
Questek 2670 XeCl excimer laser operating at 308 nm, with
10 ns pulses of 100–300 mJ energy. Shot-to-shot fluctuations
in the total output energy for a given set of operating param-
eters are on the order of Ϯ20%. The rectangular (2.5
ϫ1.3 cm) excimer beam enters a Lambda Physik FL 3002
dye laser, which converts the light to visible wavelengths
with a roughly circular profile ͑ϳ1.5 mm͒ and energies on
the order of 2 mJ per pulse. Coumarin 440 and Stilbene 420
dyes were used in this work. The visible light from the dye
laser passes through
a beta-barium borate ͑␤-BBO͒
frequency-doubling crystal to produce ultraviolet ͑UV͒
wavelengths down to about 207 nm. Unconverted visible
light and UV light are separated by passing the mixed beam
through a self-compensating set of four-prisms, and blocking
the visible light. Typical pulse energies after conversion and
separation range from 5–60 ␮J/pulse. The UV light is
steered into a scattering cell where it intersects the sample
flow at 90°.
A flowing gas system is used to prevent sample deple-
tion by the UV pulses and to avoid the buildup of significant
concentrations of photoproducts that may provide spectral
interference. The gas flow system is designed to bring the
gas phase sample into the target region of the flow cell in a
highly reproducible and controllable manner. Carrier gas
͑He͒ flows into a wide-diameter ͑ϳ2.5 cm͒ fritted bubbler
immersed in liquid ClNO held at approximately Ϫ58 °C by a
refrigerated ethanol bath. This provides a stable concentra-
tion of ClNO mixed with the carrier gas and carried into the
cell through a vertical 1/4 stainless steel tube. A roughly
Љ
cylindrical shell of carrier gas surrounds the flowing sample
to keep it collimated and localized at the center of the cell.
The flow rate of the coaxial shell is roughly twice that of the
sample gas, optimal conditions being determined beforehand
from examination of the white smoke produced by passage
of pre-mixed NH3 /HCl with helium as a carrier gas. Win-
dow purges of pure He protect the front and rear cell win-
dows and an internal light collection lens.
All gases exit through an aperture at the bottom of the
cell for release into a fume hood. During data acquisition, the
total pressure in the cell remains at approximately one atmo-
sphere.
The data acquisition system includes the light collection
optics, a grating monochromator, a photomultiplier detector,
and computer hardware/software for storing the data and run-
ning the experiment. The Czerny–Turner monochromator
͑Spex model 1402͒ has a focal length of 0.75 m. A Spec-
tronic Instruments 100 mm square holographic diffraction
grating with 2400 sinusoidal grooves per mm was used. The
optical axis of the monochromator provides a fixed spatial
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Raman investigation of ClNO
6633
FIG. 2. Resonance Raman spectra of CINO.
FIG. 1. The absorbance of the ClNO A band, measured by Ballash et al.
͑Ref. 2͒. The three vertical lines at 212.5, 219, and 222 nm indicate the
wavelengths used in this study.
tral feature and the band origin must be determined by per-
forming a rotational band contour analysis, for which the
individual rotational lines are convolved with the spectral
transfer function of the monochromator.37 Inertial constants
for selected vibrational states of ClNO are available from a
variety of sources.34,38–44 The asymmetry parameter for
vessels are thoroughly cleaned and oven dried. All materials
are then pumped out by a diffusion pump while being heated
well above 100 °C to drive away adsorbed water.
III. ClNO RAMAN SPECTRA
A. Results
TABLE I. Raman shifts from this study compared to bands from previous
IR absorption and IR Raman data ͑Refs. 23–36͒.
The absorption spectrum of ClNO from 15 000 cmϪ1 to
53 000 cmϪ1 was measured by Goodeve and Katz.1 The A
band was later remeasured by Ballash et al., and by Tyndall
et al.2,3 Figure 1 shows the spectrum of Ballash et al., with
vertical lines indicating the three wavelengths of our mea-
surements ͑212.5, 219, and 222 nm͒. These excitation wave-
lengths were chosen for ͑1͒ experimental accessibility, ͑2͒
resonant enhancement in the A band, and ͑3͒ avoidance of
absorption bands of free NO ͑see Okabe22͒ that might be
present in the cell. Raman shifts were measured out to
Vibrational frequencies ͑cmϪ1
͒
Level
This study
Previous
␯
␯
2␯
␯2ϩ␯
3␯
329.5
597.6
332.0
595.9
3
2
661.9
926.3
983.5
3
3
3
2␯
1184.2
1249.1
1506.0
1569.0
1640.4
1767.6
1186.5
2
␯2ϩ2␯
2␯2ϩ␯
␯2ϩ3␯
3
3
3
ϳ3000 cmϪ1
.
5␯
3␯
␯
3
The Raman spectra for the three excitation wavelengths
are given in Fig. 2. The spectra are vertically scaled to give
the same difference in height between the maximum at 332
cmϪ1 and the minimum at ϳ500 cmϪ1. The three spectra are
qualitatively similar, especially for the first 1000 cmϪ1 of
Raman shift. Each spectrum was collected several times to
ensure reproducibility. The centers of the Raman features for
the 212 nm excitation are given in Table I with the corre-
sponding IR absorption results and previous nonresonant Ra-
man data.23–26 The spectrum with excitation at 212 nm was
chosen for comparison since it is the least congested. This
study adds 12 previously unobserved bands to the 16 known
bands for ClNO ͑not including a difference band reported by
Jones et al.,30 in 1968͒.
