UV photodissociation of oxalyl chloride yields four fragments from one photon absorption

Article abstract of DOI:10.1063/1.473764

The photodissociation of oxalyl chloride, (ClCO)₂, has been studied near 235 nm using the photofragment imaging technique. Observed products include both ground state Cl (²P_3/2) and spin-orbit excited Cl*(²P_1/2) chlorine atoms and ground electronic state CO molecules. The rotational distribution obtained for the CO v=0 product is peaked at about J=30 and extends beyond J=50. Photofragment images were recorded for both chlorine atom fine structure components as well as many rotational levels of the CO v=0, yielding state-resolved angular and translational energy distributions. The recoil speed distribution for the Cl* exhibits a dominant fast component, with a translational energy distribution peaking at about 48 kJ/mol. The ground state chlorine atom showed two components in its speed distribution, with the slow component dominant. The corresponding translational energy distribution peaked at 10 kJ/mol but extended to 80 kJ/mol. The total average translational energy release into the Cl product is 34 kJ/mol. Similarly, the low rotational levels of the CO showed only a slow component, the intermediate rotational levels showed a bimodal speed distribution, and the highest rotational levels showed only the fast component. The fast components of both chlorine atom product and the higher rotational levels of the CO show an anisotropic angular distribution, while all slow fragments show a nearly isotropic angular distribution. These observations suggest a novel dissociation mechanism in which the first step is an impulsive three-body dissociation yielding predominantly Cl*, rotationally excited CO and chloroformyl radical ClCO, with only modest momentum transfer to the latter species. Most of the remaining ClCO undergoes subsequent dissociation yielding low rotational levels of CO and little translational energy release.

Full text of DOI:10.1063/1.473764

UV photodissociation of oxalyl chloride yields four fragments from one photon
absorption
Musahid Ahmed, David Blunt, Daniel Chen, and Arthur G. Suits
Citation: The Journal of Chemical Physics 106, 7617 (1997); doi: 10.1063/1.473764
View online: http://dx.doi.org/10.1063/1.473764
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/106/18?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Fragmentation mechanism of UV-excited peptides in the gas phase
J. Chem. Phys. 141, 154309 (2014); 10.1063/1.4897158
UV photodissociation dynamics of iodobenzene: Effects of fluorination
J. Chem. Phys. 136, 124313 (2012); 10.1063/1.3696892
Determining the CH 3 SO 2 → CH 3 + SO 2 barrier from methylsulfonyl chloride photodissociation at 193 nm
using velocity map imaging
J. Chem. Phys. 131, 044304 (2009); 10.1063/1.3159556
Nonadiabatic effects in C–Br bond scission in the photodissociation of bromoacetyl chloride
J. Chem. Phys. 125, 194305 (2006); 10.1063/1.2363991
Photodissociation of CH stretch overtone excited CH 3 Cl and CHD 2 Cl (v CH =5): Cl spin–orbit branching and
atomic fragment yields
J. Chem. Phys. 109, 7810 (1998); 10.1063/1.477427
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45

