I. Three-center versus four-center HCl-elimination in photolysis of vinyl chloride at 193 nm: Bimodal rotational distribution of HCl (v≤7) detected with time-resolved Fourier-transform spectroscopy

Article abstract of DOI:10.1063/1.1328736

The vibration-rotational emission spectra of HCl in the spectral region 2000-3310/cm were temporally resolved with a step scan Fourier transform spectrometer, after the photodissociation of vinyl chloride at 193 nm. All vibrational levels show a bimodal r

Full text of DOI:10.1063/1.1328736

I. Three-center versus four-center HCl-elimination in photolysis of vinyl chloride at 193
nm: Bimodal rotational distribution of HCl (v7) detected with time-resolved Fourier-
transform spectroscopy
Shiaw-Ruey Lin, Shih-Che Lin, Yu-Chang Lee, Yung-Ching Chou, I-Chia Chen, and Yuan-Pern Lee
Citation: The Journal of Chemical Physics 114, 160 (2001); doi: 10.1063/1.1328736
View online: http://dx.doi.org/10.1063/1.1328736
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/114/1?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Photodissociation of gaseous CH3COSH at 248 nm by time-resolved Fourier-transform infrared emission
spectroscopy: Observation of three dissociation channels
J. Chem. Phys. 138, 014302 (2013); 10.1063/1.4768872
Gas-phase photodissociation of CH3COCN at 308 nm by time-resolved Fourier-transform infrared emission
spectroscopy
J. Chem. Phys. 136, 044302 (2012); 10.1063/1.3674166
Molecular elimination in photolysis of fluorobenzene at 193 nm: Internal energy of HF determined with time-
resolved Fourier-transform spectroscopy
J. Chem. Phys. 121, 8792 (2004); 10.1063/1.1802537
Three-center versus four-center elimination of haloethene: Internal energies of HCl and HF on photolysis of CF 2
CHCl at 193 nm determined with time-resolved Fourier-transform spectroscopy
J. Chem. Phys. 117, 9785 (2002); 10.1063/1.1518028
Three-center versus four-center elimination in photolysis of vinyl fluoride and vinyl bromide at 193 nm: Bimodal
rotational distribution of HF and HBr (v5) detected with time-resolved Fourier transform spectroscopy
J. Chem. Phys. 114, 7396 (2001); 10.1063/1.1343079
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

JOURNAL OF CHEMICAL PHYSICS
VOLUME 114, NUMBER 1
1 JANUARY 2001
I. Three-center versus four-center HCl-elimination in photolysis of vinyl
chloride at 193 nm: Bimodal rotational distribution of HCl „vÏ7…
detected with time-resolved Fourier-transform spectroscopy
Shiaw-Ruey Lin, Shih-Che Lin, Yu-Chang Lee, Yung-Ching Chou, I-Chia Chen,
and Yuan-Pern Leea)
Department of Chemistry, National Tsing Hua University, 101, Sec. 2, Kuang Fu Road, Hsinchu,
Taiwan 30013, Republic of China
͑Received 20 June 2000; accepted 6 October 2000͒
Following photodissociation of vinyl chloride at 193 nm, fully resolved vibration-rotational
emission spectra of HCl in the spectral region 2000–3310 cmϪ1 are temporally resolved with a
step-scan Fourier-transform spectrometer. Under improved resolution and sensitivity, emission from
HCl up to ϭ7 is observed, with JϾ32 ͑limited by overlap at the band head͒ for ϭ1–3. All
v
v
vibrational levels show bimodal rotational distribution with one component corresponding to
ϳ500 K and another corresponding to ϳ9500 K for р4. Vibrational distributions of HCl for both
v
components are determined; the low-J component exhibits inverted vibrational population of HCl.
Statistical models are suitable for three-center ͑␣, ␣͒ elimination of HCl because of the loose
transition state and a small exit barrier for this channel; predicted internal energy distributions of
HCl are consistent but slightly less than those observed for the high-J component. Impulse models
considering geometries and displacement vectors of transition states during bond breaking predict
substantial rotational excitation for three-center elimination of HCl but little rotational excitation for
four-center ͑␣, ␤͒ elimination; observed internal energy of the low-J component is consistent with
that predicted for the four-center elimination channel. Rate coefficients 33.8 and 4.9ϫ1011 sϪ1 for
unimolecular decomposition predicted for three-center and four-center elimination channels,
respectively, based on Rice-Ramsberger-Kassel-Marcus theory are consistent with the branching
ratio of 0.81:0.19 determined by counting vibrational distribution of HCl to р6 for high-J and
v
low-J components. Hence we conclude that observed high-J and low-J components correspond to
HCl ( , J) produced from three-center and four-center elimination channels, respectively. © 2001
v
American Institute of Physics. ͓DOI: 10.1063/1.1328736͔
I. INTRODUCTION
tions of vibronic states of HCl ( р4) produced from pho-
v
tolysis of various chloroethylene at ␭у155 nm. He con-
cluded that HCl is produced primarily via a four-center ͑␣,
␤͒ elimination channel and rationalized the observed non-
statistical vibrational distribution of HCl according to an im-
pulse mechanism in which the ‘‘localized’’ energy is parti-
tioned between internal energy of HCl and relative
translational energy of the fragments. Moss et al.3 recorded
infrared ͑IR͒ emission spectra after photolysis of CH2CHCl
at 193 nm and found substantial vibrational excitation of
C2H2 and/or CHCCl and much rotational excitation in HCl.