2
1799.7
1831.2
1894.5
1
2␯2ϩ2␯
1833.9
1892.8
2089.1
3
3
␯2ϩ4␯
3
3␯2ϩ␯
3
␯1ϩ␯
2131.1
2395.6
3
2␯2ϩ3␯
2150.9
2396.7
2458.7
2723.5
␯1ϩ␯
2
␯1ϩ2␯
3
␯1ϩ␯2ϩ␯3/3␯2ϩ3␯
3
2␯
3563.8
3894.6
4159.7
4752.0
5292.9
5623.0
5888.6
6985.2
1
2␯1ϩ␯
3
2␯1ϩ␯
2
2␯1ϩ2␯
2
3␯
1
3␯1ϩ␯
3
3␯1ϩ␯
2
Since our spectrometer could not resolve individual ro-
tational lines, the relationship between the center of a spec-
4␯
1
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

6634
J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Mackey et al.
ClNO in its ground vibrational level is ␬ϭϪ0.9912, i.e.,
very close to that of a prolate symmetric top. We calculated
band contours with symmetric rotor energy levels and rela-
tive oscillator strengths, assuming a-axis type rotational tran-
sitions. The ab initio results of Bai et al. place the transition
what in question, even near band center. The existence of
two channels with different values of ␤ and different distri-
butions of ET seems to be unambiguous. Haas et al. say this
can arise either from initial excitation into two different elec-
tronic states or from initial excitation to a single excited
electronic state followed by nonadiabatic leakage of part of
the amplitude onto a second PES. They favor the first of
these possibilities and speculate that one channel may pro-
duce ground-state Cl(2P3/2) and the other excited Cl(2P1/2).
To a first approximation, the angular distribution of
products is a snapshot of the initial distribution of orienta-
tions of the transition moment in spaced-fixed coordinates
͑averaged over the ensemble of rotational states͒. Hence, the
idea that excitation at band center involves two excited elec-
tronic states in proportions of about 1:5 seems fairly compel-
ling. It is, however, at odds with the ab initio calculations of
Bai et al. on the ClNO molecule.7,8 While Bai’s results indi-
cate three excited electronic states with roughly the correct
energy, the relative intensities are calculated to be in the ratio
1:10Ϫ3:0 ͑the last value because the excited state is a trip-
let͒. Despite any shortcomings of the basis sets used and the
absence of spin–orbit terms in the ab initio calculation, it is
difficult to see how the results could be off by such a large
factor in the relative intensities. An alternative explanation,
whose plausibility is difficult to assess, is initial excitation to
a single PES with subsequent evolution into two separate
components with different average rotational periods during
photodissociation, either remaining on the same PES or with
one of the components representing nonadiabatic transfer to
another electronic state.
moment near the inertial a axis8 for the S (4 1A )
Ј
5
S (1 1A ) transition. The calculated band contours are in-
Ј
0
sensitive to the inertial asymmetry and to small b-axis com-
ponents of the transition moment. The Raman spectrum of
methane was used to obtain an experimental instrument
function for purposes of the convolution. The line shape ob-
tained for the methane fundamental ͑all Q branch͒ is com-
pletely dominated by instrumental broadening and serves
quite well as the spectral transfer function for use with
ClNO. The ClNO rotational contours thus produced are in
good agreement with the spectra actually measured. For all
the vibrational levels we simulated, the band centers corre-
sponded to band origins to well within the limits of experi-
mental error. The same is expected to hold for all of the
vibrational states measured in these experiments.
B. Relation to previous studies
It has been known since flash photolysis experiments in
the early 1960’s that photodissociation of ClNO at short
wavelengths produces vibrationally excited NO.45 The de-
tails of energy partitioning between product rotation, vibra-
tion, translation, and electronic energy appear to vary a great
deal with wavelength across the band. A good overview of
what has been done with respect to photofragment studies
can be found in the paper of Skorokhodov et al.6 and refer-
ences given therein.
Felder and Morley5 used the same apparatus as Hass,
Felder, and Huber for PTS at 248 nm. At this wavelength,
the absorption cross section is only about 3.8ϫ10Ϫ19 cm2,
which raises a question whether it should be considered part
of the same band at all. They found a bimodal ET distribu-
tion ͑actually a single peak with a long tail to low ET), with
the minor component having a relative importance of 20%.
The asymmetry parameter ␤ for the major component was
ϳ1.2. The ␤ value for the minor component was less well
determined, but appeared to be about the same. Felder and
Morley proposed a total of three excited electronic states to
account for the entire A band. They called these Aa , Ab , and
Ac in increasing order of energy, and attributed Ac to the Bai
et al. strong S (4 1A ) S (1 1A ) transition. A and Aa
At 193 nm, near the center of the band, the most defini-
tive and unambiguous results are the photofragment transla-
tional spectroscopy ͑PTS͒ results of Haas, Felder, and
Huber.14 They found a bimodal distribution in the product
translational energy ET and were able to fit their data to two
separate components with differing values of the asymmetry
parameter ␤ for the angular distribution of photoproducts. A
component with ␤Ϸ0.95 and
E
Ϸ37.4 kcal molϪ1 ac-
͘
T
͗
counted for about 17% of the products. A second component
with ␤Ϸ0.45 and
E
Ϸ43.1 kcal molϪ1 made up the re-
͗
͘
T
maining ϳ83%. The overall E was about 42.1 kcal molϪ1
.
͗
͘
T
The measured ET distribution was decomposed into indi-
vidual vibrational components by assuming that the mol-
ecules had a distribution of rotational energies with an aver-
age rotational energy ERϳ5 kcal molϪ1 and FWHM of ϳ4.6
kcal molϪ1, irrespective of vibrational state. The result is an
estimate of about 63.6 kcal molϪ1 for E . This is a signifi-
Ј
Ј
5
0
b
were assigned as S (3 1A ) S (1 1A ) and T (3 3A )
Ј
Ј
Ј
4
0
5
S (1 1A ), respectively.
Ј
0
These authors proposed that the absorption at 193 nm
owes 83% to Ac and ϳ17% to Ab . At 248 nm, they assign
80% to Aa and 20% to Ab . Noting that the average internal
͑rovibrational͒ energy of the NO is a much larger fraction of
the total energy at 193 nm than at 248 nm, they predicted
that5 ‘‘the Raman spectrum at 248 nm, in contrast to the one
obtained at 266 nm, should begin to show a weak progres-
͗
͘
V
cantly higher value than would be derived from the results of
Gillan et al.,11 who used laser induced fluorescence on rota-
tionally and translationally relaxed NO products to determine
E
. Their vibrational distribution is, in turn, skewed to-
͘
͗
V
wards higher
than the distribution given by Moser et al.,10
v
Љ
who measured infrared emission from rotationally relaxed
sion in the ␯ mode, which should become even more impor-
1
products. Gillan et al. persuasively show, however, that at
tant at shorter wavelengths.’’ This prediction is not borne
least the lowest
states observed by Moser et al., were
out by the present investigation. The ␯ mode is seen only in
v
Љ
1
produced by secondary processes and were not present in the
nascent distribution. Thus, the details of the energy distribu-
tion amongst the product’s degrees of freedom is still some-
combination bands with the other modes.
Skorokhodov et al.6 performed two kinds of measure-
ments that allowed specific detection of Cl(2P3/2) and
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Raman investigation of ClNO
6635
Cl ( P1/2): ͑1͒ Photofragment imaging46 using REMPI de-
2
TABLE II. Dunham coefficients for ClNO: Jones et al. ͑Ref. 30͒ this study,
and standard deviations for this study. All values are in cmϪ1
*
.
*
tection, and ͑2͒ VUV–LIF determination of the Cl /Cl ratio.
*
The results indicated a Cl /Cl ratio of ϳ0.9 for dissociation
Coefficient
Jones et al.
This work
Std. dev.
at 212.56 nm, of ϳ0.45 at 235 nm and of ϳ1.1 at 248 nm.
␻
␻
␻
x11
x12
x13
x22
x23
x33
1835.6
603.2
336.4
Ϫ17.8
0.0
Ϫ0.6
Ϫ2.6
Ϫ4.3
Ϫ1.0
1831.6
600.4
331.2
Ϫ17.4
1.2
3.8
3.9
2.9
0.5
0.9
1.0
0.7
0.7
0.3
1
*
For photodissociation at ϳ235 nm, both Cl and Cl have
2
*
bimodal distributions of ET . The low ET component for Cl
3
had ␤Ϸ1.8 and the high ET component had ␤Ϸ1.5. The
angular distribution for Cl was similar, leading to the inter-
pretation that all four components arise from a single elec-
tronic transition, i.e., the state called Ab by Felder and Mor-
ley. The signal for dissociation at ϳ211 nm was too poor to
obtain a definite value of ␤, though it was reported to be low.
This similarity to the situation at 193 nm led Skorokhodov
et al., to assign all the action at both 193 nm and 211 nm to
the single dominant surface Ac . They attribute the dissocia-
tion at 235 nm to Ab , and that at 248 to Aa and Ab .
The lack of any significant electronic interference in the
Raman fundamentals measured here suggests that only one
state is dominant ͑although possibly multiple states charac-
terized by nearly parallel PES’s͒ in the region 212–222 nm.
This abuts the wavelength region assigned by Skorokhodov
et al., and so would be Ac under their assignment. This then
requires that the crossover to dominance of the Ab transitions
must occur in the narrow wavelength range 222–235 nm.
Further analysis of the electronic state implications must
await quantitative modeling that is beyond the scope of the
present paper.
0.7
Ϫ2.7
Ϫ3.1
Ϫ0.5
The results of our fit are compared with the observed
peak positions in Table III. The largest magnitude of devia-
tion is 4.1 cmϪ1. For features that were present both in our
data and the previous data, the more precise IR data were
used in the fitting procedure.
The ability of this simple Dunham expansion to fit the
levels might be taken as an indication that CINO could ad-
equately be described in terms of normal coordinates. We
know, however, that the highest energy levels observed must
correspond to large amplitude excursions from the equilib-
rium geometry—a situation that is seldom if ever well de-
scribed by normal coordinates. What is more, even in the
lower-energy ground state vibrations, normal coordinate
analyses have produced controversy about the assignment of
the modes. McDonald et al., have recently and definitively
assigned ␯ ͑330 cmϪ1͒ as a Cl–N stretching mode and ␯
3
2
͑600 cmϪ1͒ as a ‘‘scissoring’’ motion pivoting on nitrogen,
IV. VIBRATIONAL ANALYSIS
A. Background
TABLE III. Spectral features, the predicted locations from our Dunham
Bailey and Cassie measured the first IR absorption spec-
trum of ClNO in 1934.47 Since that time, upwards of 20
individual studies have dealt with various aspects of the vi-
brational and rotational absorption spectrum. A comprehen-
sive overview of work up to 1986 is given by McDonald
et al.,33 although a number of further studies have occurred
since this paper was published.