Donaldson and Leone4 used a Fourier-transform infrared
͑FTIR͒ spectrometer that synchronously triggers the photoly-
sis laser at 193 nm to record time-resolved IR emission from
dissociation of CH2CHCl. They determined a vibrational dis-
Vinyl chloride (CH2CHCl) serves as a model unsatur-
ated hydrocarbon for which dissociation might proceed via
multiple channels during photolysis. A recent review is given
by Blank et al.1 who employed photofragment translational
spectroscopy ͑PTS͒ using vacuum ultraviolet ͑VUV͒ syn-
chrotron radiation for ionization of products and observed
five primary dissociation channels following excitation of
CH2CHCl at 193 nm. Most Cl atoms are translationally hot
and originate from dissociation via a hypersurface of an elec-
tronically excited state. The remaining channels proceed on
the ground electronic surface following internal conversion
from the optically prepared state; these include elimination
of H, H2, HCl, and Cl ͑translationally cold͒. Two secondary
channels eliminating H after elimination of Cl, and Cl after
elimination of H from CH2CHCl, are also identified. In this
paper, we focus only on channels that eliminate HCl from
CH2CHCl; these channels might proceed via a transition
state involving three or four centers.
tribution of HCl (1р р4) slightly cooler than that reported
v
by Berry,2 and proposed that available energy is partitioned
statistically into the product modes of a transition state that
has a small vibrational wave number (ϳ1500 cmϪ1) for
H–Cl stretching. Blank et al.1 reported that the photoioniza-
tion energy threshold of HCl produced from photolysis of
vinyl chloride decreases by ϳ2.2 eV from thermalized HCl,
indicating that the HCl fragment has substantial internal
energy.
Berry2 employed a chemical laser to determine distribu-
a͒
Author to whom correspondence should be addressed; jointly appointed by
the Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei,
Taiwan. Electronic mail: yplee@mx.nthu.edu.tw
0021-9606/2001/114(1)/160/9/$18.00
160
© 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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

J. Chem. Phys., Vol. 114, No. 1, 1 January 2001
Photolysis of vinyl chloride
161
The only report on rotational energy distribution of HCl
was by Reilly et al.5 who performed pump-probe experi-
ments with a molecular beam containing CH2CHCl and de-
tected Cl and HCl after photolysis at 193 nm with
(2ϩ1)-REMPI ͑resonantly enhanced multiphoton ioniza-
tion͒ followed by time-of-flight ͑TOF͒ detection. They found
a bimodal rotational distribution with temperatures corre-
three-center over four-center elimination of CH2CHCl. Fol-
lowing further experiments and a theoretical study by Riehl
and Morokuma,12 Huang et al.7 proposed a mechanism that
includes ␣, ␤-migration of H that competes with the three-
center ␣, ␣-elimination of HCl, with no significant contribu-
tion from the four-center elimination.
We have demonstrated that step-scan time-resolved
Fourier-transform spectroscopy ͑TR-FTS͒ provides much
improved resolution and sensitivity over previous IR emis-
sion techniques.13,14 Here we report emission from HCl ͑1
sponding to 340 and 22 600 K for HCl in the ϭ0 state. In
v
contrast, they observed a Boltzmann-like distribution corre-
sponding to rotational temperatures of 2100Ϯ250 and 1850
Ϯ140 K for ϭ1 and 2 states of HCl, respectively. Huang
v
р р7, Jр32͒ during photolysis of CH CHCl at 193 nm by
means of TR-FTS. A bimodal rotational distribution of HCl
is observed for all vibrational states, yielding enhanced un-
derstanding of HCl-elimination channels.
v
2
et al.6 photolyzed d1-vinyl chloride (CH2CDCl) at 193 nm
and found that the rotational state distributions of HCl prod-
uct are nearly identical for CH2CDCl and CH2CHCl.
Later, they used velocity-aligned Doppler spectroscopy
II. EXPERIMENT
͑VADS͒ to determine the speed distribution of HCl ͑
v
ϭ0–2, J͒ produced in photolysis of CH2CHCl at 193 nm
and found that released kinetic energy much exceeds that
The apparatus employed to obtain step-scan time-
resolved Fourier-transform spectra resembles that described
previously.13,14 An ArF excimer laser ͑Lambda Physik
LPX120i͒, operated at 30–60 Hz with pulse energy 8–10 mJ
was employed as a photolysis source. A telescope served to
focus the laser beam to about 20 mm2 at the reaction center
with a fluence of 40–50 mJ cmϪ2. We estimated a photolysis
yield Ͻ50% in the irradiated region based on an absorption
cross section of 1.9ϫ10Ϫ17 cm2 for CH2CHCl.2 IR emission
was collected with a set of Welsh mirrors and directed into
the Fourier-transform spectrometer ͑Bruker IFS66v͒ through
two CaF2 lenses. A CaF2 beam splitter and an InSb detector
predicted with
a
statistical model for three-center
elimination.7 To reconcile this situation they proposed that
three-center elimination and isomerization from vinylidene
(CCH2) to acetylene ͑HCCH͒ occur in a concerted but non-
synchronous fashion. The isomerization is expected to be
rapid enough to share its exothermicity with nearby HCl
fragment so that HCl departs with augmented vibrational and
translational energies. The dichotomy between populations
of rotational states of HCl with ϭ0 and HCl with vϾ0 is
v
explained according to a vibrationally adiabatic mechanism
in which the adiabatic barrier disappears entirely for chan-
cooled to 77
K were used. Filters passing either
2850–3310 cmϪ1 ͑OCLI, W03024-6 and W03999-4͒ or
2000–2900 cmϪ1 ͑OCLI, W04212-4͒ and an iris were placed
in the sample compartment of the spectrometer. The detected
transient signal, amplified with a gain of 1ϫ105 V/A ͑band-
width 1.5 MHz͒, was further amplified with a low-noise volt-
age amplifier ͑bandwidth 1 MHz, gain typically set at 50͒,
and sent to the internal A/D converter ͑16 bit, 200 kHz͒ of
the spectrometer. The response time of the IR detector is
ϳ0.7 ␮s. The interval of data acquisition in each time slice
is 5 ␮s; hence a datum at t ␮s represents an average of signal
in a range tϮ2.5 ␮s. The first datum was acquired ϳ2.5 ␮s
͑designated as 0–5 ␮s͒ after laser irradiation. Typically 300–
350 time slices were acquired at 5 ␮s intervals and the signal
was averaged for 50–60 laser pulses at each scan step; 3416
and 5979 scan steps were performed to yield a spectrum with
resolution 0.3 cmϪ1 in the spectral range 2850–3310 and
2000–2900 cmϪ1, respectively.
nels associated with HCl with Ͼ0.
v
The relatively large kinetic energy of HCl derived from
VADS is consistent with results from PTS. Average energies
determined with VADS are 25Ϯ2, 21Ϯ2, and 18Ϯ1
kcal molϪ1 for HCl ( ϭ0–2), respectively.7 Umemoto
v
et al.8 photolyzed CH2CHCl at 193 nm and determined the
kinetic energy distribution of HCl with an average transla-
tional energy of 15Ϯ1 kcal molϪ1 by means of PTS at a
fixed scattering angle of 90°. An average energy of
18Ϯ1 kcal molϪ1 deduced by Blank et al.1 is consistent with
results from VADS if substantial population of HCl with
v
Ͼ2 is assumed.
Varied branching ratios between channels for three-
center and four-center elimination of HCl during photolysis
of vinyl chloride at several wavelengths were reported.
Berry2 reported a quantum yield of HCl from CH2CDCl 92%
of that from CH2CHCl for ␭у155 nm, indicating that four-
center HCl-elimination is the major channel. Ausloos et al.9
observed a nearly equal yield from three-center and four-
center elimination in the spectral region 200–220 nm, with
the former decreasing in importance at shorter wavelengths.
Experiments using GC-MS ͑gas chromatography—mass
spectrometry͒ detection of C2H2 and C2HD upon IR multi-
photon dissociation of deuterated isotopomers of CH2CHCl
reveal a preference for a three-center elimination of HCl; a
ratio of 2.3 was reported for three-center to four-center HCl-
elimination channels.10,11 REMPI measurements of Huang
et al.6 produced ratios of 3.9–4.0 for yields of HCl produced
from CH2CHCl to that from CH2CDCl. These authors origi-
nally interpreted the results as signifying a 3:1 preference for
In some experiments we employed the same InSb detec-
tor and amplifiers as described previously but digitized the
signal with an external board ͑PAD1232, 40 MHz, 12 bit
ADC͒ at 25 ns resolution. The first 40 spectra thus obtained
were subsequently averaged to yield a satisfactory spectrum
representing emission in the period 0–1 ␮s after photolysis.
A blackbody source ͑Graseby, model 564, maintained at
1273 K͒ was employed to determine the response function of
the instrument, as described previously.13,14
The partial pressure of CH2CHCl was kept in the range
110–180 mTorr. Ar ͑240–420 mTorr͒ was added near the
entrance photolysis port to avoid brown deposit on the quartz
window. The pressure of the system was measured with a
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

162
J. Chem. Phys., Vol. 114, No. 1, 1 January 2001
Lin et al.
III. RESULTS
Data acquisition in two modes was performed with iden-
tical detector and amplifiers. The internal 16 bit A/D con-
verter of the spectrometer provides temporal resolution of 5
␮s with satisfactory sensitivity, whereas the external
PAD1232 board provides an effective temporal resolution
ϳ1 ␮s with decreased signal-to-noise ͑S/N͒ ratio. In this pa-
per, we focus on only the first available acquisition window
that corresponds to a near collisionless condition to obtain
information on nascent vibration-rotational distributions of
HCl after photolysis. Modeling of temporal profiles to pro-
vide kinetic information and branching ratios of other chan-
nels will be presented in a forthcoming paper.15
Figure 1 shows representative emission spectra recorded
at a resolution of 8 cmϪ1 after photolysis of a flowing mix-
ture of CH2CHCl and Ar ͑0.13 and 1.57 Torr, respectively͒
at 193 nm. Emission from highly excited HCl and C2H2 were
observed immediately after laser photolysis ͑trace A͒; at a
later stage ͑traces B and C͒ these molecules were quenched
FIG. 1. Infrared emission spectra of HCl recorded at varied intervals after
photolysis of CH2CHCl ͑130 mTorr͒ in Ar ͑1.57 Torr͒ at 193 nm. ͑a͒ 2 ␮s,
͑b͒ 17 ␮s, ͑c͒ 47 ␮s. Spectral resolution 8 cmϪ1; averaged over 50 laser
pulses at each interferometer scan step.
to lower vibrational states. Overtone (⌬ ϭϪ2) emission of
v
HCl in the range 4500–6000 cmϪ1 was also observed.
To limit the period of data acquisition to ϳ2 h so as to
assure stability of experimental conditions, high-resolution
spectra were recorded in two sections ͑2000–2900 and
2850–3310 cmϪ1) with proper filters placed in the optical
path. The variations in relative distribution of duplicated
lines in overlapped sections are typically within 5%. For im-
proved accuracy, we normalized these two spectra according
to overlapped lines.
capacitance manometer ͑MKS Baratron, model 121, 0–10
Torr͒ and flow rates were measured with flow meters cali-
brated according to standard procedures. CH2CHCl ͑Merck,
99.95%͒ was used without purification except for degassing;
no impurity was detected in IR spectra.
FIG. 2. High-resolution emission spectra of HCl in two spectral regions, 3200–2850 and 2900–2510 cmϪ1, recorded 0–5 ␮s after photolysis of CH2CHCl
͑130 mTorr͒ in Ar ͑380 mTorr͒. Spectral resolution 0.3 cmϪ1; averaged 100 laser pulses at each scan step. The instrument response is corrected. Assignments
are shown in stick diagrams.
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

J. Chem. Phys., Vol. 114, No. 1, 1 January 2001
Photolysis of vinyl chloride
163
eters reported by Arunan et al.16 and Coxon and
Roychowdhury.17
To study the quenching effect, we recorded emission
spectra at 25 ns intervals with an external PAD1232 board.