Jones et al., published the first Dunham expansion for
the vibrational terms of ClNO in 1968, giving harmonic fre-
quencies and first order anharmonicity constants.30 Their fit
used 17 transitions for 16O14N35Cl ͑all that were known at the
expansion, and differences. All values are in cmϪ1
.
Level
Observed
Calculated
Difference
␯
␯
332.0
595.9
332.6
598.3
0.6
2.4
3
2
2␯
␯2ϩ␯
3␯
661.9
926.3
983.5
659.2
922.6
985.9
Ϫ2.7
Ϫ3.7
2.4
3
3
3
2␯
1186.5
1249.1
1506.0
1569.0
1640.4
1767.6
1799.7
1833.9
1892.8
2089.1
2131.1
2150.9
2395.6
2458.7
3563.8
3894.6
4159.7
4752.0
5292.9
5623.0
5888.6
6985.2
1186.0
1246.9
1507.9
1571.1
1639.1
1769.0
1801.8
1829.8
1895.4
2088.6
2130.3
2151.8
2396.0
2458.7
3563.7
3894.0
4159.8
4751.2
5291.5
5623.6
5889.5
6985.4
Ϫ0.5
Ϫ2.2
1.9
2.1
Ϫ1.3
1.4
2.1
Ϫ4.1
2.6
Ϫ0.5
Ϫ0.8
0.9
2
␯2ϩ2␯
2␯2ϩ␯
␯2ϩ3␯
5␯
3␯
3
3
3
time͒, including the ␯1 –␯ difference band. We have com-
2
3
bined our Raman data with earlier data to produce a revised
Dunham expansion:
2
␯
1
2␯2ϩ2␯
␯2ϩ4␯
3␯2ϩ␯
␯1ϩ␯
2␯2ϩ3␯
␯1ϩ␯
␯1ϩ2␯
2␯
2␯1ϩ␯
2␯1ϩ␯
2␯1ϩ2␯
3␯
3␯1ϩ␯
3␯1ϩ␯
4␯
3
3
3
3
3
3
G
,
,
͒ϭ
␻
ϩ1/2͒ϩ
x
kϩ1/2͒
͑v
͑v1 v2 v3
͑vk
3
͚
͚ ͚
k
kj
kϭ1
k
jуk
3
ϫ
͑vj
ϩ1/2͒.
͑1͒
0.5
0.0
2
3
Table II gives the Dunham coefficients found by Jones et al.,
our Dunham coefficients obtained from the larger set of lev-
els, and the standard deviations for the parameters in our fit.
There are nine independently adjustable parameters to fit 28
spectral features. All ␻ values are well determined, as are
Ϫ0.1
Ϫ0.6
0.1
Ϫ0.8
Ϫ1.4
0.6
1
3
2
2
1
3
x
11 , x22 , and x23 . The values of x12 , x13 , and x33 are poorly
0.9
0.2
2
determined. The value of x13 is not significantly different
1
from zero.
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

6636
J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Mackey et al.
since the angular momentum of the parent molecule must be
conserved.͒ Jacobi coordinates have been used to describe
the dissociation in other bands of CINO which show
little rotational excitation beyond the initial thermal
distribution.49–51
The lengths r1 and r3 and the angle ␪ between rជ1 and rជ3
provide three ͑curvilinear but orthogonal͒ vibrational coordi-
nates for describing the vibrational modes. ͑The subscripts
on the Jacobi coordinates are chosen so that r1 correlates to
the ␯ vibrational mode and r3 to ␯3 .) In general, all three
1
Jacobi coordinates will be coupled in the molecular region of
the potential surface, though there are some indications that
r1 is only weakly coupled to the other two coordinates: the
NO bond is found to have similar lengths in the diatomic
͑1.1508 Å, Herzberg52͒ and triatomic ͑1.1405 Å, Jones30͒
FIG. 3. The ClNO Jacobi coordinates.
molecules; the ␯ mode is found to be a very minor player in
1
but they point out that this leads to an unusually low stretch-
ing frequency and an unusually high bending frequency for a
molecule with constituent atoms of these masses.33 It is rare
but not unknown for a bending mode to have a higher fre-
quency than a stretching mode. Our Raman spectra with A
band excitation reflect dynamics in the excited state͑s͒ fol-
lowing absorption of a photon. These spectra show more
the resonance Raman spectra, indicating that there is not a
great deal of change of the NO bond length during the early
femtoseconds of the excited state dynamics.
For zϭcos(␪) and for volume element dr1 dr2 dz, the
vibrational kinetic energy in Jacobi coordinates is
ប2 ץ2
ប2 ץ2
TϭϪ
Ϫ
ץr21
ץr23
2␮
2␮
1
3
overtone and combination activity in ␯ than in ␯2 , thus
3
supporting the assignment of this mode as largely a Cl–N
stretching motion.