Four separate spectra, recorded under identical conditions,
were summed to improve the S/N ratio. Further averaging of
spectra recorded in time slices 1–24, 1–40, 41–80, and 80–
160 yields spectra corresponding to temporal ranges 0–0.6,
0–1, 1–2, and 2–4 ␮s, respectively. Each vibration-
rotational line in the R-branch was normalized with the in-
strument response function, integrated, and divided by its
respective Einstein coefficient16 to yield relative population
Pv(J). For partially overlapped lines, deconvolution was ap-
plied before integration.
We found that quenching is non-negligible for data re-
corded with 5 ␮s resolution; nearly collisionless conditions
are attained for tр1 ␮s. A more detailed discussion and list-
ing of data may be found in the Electronic Physics Auxiliary
Publication Service ͑EPAPS͒.18 The semilogarithmic plots of
P (J)/(2Jϩ1) of HCl recorded in the 0–1 ␮s range for
v
v
ϭ1–6 produced from CH2CHCl at 0.110 Torr and Ar at
0.240 Torr are shown in Fig. 3, in which and J indicate
v
vibrational and rotational levels of the emitting state. Bimo-
dal rotational distributions observed for all vibrational levels
are fitted with biexponential functions to yield two rotational
temperatures. We denote these two components as high-J
and low-J components. Fitted rotational temperatures and
population of both high-J and low-J components in each
vibrational state observed in a few typical experiments are
compared in Table I. Figure 4 illustrates the deconvoluted
high-J and low-J components and the fitted overall popula-
tion distribution for experiment 1b.
FIG. 3. Semilogarithmic plots of relative rotational population of HCl ͑
v
ϭ1–6, symbol ᭺͒ averaged 0–1␮s after photolysis of CH2CHCl ͑110
mTorr͒ in Ar ͑240 mTorr͒ at 193 nm. Data reported by Reilly et al. ͑Ref. 5,
symbol ⌬͒ are shifted vertically for clarity. Solid lines represent least-
squares fits of low-J and high-J components.
Figure 2 shows a partial emission spectrum of HCl re-
corded 0–5 ␮s after photolysis of CH2CHCl ͑0.130 Torr͒
and Ar ͑0.380 Torr͒ at a resolution of 0.3 cmϪ1. The spec-
trum clearly illustrates highly rotational excitation of HCl
IV. DISCUSSION
A. Bimodal rotational distribution
with obvious R bandheads for ϭ1–3. Lines associated
v
Ј
with
up to 7 were observed. Partial assignments are
It is clear from Figs. 3 and 4 and Table I that photolysis
of CH2CHCl produces highly rotationally excited HCl in vi-
v
Ј
shown in Fig. 2 as stick diagrams, based on spectral param-
TABLE I. Experimental conditions, fitted rotational temperature and total rotational population of HCl ( ) after photolysis of CH CHCl at 193 nm.
v
2
High-J component
Low-J component
a
a
Expt.
no.
t
PCH2CHCl
/mTorr
PAr
/mTorr
P J͒
P J͒
͑
v
͑
v
͚
͚
/␮s
Temp/K
J
Erot /kJ molϪ1
Temp/K
J
Erot /kJ molϪ1
v
c
c
1b
110
240
0–1.0
1
2
3
4
5
6
1
2
3
4
5
6
9200Ϯ1500
11 000Ϯ1800
10 000Ϯ1800
10 000Ϯ3500
6000Ϯ2100
3500Ϯ1200
8500Ϯ1600
11 000Ϯ2000
6000Ϯ1200
4500Ϯ800
2670b͑2320͒
2410͑2105͒
2015͑1730͒
1640͑1195͒
1200͑1080͒
630͑630͒
60.3b͑45.8͒
530Ϯ200
500Ϯ150
440Ϯ140
580Ϯ200
520Ϯ200
360Ϯ200
520Ϯ220
580Ϯ160
520Ϯ140
400Ϯ100
320Ϯ100
420Ϯ120
520
700
670
420
180
105
1100
1820
1470
660
3.7
3.6
3.3
3.6
2.4
3.1
5.1
4.4
4.2
3.0
2.0
3.0
60.6͑48.8͒
49.7͑38.9͒
42.3͑27.5͒
28.0͑23.8͒
15.3͑15.3͒
57.2͑44.0͒
58.1 ͑47.2͒
36.4͑29.8͒
29.1͑12.9͒
20.5͑18.6͒
17.6͑17.6͒
2
115
236
0–5.0
5385͑4755͒
4070͑3600͒
2620͑2415͒
2560͑2240͒
1555͑1495͒
820͑820͒
3200Ϯ800
4200Ϯ1200
380
290
aPv(J)ϭ͑integrated emittance͒ϫ10 000/͓͑instrumental response factor͒͑Einstein coefficient͔͒.
bFitted values; extrapolated populations up to Jϭ42 are included. See text.
cSummed values are listed in parentheses; only observed levels are summed. See text.
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

164
J. Chem. Phys., Vol. 114, No. 1, 1 January 2001
Lin et al.
clarity, are consistent with our data except for the absence of
high-J parts. The absence of high-J lines of HCl in the
v
Ͼ0 states in their work might be due to lack of sensitivity
for these transitions in the REMPI scheme. Observation of
low-J components of HCl in their experiments under colli-
sionless conditions in a supersonic jet also supports our con-
clusion that observed low-J components are nascent, rather
than from rotational quenching.
The average rotational energy of each vibrational state of
HCl is calculated on summing a product of rotational energy
and normalized population for each observed rotational
level, Erotϭ͚JPv(J)Er(J)/͚JPv(J), and is referred to as a
‘‘summed value,’’ as listed parenthetically in columns Erot of
Table I. Lines associated with large J in a state with large
v
are more difficult to detect because sensitivity is limited. We
assume a Boltzmann distribution for high-J components and
associate an extrapolated population with unobserved lines
up to Jmax( )ϭ42, 38, 35, 31, 26, 20 for ϭ1–6, respec-
v
v
tively; they correspond to a vibration-rotational energy of
ϳ19 300 cmϪ1, the highest observed value in ϭ6. Rota-
v
tional energies thus obtained are referred to as ‘‘fitted val-
ues’’ in Table I. Fitted and summed values for the low-J
component are nearly identical; hence only fitted values are
listed in Table I. The average rotational energy of the high-J
component ͑averaged for experiments 1a and 1b͒ is obtained
by multiplying Erot in Table I by associated population for
each vibrational level; energies 47Ϯ2 and 37Ϯ1 kJ molϪ1
are determined from fitted and summed values, respectively.
The average rotational energy of the low-J component is
3.8Ϯ0.3 kJ molϪ1; the error limit only reflects deviations in
averaging.
FIG. 4. Deconvoluted relative rotational distributions of high-J and low-J
components of HCl ( ϭ1–6) recorded 0–1 ␮s after photolysis of
v
CH2CHCl ͑110 mTorr͒ in Ar ͑240 mTorr͒. Solid lines represent fitted over-
all population and open circles are experimental data.
brational states р6. As pressure and the time of the detec-
v
tion window are decreased to attain nearly collisionless con-
ditions as in experiment 1b, the rotational temperature of
high-J and low-J components converge to ϳ9500Ϯ2000 K
B. Vibrational distribution of HCl„v…
We summed and normalized relative populations Pv(J)
of observed rotational lines associated with each vibrational
state to obtain a relative vibrational population ͑referred to as
‘‘summed population’’͒. Summation of observed and ex-
and ϳ500Ϯ200 K for HCl ( р4), respectively. Hence data
v
in experiments nos. 1a ͑0–0.6 ␮s͒ and 1b ͑0–1 ␮s͒ are used
to determine nascent populations.
Relative rotational populations reported by Reilly et al.,5
shown in Fig. 3 with symbol ⌬ but shifted downward for
trapolated populations for J up to Jmax ( ) yields an ex-
pected vibrational population ͑referred to as ‘‘fitted popula-
v
TABLE II. Relative vibrational populations of high-J and low-J components of HCl from photolysis of CH2CHCl at 193 nm.
Expt. no.
ϭ1
ϭ2
ϭ3
ϭ4
ϭ5
ϭ6
v
v
v
v
v
v
High-J component
0–1 ␮sa
0.264Ϯ0.011c
(0.265Ϯ0.012)d
0.310Ϯ0.007c
(0.308Ϯ0.002)d
0.393
0.213Ϯ0.016
(0.218Ϯ0.021)
0.239Ϯ0.001
(0.240Ϯ0.005)
0.254
0.184Ϯ0.006
(0.182Ϯ0.013)
0.165Ϯ0.011
(0.169Ϯ0.011)
0.160
0.156Ϯ0.001
(0.140Ϯ0.011)
0.154Ϯ0.004
(0.148Ϯ0.002)
0.099
0.118Ϯ0.005
(0.121Ϯ0.002)
0.084Ϯ0.008
(0.087Ϯ0.011)
0.059
0.066Ϯ0.006
(0.076Ϯ0.006)
0.049Ϯ0.001
(0.053Ϯ0.000)
0.034
0–5 ␮sb
e
PST͑3-center͒
e
SSE͑3-center͒
0.354
0.246
0.168
0.112
0.073
0.047
Low-J component
0–1 ␮sa
0–5 ␮sb
0.194Ϯ0.005
0.190Ϯ0.002
0.272Ϯ0.002
0.325Ϯ0.006
0.255Ϯ0.005
0.260Ϯ0.003
0.151Ϯ0.013
0.120Ϯ0.005
0.078Ϯ0.010
0.066Ϯ0.001
0.051Ϯ0.011
0.047Ϯ0.004
aAverage of experiments 1a and 1b.
bAverage of experiments 2 and 3.
cDerived from fitted values; see text.
dDerived from summed values; see text.
eFor three-center elimination with ␯HClϭ1500 cmϪ1 and Eavaϭ341 kJ molϪ1
.
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

J. Chem. Phys., Vol. 114, No. 1, 1 January 2001
Photolysis of vinyl chloride
165
TABLE III. Comparison of average internal energy of high-J and low-J
components of HCl with statistical and impulse model calculations.
Three-center
eliminationa
Four-center
eliminationb
High-J
Low-J
component component PST SSE Impulsec
Impulsec
2.8
Erot /kJ
Evib /kJ
3.8Ϯ0.3
81Ϯ2f
37
55
36
66
52
47Ϯ2
(37Ϯ1)d
74Ϯ3e
aAvailable energy 341 kJ molϪ1
bAvailable energy 515 kJ molϪ1
.
.
cAssuming that the impulse force is along the displacement vector corre-
sponding to the imaginary mode of the transition state predicted with the
B3LYP method; see text.
dFrom summed values; see text.
eIncluding only ϭ0–6; Eϭ80 kJ molϪ1if extrapolated population of
v
v
v
ϭ7–8 are included.
fIncluding only ϭ0–6; Eϭ88 kJ molϪ1 if extrapolated population of
v
ϭ7–8 are included.
Ϯ1 kJ molϪ1 ͑summed value͒ for the high-J component in-
clude ϭ0 to 6 levels of HCl. If vibrational distributions are
v
extrapolated to ϭ8, average vibrational energies 80 ͑high-
v
J͒ and 88 kJ molϪ1 ͑low-J͒ are derived.
FIG. 5. Comparison of relative vibrational distributions of HCl for both
high-J and low-J components after photolysis of CH2CHCl ͑110 mTorr͒ in
Ar ͑240 mTorr͒ at 193 nm. ᭡: this work ͑from ‘‘fitted population’’͒; छ:
Donaldson and Leone ͑Ref. 4͒; ᭺: Berry ͑Ref. 2͒; solid lines: SSE calcula-
tion; dotted lines: PST calculations; see text and Table II.