ប2
1
1
ץ
ץ
Ϫ
ϩ
1Ϫz2͒
,
͑2͒
͑
ͩ
ͪ
␮1r21 ␮3r23
2
ץz
ץz
In their exploration of the stretch vs bend issue,
McDonald et al., performed a number of calculations.33
where ␮ is the reduced mass of NO and ␮ is the reduced
1
3
First, they attempted a normal mode fit assigning the ␯
3
mass of Cl and NO. Kinetic coupling between the coordi-
nates occurs only from the rϪ1 2 and rϪ3 2 terms in the bending
mode as the bend and the ␯ mode as the N–Cl stretch. The
2
Ϫ2
1e
Ϫ2
3e
result failed to reproduce the observed rotational distortion
kinetic energy. Replacing these by r
tively, allows an approximately separable model for describ-
ing the vibrational levels.
and r , respec-
constants. Another attempt, using the bend for ␯ and the
2
N–Cl stretch for ␯ produced satisfactory results with four
3
adjustable parameters, including an interaction term coupling
the N–Cl stretch and the bend. Finally, a third force field was
generated, using the three bond lengths ͑N–Cl, N–O, and
Cl–O͒ as ͑nonorthogonal͒ coordinates. A potential energy
For a crude first approximation employing this separable
kinetic energy expression, we used the potential
2
V ϭc
Ј
1Ϫexp Ϫa r Ϫr ͒͒ 2ϩc020 r Ϫr
͒
͑ ͑ ͔ ͑
1 1 1e 3 3e
͓
200
ϩc002 zϪz ͒2,
͑3͒
distribution analysis for this model indicated that the ␯ nor-
2
͑
e
mal mode was a strong mixture of the N–Cl and N–O
stretches. In their words ‘‘It is clear that the choice of a
model and the interpretation of the physical significance of a
model must be done carefully especially for a molecule like
ONCI.’’
where r1eϭ1.1405 Å, r3eϭ2.2828 Å, and zeϭϪ0.607 49.
The remaining parameters were adjusted to optimize a fit to
the known bands. Since the anharmonicity of ␯ is well-
1
determined from the Dunham expansion, the r1 mode was
modelled as a Morse oscillator. This procedure yielded a
physically reasonable starting point in the iterative optimiza-
tion of parameters for the more general potential of the form
B. Jacobi coordinates
We have chosen to describe the system in Jacobi coor-
dinates. This choice is natural for describing the atom–
diatomic molecule products as the system proceeds towards
dissociation but also turns out to provide an adequate de-
scription in the neighborhood of the ground-state equilibrium
geometry. Jacobi coordinates explicitly include the ability to
describe large amplitude motions.48 In the specific case of
ClNO, the two Jacobi vectors are rជ1 , which points from N to
O, and rជ3 which points from the NO center of mass to Cl
͑see Fig. 3͒. In a purely impulsive model of the dissociation
taken strictly along the rជ3 direction, there is no torque on the
NO fragment upon dissociation since the impulse is‘‘a-
gainst’’ the NO center of mass. ͑This does not mean, how-
ever, that the NO fragment will have no rotational motion,
i
Vϭ
c
1Ϫexp Ϫa r Ϫr ͒͒
͓ ͑ ͑ ͔
ijk 1 1 1e
͚ ͚ ͚
i
j
k
j
k
ϫ r Ϫr ͒ zϪz ͒
͑4͒
͑
͑
3
3e
e
suitable for modelling all observed bands including combi-
nation bands. The kinetic coupling in Eq. ͑2͒ was also fully
included at the same time. Vibrational eigenvalues for this
potential were then determined using a large-amplitude vi-
brational basis.
We followed the strategy of Johnson and Reinhardt53,54
in using Laguerre bases of exponential arguments for radial
coordinates and Jacobi polynomial bases for the angular co-
ordinate. Basis set parameters and number of configurations
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Raman investigation of ClNO
6637
TABLE IV. Nonzero potential parameters for the fit in Jacobi coordinates.
a1
2.7924 ÅϪ1
c200
c030
c004
c022
47 874 cmϪ1
c020
c002
c202
26 325 cmϪ1 ÅϪ2
74 555 cmϪ1
4750 cmϪ1
Ϫ336.3 cmϪ1 ÅϪ3
0.7083 cmϪ1
170.8 cmϪ1 ÅϪ2
͑products of primitive basis functions in the three individual
vibrational coordinates͒ were optimized in bootstrap fashion
to ensure convergence to within ϳ0.1 cmϪ1 of the eigenval-
ues corresponding to all observed vibrational transitions. A
total of 898 basis functions were ultimately used. For a fixed
set of parameters describing the basis functions and a fixed
basis size, the potential parameters were varied. The optimi-
zation was occasionally interrupted to verify that all of the
relevant eigenvalues were well converged. The potential op-
timization, as judged by mean square error of the calculated
transitions, first emphasized the fundamentals. Then over-
tones were added, and finally the combination bands. Several
higher order and cross terms in the potential were tested
individually and in combination to see whether they signifi-
cantly improved the fit.
The best fit balancing the competing goals of minimum
error and minimum number of parameters was achieved with
a potential that had eight adjustable parameters. The best fit
values of these parameters are given in Table IV. Table V
compares the fit to experimental data. The rms deviation for
the whole set of vibrational terms is 2.71 cmϪ1. The largest
deviation from a spectral feature is an underestimate by
FIG. 4. The ground state wave function contours. ͑a͒ r1 vs ␪ with r3 fixed
at 2.2 Å. ͑b͒ r3 vs ␪ with r1 at fixed at 1.14 Å. ␪ is in radians, r1 and r3 are
in Å.
about 5 cmϪ1 for the 3␯1ϩ␯ combination band. Contour
2
plots of the ground state and selected excited state wave
functions are shown in Figs. 4–6.