The vibrational population of the low-J component is
inverted, with ϭ2 having the greatest population. The
v
population of ϭ0 is estimated by fitting Fig. 5 with a
v
smooth curve; relatively large uncertainties of the small
value 0.13Ϯ0.05 have little effect on calculations of average
vibrational energy. The relative distribution for у3 corre-
v
tion’’͒. Averaged vibrational distribution for high-J and low-
J components recorded in a short period ͑experiments 1a and
1b͒ and a long period ͑experiments 2 and 3͒ are listed in
Table II for comparison. The vibrational distributions thus
derived from summed population and fitted population of the
high-J component are similar. Distributions derived from fit-
ted values are used for discussion, unless noted.
Vibrational distributions of high-J and low-J compo-
nents are compared with previous reports and statistical pre-
dictions in Fig. 5. The vibrational distribution of the high-J
component is nearly Boltzmann, corresponding to a vibra-
sponds to a temperature slightly smaller than that observed
for the high-J component. Consequently, the averaged vibra-
tional energy 81Ϯ2 kJ molϪ1 for the low-J component (
v
ϭ1–6) is only slightly greater than that of the high-J com-
ponent. Observed averaged rotational and vibrational ener-
gies of both high-J and low-J components are compared
with those according to various model calculations for three-
center and four-center elimination channels in Table III.
C. Statistical model calculations for the three-center
elimination of HCl
tional temperature of 26 000Ϯ3000 K for HCl with
v
ϭ1–5. Observed population of ϭ6 is slightly lower than
For a dissociation channel without exit barrier, statistical
v
19,20
that expected for a Boltzmann distribution; presumably
quenching of this state is still not negligible. Observed dis-
tribution agrees well with those reported by Berry.2 Vibra-
tional distribution reported by Donaldson and Leone4 ap-
pears to correspond to a slightly smaller temperature,
presumably due to vibrational quenching in a period 15–30
␮s before their data acquisition.
theories such as phase space theory ͑PST͒
and separate
21
statistical ensemble ͑SSE͒ are expected to predict satisfac-
torily distribution of rotational and vibrational states of prod-
ucts. Energies of three- and four-center HCl-elimi-
nation channels are shown in Fig. 6 according to theoreti-
cal calculations of Riehl and Morokuma12 using
QCISD͑T͒/6-311ϩG(d,p) at MP2/6-31G(d,p) optimized
geometries, with zero-point energy corrected. The three-
center elimination channel has a small (ϳ11 kJ molϪ1) exit
Based on observed vibrational distribution for ϭ1–5
v
of the high-J component, we estimate by linear extrapolation
the relative population of ϭ0 of HCl to be 0.32Ϯ0.03 in
͑reverse͒ barrier. Given the large amount (ϳ341 kJ molϪ1
)
v
Fig. 5. Because the population in Fig. 5 was normalized for
of excess energy for this channel, one expects statistical
theories to describe adequately the distribution of product
states. Because experiments were performed under bulk con-
ditions in which parent molecules were prepared with me-
dium rotational angular momentum, constraints on angular
momentum affect negligibly the observed population in this
work, as suggested in the case of HF-elimination from
HCl ( ϭ1–6), this value indicates that about 24.2% ͑0.32/
v
1.32͒ of the total population of HCl is in its ϭ0 state.
v
The average vibrational energy may be determined from
observed and extrapolated vibrational population. Berry re-
ported a vibrational energy of 58 kJ molϪ1 for HCl (
v
ϭ0–4). Our values of 74Ϯ3 ͑fitted value͒ and 72
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

166
J. Chem. Phys., Vol. 114, No. 1, 1 January 2001
Lin et al.
FIG. 7. Geometries of transition states for three-center elimination ͑TS3͒
and four-center elimination ͑TS4͒ of HCl from vinyl chloride predicted with
the B3LYP/aug-cc-pVTZ method. Results by Riehl and Morokuma ͑Ref.
12͒ using MP2/6-31G(d,p) are also shown in parentheses for comparison.
Bond lengths are in Å. Displacement vectors corresponding to imaginary
wave numbers are shown in the lower part.
FIG. 6. Energies ͑in kJ molϪ1
and dissociation products relative to CH2CHCl. Values using
QCISD͑T͒/6311ϩG(d,p) at MP2/6-31G(d,p) optimized geometries are
taken from Riehl and Morokuma ͑Ref. 12͒ with zero-point energy corrected.
͒ of transition states ͑TS3 and TS4͒
greater than that predicted for three-center HCl-elimination,
but similar to that predicted with the PST model for the
four-center elimination. The rotational energy of the low-J
component is much smaller than that predicted statistically
for either elimination channel. This suggests that the low-J
component may be associated with a dynamically controlled
dissociation process that typically occurs with a tight transi-
tion state.
CH2CFCl.22 Hence, the total angular momentum was not
constrained in our model calculations. We employed vibra-
tional wave number of vinylidene predicted by Chang et al.23
and the energies ͑corrected for zero-point energy͒ of precur-
sor and products predicted by Riehl and Morokuma.12 Vibra-
tional wave numbers 1500 cmϪ1 were used for HCl, as sug-
gested by Donaldson and Leone.4
The vibrational distribution of HCl was calculated with
both PST and SSE models. The SSE ͑PST͒ model predicts an
average vibrational energy of 66 (55) kJ molϪ1, slightly
smaller than the experimental value of 74Ϯ3 kJ molϪ1 for
the high-J component. Statistical models predict an average
rotational energy of ϳ37 kJ molϪ1 for HCl, smaller than the
fitted experimental value of 47Ϯ2 kJ molϪ1 but similar to the
summed value of 37Ϯ1 kJ molϪ1 for the high-J component.
Observed internal energies of the high-J component
agree satisfactorily with those predicted for the three-center
elimination. That observed experimental values are slightly
larger than statistical predictions is consistent with a model
proposed previously to explain why observed translational
energy of HCl is greater than statistical prediction.1,7 The
model indicates that the exothermicity of isomerization of
vinylidene to acetylene may be utilized to excite HCl in the
three-center channel.