TABLE V. Observed spectral features, calculated values from fit, and dif-
A small-amplitude expansion of our potential allows a
normal modes calculation that can be compared directly to
that of other investigators. Figure 7 compares our stick dia-
grams obtained in this way with those arrived at by Mc-
Donald et al.31 The two sets of normal coordinates are quite
ferences, in units of cmϪ1
.
Level
Observed
Calculated
Difference
␯
␯
2␯
␯2ϩ␯
3␯
332.0
595.9
661.9
926.3
983.5
329.5
595.6
658.0
922.4
985.8
Ϫ2.5
Ϫ0.3
Ϫ3.8
Ϫ3.9
2.3
3
2
similar. The two ␯ normal coordinates are virtually identi-
1
3
cal, as they should be since both coordinate systems take q1
as the pure N–O stretch. Perhaps the most prominent differ-
3
3
2␯
1186.5
1249.1
1506.0
1569.0
1640.4
1767.6
1799.7
1833.9
1892.8
2089.1
2131.1
2150.9
2395.6
2458.7
3563.8
3894.6
4159.7
4752.0
5292.9
5623.0
5888.6
6985.2
1185.1
1248.4
1509.4
1573.5
1638.5
1768.5
1800.2
1832.7
1897.6
2090.2
2129.7
2155.1
2394.5
2458.3
3567.9
3895.3
4159.4
4749.0
5295.4
5624.0
5883.6
6989.0
Ϫ1.4
Ϫ0.7
3.4
4.5
Ϫ1.9
0.9
0.5
Ϫ1.2
4.8
1.1
Ϫ1.4
4.2
2
ence in the diagrams is the ␯ mode. Our assignment for the
3
␯2ϩ2␯
2␯2ϩ␯
␯2ϩ3␯
5␯
3␯
3
3
3
␯ fit is obviously the Jacobi stretch, involving no change in
3
the N–O distance. The results of Jones et al. for ␯ is a mode
3
that appears to have some slight bending character, with ni-
trogen relatively stationary while chlorine and oxygen move
generally away ͑and toward͒ each other.
3
2
␯
1
2␯2ϩ2␯
␯2ϩ4␯
3␯2ϩ␯
␯1ϩ␯
2␯2ϩ3␯
␯1ϩ␯
␯1ϩ2␯
2␯
2␯1ϩ␯
2␯1ϩ␯
2␯1ϩ2␯
3␯
3␯1ϩ␯
3␯1ϩ␯
4␯
3
3
Because of the relatively weak coupling between the Ja-
cobi coordinates in the numerically converged large-
amplitude vibrational calculations, the Jacobi coordinates are
identified as ‘‘good’’ coordinates for the ground electronic
state of nitrosyl chloride. The potential in these coordinates
is physically reasonable, has relatively small coupling con-
stants, and sufficiently describes all measured vibrational
spectra to good accuracy. Moreover, the modal structure of
the vibrational wave functions ͑Figs. 4–6͒ clearly aligns
along the appropriate coordinate axes with no significant dis-
tortion. This is salutary because Jacobi coordinates are the
natural choice for describing the photodissociation process in
the A band.
3
3
3
Ϫ1.1
Ϫ0.4
4.1
2
3
1
0.8
3
Ϫ0.3
Ϫ3.0
2.5
1.0
Ϫ5.0
3.8
2
2
1
3
2
1
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

6638
J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Mackey et al.
FIG. 6. The 17th excited state, corresponding to one quantum of excitation
in the ␯ mode and three quanta in the ␯ mode. Shown is r3 vs ␪. ␪ is in
radians, r3 in Å. ͑a͒ r1ϭ1.14 Å, ͑b͒ r1ϭ1.2 Å, ͑c͒ r1ϭ1.25 Å. This cor-
responds to the 15th entry in Table V.
3
2
FIG. 5. The first excited state, corresponding to excitation of the ␯ mode,
showing ␪ vs r3 . ␪ is in radians, r3 in Å. ͑a͒ r1ϭ1.14 Å, ͑b͒ r1ϭ1.2 Å, ͑c͒
r1ϭ1.23 Å.
3
V. DISCUSSION
of the strong evidence of vibrational excitation in the NO
Resonance Raman spectra have been obtained for CINO
at three wavelengths in the ultraviolet A band. Twelve pre-
viously unobserved vibrational levels in the ground state
have been observed. These new levels, together with previ-
ously known vibrational levels, allowed the construction of a
large amplitude vibrational model valid in and beyond the
Franck–Condon region. The ground state surface shows a
large degree of separability in Jacobi coordinates even in the
triatomic molecular region; this finding bears directly upon
previous difficulties in identifying the ␯ and ␯ modes. This
photoproduct. The ␯ mode does show up in overtones and
1
combination bands. This is reminiscent of the behavior of the
␯
umbrella bending mode in CH3I,55 i.e., absence of the
2
fundamental, but appearance of the mode in combination
with the C–I stretch. This was taken to indicate that motion
in the umbrella angle did not occur immediately upon exci-
tation, but started to develop only when the C–I bond angle
had lengthened from its ground-state equilibrium value. The
same reasoning would apply for CINO. The behavior of the
␯ mode seems to indicate that motion in the r1 coordinate is
1
2
3
ground-state PES and its vibrational eigenvalues will become
the foundation of an effort to model the photodissociation
dynamics in the upper electronic state͑s͒.