D. Transition states and the impulse model
Although Riehl and Morokuma12 reported geometry and
energy of transition states for various unimolecular elimina-
tion channels of vinyl chloride, vibrational wave numbers of
transition states were not listed. We performed calculations
on transition states TS3 and TS4, respectively, of three-
center and four-center HCl-elimination channels with
B3LYP/aug-cc-pVTZ density functional theory24,25 using
GAUSSIAN 98 programs.26 Predicted geometry of TS3 and
TS4 is shown in Fig. 7 with structural parameters indicated.
Parameters predicted by Riehl and Morokuma with a frozen-
core Moller-Plesset second-order perturbation method,27
MP2/6-31G(d,p), are also listed parenthetically for com-
parison. Vibrational wave numbers of TS3 and TS4 are listed
in Table IV. The imaginary vibrational wave numbers are
219i and 1559i for TS3 and TS4, respectively; displacement
vectors of corresponding motion of these modes are shown
in the lower part of Fig. 7.
Statistical calculations were also performed for the four-
center channel for comparison, even though one expects un-
satisfactory results in view of the tight transition state and a
large exit barrier for this channel. Rotational energy of 50
(51) kJ molϪ1 and vibrational energy of 97 (82) kJ molϪ1
were derived with the SSE ͑PST͒ model.
The inverted vibrational distribution of the low-J com-
ponent cannot be reproduced with statistical models. Ob-
served average vibrational energy of 81Ϯ2 kJ molϪ1 is much
The rotational energy may be predicted with a model
considering motions of the reaction coordinates derived from
displacement vectors corresponding to the imaginary modes
of both transition states. If a molecule dissociates instanta-
neously once it reaches the transition state, no energy flows
beyond the saddle point. The direction of the repulsive force
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

J. Chem. Phys., Vol. 114, No. 1, 1 January 2001
Photolysis of vinyl chloride
167
TABLE IV. Vibrational wave numbers ͑cmϪ1͒ of transition states TS3 and
TS4 predicted with the B3LYP/aug-cc-pVTZ method.
slightly more proportionate amount of four-center elimina-
tion, but the discrepancies are probably within experimental
uncertainties.
TS3
3115
1239
684
TS4
Based on discussions of Secs. IV C–IV E, we conclude
that observed high-J and low-J components correspond well
with HCl produced via three-center and four-center elimina-
tion, respectively. A model with vibrational adiabaticity was
proposed to explain reported dichotomy between the rota-
tional state distribution of HCl ( ϭ0) and HCl ( Ͼ0).5,7
3192
1463
805
1734
831
477
3414
1548
801
3298
861
695
280
1842
836
624
266
151
344
219i
1559i
v
v
Our data show no obvious variation in bimodal rotational
distributions for all observed vibrational levels.
is assumed to follow the displacement vectors. Hence dis-
placement vectors shown in Fig. 7 imply substantial rota-
tional excitation of HCl in the three-center elimination be-
cause the H atom moves nearly parallel to the C–Cl bond
with a large impact parameter toward the Cl atom. If we
distribute the available energy between H and C2H2 accord-
ing to classical mechanics and calculate the dynamics after
the energized H atom moves toward the Cl atom along the
displacement vector, a rotational energy of 52 kJ molϪ1 is
predicted for three-center elimination with an available en-
ergy of 341 kJ molϪ1. This value is close to the experimental
Assuming that previously reported1,7 average transla-
tional energy ϳ75 kJ molϪ1 is mainly due to the major three-
center elimination channel, our observation of an internal
energy ϳ120 kJ molϪ1 for HCl implies an internal energy
ϳ146 kJ molϪ1 for vinylidene after photolysis at 193 nm.
This value is much smaller than one typically expects, con-
sidering the complexity of vinylidene as compared with HCl.
However, if the exothermicity of isomerization of vinylidene
to acetylene is included in the available energy, the internal
energy of acetylene product becomes ϳ320 kJ molϪ1. For
the four-center elimination channel, if an average transla-
tional energy ϳ75 kJ molϪ1 is used, our results imply an
internal energy of ϳ355 kJ molϪ1 for acetylene. Unfortu-
nately, because of its complexity, we are unable to assign or
simulate observed emission spectrum of acetylene in the 3
␮m region at this stage.
value of 47Ϯ2 kJ molϪ1
.
In contrast, substantial vibrational excitation but little
rotational excitation is predicted according to displacement
vectors of four-center HCl-elimination in Fig. 7 because the
H atom is moving toward Cl with a small impact parameter.
With an available energy of 515 kJ molϪ1 a rotational energy
of 2.8 kJ molϪ1 is predicted for this channel, consistent with
the experimental value of 3.8Ϯ0.3 kJ molϪ1. Hence, the im-
pulse model using displacement vectors of imaginary fre-
quencies describes satisfactorily rotational and vibrational
distributions observed for the low-J components of HCl as
resulting from four-center elimination.
V. CONCLUSION
Rotationally resolved emission from HCl up to ϭ7 is
v
observed in the spectral range 2000–3310 cmϪ1 after pho-
tolysis of vinyl chloride at 193 nm. All vibrational levels
show a bimodal rotational distribution with one component
corresponding to ϳ500 K and the other corresponding to
ϳ9500 K. The two components with low and high rotational
E. Branching ratio and RRKM rates of dissociation
temperatures correspond to HCl ( ,J) produced from four-
v
We estimate the rate of dissociation of vinyl chloride
irradiated at 193 nm with a microcannonical transition state
theory. The Whitten–Robinovitch equations28 were used to
calculate the density of states and number of transition states.
Because elimination of HCl occurs on the ground electronic
surface, we use vibrational wave number of the ground state
of vinyl chloride. The rate of dissociation for three-center
elimination with an energy 330 kJ molϪ1 on the transition
state and four-center elimination with an energy
center and three-center HCl-elimination channels, respec-
tively. The four-center channel produces HCl with little ro-
tation and inverted vibrational population. Statistical models
predict satisfactorily internal energy distribution of HCl from
three-center elimination. Impulse models considering geom-
etries and displacement vectors of transition states during
bond breaking predict the rotational distribution of both
channels satisfactorily. The branching ratio of 0.81:0.19 de-
termined for high-J and low-J components is consistent with
rate coefficients for unimolecular decomposition via three-
center:four-center elimination predicted with the RRKM
theory.
295 kJ molϪ1 are calculated to be 33.8 and 4.9ϫ1011 sϪ1
,
respectively. Accordingly, the branching ratio for formation
of HCl according to three- and four-center processes is esti-
mated to be 0.87:0.13.
Estimates of populations of HCl ( ϭ0) for low-J and
high-J components are described in Sec. IV B. The branch-
v
ACKNOWLEDGMENTS
ing ratio is determined by comparing ͚v,JPv(J) for
v
We thank the National Science Council of the Republic
of China ͑Grant No. NSC90-2113-M-007-008͒ for support
and the National Center for High Performance Computing
for computer time.
ϭ0–6 ͑Table I͒ of high-J and low-J components of HCl.
The estimated ratio 0.81:0.19 is consistent with that pre-
dicted for three-center and four-center elimination channels
with RRKM theory. If population of higher vibrational states
is linearly extrapolated to ϭ8, the branching ratio becomes
0.84:0.16. Previous reports on photolysis of CH2CDCl at 193
nm: 0.75:0.25 ͑Ref. 6͒ and 0.70:0.30 ͑Ref. 11͒ show a
v
1 D. A. Blank, W. Sun, A. G. Suits, and Y. T. Lee, J. Chem. Phys. 108,
5414 ͑1998͒.
2 M. J. Berry, J. Chem. Phys. 61, 3114 ͑1974͒.
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27

168
J. Chem. Phys., Vol. 114, No. 1, 1 January 2001
Lin et al.
3 M. G. Moss, M. D. Ensminger, and J. D. McDonald, J. Chem. Phys. 74,
6631 ͑1981͒.
18 See EPAPS Document No. E-JCPSA6-114-018101 for a complete listing
of data and discussion on quenching effects. This document may be re-
trieved 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.
4 D. J. Donaldson and S. R. Leone, Chem. Phys. Lett. 132, 240 ͑1986͒.
5 P. T. A. Reilly, Y. Xie, and R. J. Gordon, Chem. Phys. Lett. 178, 511
͑1991͒.
6 Y. Huang, Y. A. Yang, G. X. He, and R. J. Gordon, J. Chem. Phys. 99,
2752 ͑1993͒.
19 P. Pechukas and J. C. Light, J. Chem. Phys. 42, 3281 ͑1965͒; C. E. Kolts,
J. Phys. Chem. 75, 1526 ͑1971͒.
7 Y. Huang, Y. A. Yang, G. X. He, S. Hashimoto, and R. J. Gordon, J.
Chem. Phys. 103, 5476 ͑1995͒.
20 M. Hunter, S. A. Reid, D. C. Robie, and H. Reisler, J. Chem. Phys. 99,
1093 ͑1993͒.
8 M. Umemoto, K. Seki, H. Shinohara, U. Nagashima, and N. Nishi, J.
Chem. Phys. 83, 1657 ͑1985͒.
21 C. Wittig, I. Nadler, H. Reisler, M. Noble, J. Catanzarite, and G.
Radhakrishnan, J. Chem. Phys. 83, 5581 ͑1985͒.
9 P. Ausloos, R. E. Rebberts, and M. H. Wijnen, J. Res. Natl. Bur. Stand.,
Sect. A 77, 243 ͑1973͒.
22 T. R. Fletcher and S. R. Leone, J. Chem. Phys. 88, 4720 ͑1988͒.
23 N.-y. Chang, M.-y. Shen, and C.-h. Yu, J. Chem. Phys. 106, 3237 ͑1997͒.
24 A. D. Becke, J. Chem. Phys. 98, 5648 ͑1993͒.
10 F. M. Lussier and J. I. Steinfeld, Chem. Phys. Lett. 50, 175 ͑1977͒.
11 C. Reiser, F. M. Lussier, C. C. Jensen, and J. I. Steinfeld, J. Am. Chem.
Soc. 101, 350 ͑1979͒.
25 C. Lee, W. Yang, and R. G. Parr, Phys. Rev. B 37, 785 ͑1988͒.
26
GAUSSIAN 98 ͑Revision A.1͒, M. J. Frisch, G. W. Trucks, H. B. Schlegel
12 J.-F. Riehl and K. Morokuma, J. Chem. Phys. 100, 8976 ͑1994͒.
13 P. S. Yeh, G. H. Leu, Y.-P. Lee, and I-C. Chen, J. Chem. Phys. 103, 4879
͑1995͒.
et al., Gaussian Inc., Pittsburgh, PA, 1998.
27 W. J. Hehre, R. Dichtfield, and J. A. Pople, J. Chem. Phys. 56, 2257
͑1972͒; P. C. Hariharan and J. A. Pople, Theor. Chim. Acta 28, 213
͑1973͒; M. S. Gordon, Chem. Phys. Lett. 76, 163 ͑1980͒; M. J. Frisch, J.
A. Pople, and J. S. Binkley, J. Chem. Phys. 80, 3265 ͑1984͒.
28 K. A. Holbrook, M. J. Pilling, and S. H. Robertson, Unimolecular Reac-
tions, 2nd ed. ͑Chichester, New York, 1996͒.
14 S.-R. Lin and Y.-P. Lee, J. Chem. Phys. 111, 9233 ͑1999͒.
15 S.-R. Lin, S. C. Lin, and Y.-P. Lee, to be published.
16 E. Arunan, D. S. Setser, and J. F.Ogilvie, J. Chem. Phys. 97, 1734 ͑1992͒.
17 J. A. Coxon and U. K. Roychowdhury, Can. J. Phys. 63, 1485 ͑1985͒.
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:
137.149.200.5 On: Sat, 29 Nov 2014 13:30:27