Although the work reported here concentrates primarily
on the vibrational levels of the electronic ground state, there
are a couple of clues to the dynamic behavior of electroni-
cally excited CINO in the pattern of levels observed and
their relative intensities. The absence of any observable ac-
not occurring in the first few fs of evolution, but develops
later in the course of the photodissociation.
The Raman spectra may also have some bearing on the
question of whether more than one electronically excited
state is operative near the center of the A band. Simultaneous
excitation of more than one excited electronic state sets up
the possibility of interference between amplitudes evolving
in the two different states. This has clearly been seen in the
case of CH3I ͑Ref. 14͒ and higher alkyl iodides.13 No such
tivity in the ␯ fundamental is somewhat surprising in view
1
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Raman investigation of ClNO
6639
and Fourier transformation to obtain Raman scattering
amplitudes15–17,57 provides for comparison with the observed
patterns of intensity.
ACKNOWLEDGMENTS
Grateful acknowledgment is made to the National Sci-
ence Foundation and the Robert A. Welch Foundation for
support of this work.
1 C. F. Goodeve and S. Katz, Proc. R. Soc. London, Ser. A 172, 432 ͑1939͒.
2 N. M. Ballash and D. A. Armstrong, Spectrochim. Acta, Part A 30, 941
͑1974͒.
3 G. S. Tyndall, K. M. Stedman, W. Schneider, J. P. Burrows, and G. K.
Moortgat, J. Photochem. 36, 133 ͑1987͒.
4 B.-M. Haas, P. Felder, and J. R. Huber, Chem. Phys. Lett. 180, 293
͑1991͒.
5 P. Felder and G. P. Morley, Chem. Phys. 185, 145 ͑1994͒.
6 V. Skorokhodov, Y. Sato, K. Suto, Y. Matsumi, and M. Kawasaki, J.
Phys. Chem. 100, 12321 ͑1996͒.
7 D. Solgadi, F. Lahmani, C. Lardeux, and J. P. Flament, Chem. Phys. 79,
225 ͑1983͒.
8 Y. Y. Bai, A. Ogai, C. X. W. Qian, L. Iwata, G. A. Segal, and H. Reisler,
J. Chem. Phys. 90, 3903 ͑1989͒.
9 L. Werner, B. Wunderer, and H. Walther, Chem. Phys. 60, 109 ͑1981͒.
10 M. D. Moser, E. Weitz, and G. C. Schatz, J. Chem. Phys. 78, 757 ͑1983͒.
11 I. T. F. Gillan, D. J. Denvir, H. F. J. Cormican, I. Duncan, and T. Morrow,
Chem. Phys. 167, 193 ͑1992͒.
12 H. L. Kim, Y. Mo, Y. Matsumi, and M. Kawasaki, Bull. Korean Chem.
Soc. 13, 162 ͑1992͒.
13 D. L. Phillips and A. B. Myers, J. Chem. Phys. 95, 226 ͑1991͒.
14 G. E. Galica, B. R. Johnson, M. O. Hale, and J. L. Kinsey, J. Phys. Chem.
95, 7994 ͑1991͒.
FIG. 7. Comparing the normal modes of McDonald et al., with the Jacobi
coordinates found in this study. ͑a͒ and ͑b͒ are the N–O stretches in each
coordinate system and are virtually identical. ͑c͒ is McDonald’s Cl–N–O
bend. ͑d͒ is the Jacobi bend. ͑e͒ is McDonald’s N–Cl stretch, while ͑f͒ is the
Jacobi Cl–͑NO͒ stretch. In ͑a͒ through ͑d͒ the chlorine atom appears to be
stationary because to make its motion visible would make the other motion
arrows extend beyond the figure’s border.
15 E. J. Heller, Acc. Chem. Res. 14, 368 ͑1981͒.
16 E. J. Heller, in Potential Energy Surfaces and Dynamics Calculations,
edited by D. G. Truhlar ͑Plenum, New York, 1981͒.
17 E. J. Heller, R. L. Sundberg, and D. Tannor, J. Phys. Chem. 86, 1822
͑1982͒.
18 B. R. Johnson, C. Kittrell, P. B. Kelly, and J. L. Kinsey, J. Phys. Chem.
100, 7743 ͑1996͒.
19 A. J. Bell and J. G. Frey, Mol. Phys. 69, 943 ͑1990͒.
20 D. A. Baugh ͑private communication͒.
interference effects are evident in the CINO results. While
absence of interferences does not conclusively rule out par-
ticipation of more than one excited electronic state in the
absorption, it weighs in on that side of the issue, at least
insofaras the absorption and early dynamical evolution is
concerned. On the basis of this evidence, therefore, we prefer
a picture in which the initial excitation for these wavelengths
is predominantly to a single excited electronic state, Bai’s
21 Houston Valve Company, Corrosion Charts, 1996.
22 H. Okabe, Photochemistry of Small Molecules ͑Wiley, New York, 1978͒.
23 C. M. Beeson and D. M. Yost, J. Chem. Phys. 7, 44 ͑1939͒.
24 P. Botschwina, H. Haertner, and W. Sawodny, Chem. Phys. Lett. 74, 156
͑1980͒.
25 W. G. Burns and H. J. Bernstein, J. Chem. Phys. 18, 1669 ͑1950͒.
26 J. R. Durig and R. C. Lord, Spectrochim. Acta 19, 421 ͑1963͒.
27 W. H. Eberhardt and T. G. Burke, J. Chem. Phys. 20, 529 ͑1952͒.
28 M. Feuerhahn, W. Hilbig, R. Minkwitz, and U. Engelhardt, Spectrochim.
Acta, Part A 34, 1065 ͑1978͒.
S (4 1A ). In order to account for the observed bimodal dis-
Ј
5
29 I. C. Hisatsune and P. Miller, Jr., J. Chem. Phys. 38, 49 ͑1963͒.
30 L. H. Jones, R. R. Ryan, and L. B. Asprey, J. Chem. Phys. 49, 581 ͑1968͒.
31 L. H. Jones and B. I. Swanson, J. Phys. Chem. 95, 86 ͑1991͒.
32 L. Landau and W. H. Fletcher, J. Mol. Spectrosc. 4, 276 ͑1960͒.
33 J. K. McDonald, J. A. Merritt, V. F. Kalasinsky, H. L. Heusel, and J. R.
Durig, J. Mol. Spectrosc. 117, 69 ͑1986͒.
tribution in ET and the different ␤ parameters for the two
components, one would then have to postulate either bifur-
cation of the evolving wave packet into two pieces on that
surface or nonadiabatic transfer of part of the amplitude to
another electronic PES. In either case, the difference in ␤
parameters could only be understood if the two channels
involved significant differences in the amount of molecular
rotation ͑of the triatomic entity͒ during the course of the
dissociation.
We are in the process of constructing a code for time-
dependent propagation in curvilinear coordinates using
wavelets.56 This will be used to model the observed Raman
spectra, determining the minimum number of PES’s re-
34 A. M. Mirri and E. Mazzariol, Spectrochim. Acta 22, 785 ͑1966͒.
35 A. G. Pulford and A. Walsh, Trans. Faraday Soc. 47, 347 ͑1951͒.
36 J. H. Wise and J. T. Elmer, J. Chem. Phys. 18, 1411 ͑1950͒.
37 J. E. Stewart, Infrared Spectroscopy ͑Marcel Decker, New York, 1970͒.
38 G. L. Bendazzoli, G. Cazzoli, C. D. Esposti, G. Fano, F. Ortolani, and P.
Palmieri, J. Chem. Phys. 84, 5351 ͑1986͒.
39 G. Cazzoli, C. Degli Esposti, D. Dmiani, and A. M. Mirri, Chem. Phys.
Lett. 80, 365 ͑1981͒.
40 G. Cazzoli, C. Degli Esposti, P. Palmierii, and S. Simeone, J. Mol. Spec-
trosc. 97, 165 ͑1983͒.
41 B. Gatehouse, H. S. P. Muller, N. Heineking, and M. C. L. Gerry, J.
Chem. Soc., Faraday Trans. 91, 3347 ͑1995͒.
¨
quired. Solution of the time-dependent Schrodinger equation
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31

6640
J. Chem. Phys., Vol. 114, No. 15, 15 April 2001
Mackey et al.
42 A. M. Mirri, R. Carvellati, and G. Cazzoli, J. Mol. Spectrosc. 71, 386
͑1978͒.
50 C. X. W. Qian, A. Ogai, L. Iwata, and H. Reisler, J. Chem. Phys. 92, 4296
͑1990͒.
43 A. M. Mirri, R. Cervellati, and G. Cazzoli, J. Mol. Spectrosc. 83, 202
͑1980͒.
51 D. Solter, H.-J. Werner, M. von Dirke, A. Untch, A. Vegiri, and R.
Schinke, J. Chem. Phys. 97, 3357 ͑1992͒.
44 J. D. Rogers, W. J. Pietenpol, and D. Williams, Phys. Rev. 83, 431 ͑1951͒.
45 N. Basco and R. G. W. Norrish, Proc. R. Soc. London, Ser. A 268, 291
͑1962͒.
52 G. Herzberg, Molecular Spectra and Molecular Structure I. Spectra of
Diatomic Molecules, 2nd ed. ͑Krieger, Melbourne, 1989͒.
53 B. R. Johnson and W. P. Reinhardt, J. Chem. Phys. 85, 4538 ͑1986͒.
54 B. R. Johnson, J. Chem. Phys. 92, 574 ͑1990͒.
55 D. Imre, J. L. Kinsey, A. Sinha, and J. Krenos, J. Phys. Chem. 88, 3956
͑1984͒.
56 B. R. Johnson, J. L. Mackey, and J. L. Kinsey, J. Comp. Phys. ͑to be
published͒.
57 E. J. Heller, J. Chem. Phys. 62, 1544 ͑1975͒.
46 D. W. Chandler and P. L. Houston, J. Chem. Phys. 87, 1445 ͑1987͒.
47 C. R. Bailey and A. B. D. Cassie, Proc. R. Soc. London 145A, 336 ͑1934͒.
48 R. Schinke, Photodissociation Dynamics: Spectroscopy and Fragmenta-
tion of Small Polyatomic Molecules ͑Cambridge University Press, New
York, 1993͒.
49 U. Manthe, H.-D. Meyer, and L. S. Cederbaum, J. Chem. Phys. 97, 3199
͑1992͒.
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:
202.28.191.34 On: Sun, 21 Dec 2014 22:06:31