UV photodissociation of oxalyl chloride yields four fragments
from one photon absorption
Musahid Ahmed, David Blunt, Daniel Chen, and Arthur G. Suitsa)
Chemical Sciences Division, Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley,
California 94720
͑Received 4 December 1996; accepted 6 February 1997͒
The photodissociation of oxalyl chloride, ͑ClCO͒2, has been studied near 235 nm using the
photofragment imaging technique. Observed products include both ground state Cl (2P3/2) and
2
*
spin-orbit excited Cl ( P1/2) chlorine atoms and ground electronic state CO molecules. The
rotational distribution obtained for the CO vϭ0 product is peaked at about Jϭ30 and extends
beyond Jϭ50. Photofragment images were recorded for both chlorine atom fine structure
components as well as many rotational levels of the CO vϭ0, yielding state-resolved angular and
*
translational energy distributions. The recoil speed distribution for the Cl exhibits a dominant fast
component, with a translational energy distribution peaking at about 48 kJ/mol. The ground state
chlorine atom showed two components in its speed distribution, with the slow component dominant.
The corresponding translational energy distribution peaked at 10 kJ/mol but extended to 80 kJ/mol.
The total average translational energy release into the Cl product is 34 kJ/mol. Similarly, the low
rotational levels of the CO showed only a slow component, the intermediate rotational levels
showed a bimodal speed distribution, and the highest rotational levels showed only the fast
component. The fast components of both chlorine atom product and the higher rotational levels of
the CO show an anisotropic angular distribution, while all slow fragments show a nearly isotropic
angular distribution. These observations suggest a novel dissociation mechanism in which the first
*
step is an impulsive three-body dissociation yielding predominantly Cl , rotationally excited CO
and chloroformyl radical ClCO, with only modest momentum transfer to the latter species. Most of
the remaining ClCO undergoes subsequent dissociation yielding low rotational levels of CO and
little translational energy release. © 1997 American Institute of Physics.
͓S0021-9606͑97͒01418-9͔
I. INTRODUCTION
chloride have been the subject of a number of studies over
the past forty years, owing to interest in the nature of the
internal rotations about the carbon-carbon bond and the ex-
istence and identity of conformers beyond the lowest energy
anti form.11–14 Surprisingly, there have been no studies of
oxalyl chloride photochemistry in the ultraviolet. One study
reported the thermal decomposition of oxalyl chloride in-
duced by a CO2 laser, with detection of the reactant
(COCl)2 and product COCl2 by infrared absorption.15 The
results were unaltered following addition of radical scaven-
gers NO and propene, suggesting a concerted thermal de-
composition: (COCl)2→COϩCOCl2 . The UV-induced pho-
toisomerization of oxalyl chloride was studied in an argon
matrix using Fourier-transform IR spectroscopy.16 The au-
thors found evidence of a cis form and interconversion be-
tween the cis and a gauche conformer. In addition, UV irra-
diation of oxalyl chloride in a xenon matrix led to the
formation of CO and phosgene; however, this photochemical
process was not the focus of the study and was not reported
in detail.
When a photochemical event leads to the breaking of
several chemical bonds, the timing of the bond scissions and
the energy release to the fragments provide a window into
the chemical dynamics underlying these unusual processes.
This has motivated considerable interest and extensive study
for such systems as s-tetrazine,1 glyoxal2–4 and acetone,5–7
where absorption of a single photon can lead to three product
fragments. To our knowledge, no reports have appeared to
date for systems dissociating to four fragments following
absorption of a single UV photon. The reason for this is
simply that the energetics of breaking three bonds usually
makes it impossible. Oxalyl chloride, ͑ClCO͒2, the chlorine-
substituted analogue of glyoxal, possesses bond dissociation
energies (Do ClCO–ClCOϭ322 kJ/mol and Do Cl–COϭ 25
8
kJ/mol in ClCO͒ such that dissociation into the four frag-
ments ClϩClϩCOϩCO is possible at 235 nm with a further
167 kJ/mol available energy. All of the fragments, further-
more, can be detected via resonant multiphoton ionization
͑REMPI͒ in the vicinity of 235 nm. Oxalyl chloride is thus
an ideal candidate for investigation of multiple bond-
breaking processes via the photofragment imaging
technique.9,10
A recent study using ion time-of-flight explored the pho-
todissociation dynamics of phosgene,17 a closely related sys-
*
tem. In the phosgene study, both Cl and Cl were detected,
The structure and conformational properties of oxalyl
and both were found to exhibit bimodal translational energy
distributions. The ground state chlorine atom distribution
was found to be dominated by the slow component, while for
a͒
Author to whom correspondence should be addressed.
J. Chem. Phys. 106 (18), 8 May 1997
0021-9606/97/106(18)/7617/8/$10.00
© 1997 American Institute of Physics
7617
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45

7618
Ahmed et al.: Photodissociation of oxalyl chloride
in a counterpropagating configuration, are each pumped by
700 l/s high-throughput compound turbomolecular pumps
͑Osaka TG-703M͒, while the main chamber is pumped by a
1000 l/s magnetic bearing turbomolecular pump ͑Seiko-Seiki
STP-1000͒. The latter pump is backed by an oil-free scroll
pump ͑Galiso-Nuvac NDP-18HV͒ providing a completely
oil-free system in the ionization region. The molecular
beams cross at the center of the main chamber on the axis of
a time-of-flight mass spectrometer. The products are ionized
either by REMPI using UV lasers, or direct one-photon ion-
ization using VUV undulator radiation typically introduced
via the smaller ports at 45 deg. Product ions are accelerated
by repeller and acceleration fields into a one meter flight tube
perpendicular to the plane of the beams. Product ions strike
the position-sensitive detector, which is an 80-mm diameter
dual microchannel plate coupled to a phosphor screen ͑Gali-
leo 3075FM͒. Mass-selected images were obtained by puls-
ing the microchannel plate, typically pulsing from a DC
value of Ϫ1100 V to Ϫ1900 V, with 1 microsecond dura-
tion. The detector is viewed by an integrating fast-scan video
camera system ͑Hammamatsu͒ employing thresholding in
conjunction with a linear video look-up table. Future
crossed-beam experiments will employ a custom-built video
system capable of integration with 32-bit depth and opera-
tion at frame rates to 220 Hz ͑Data Design AC-101M͒ and
linear or, in some cases a binary, look-up table for single ion
accumulation. Typical accumulation times were 10 minutes
for each image. Some of the images were re-recorded on a
different apparatus, to be described in detail in a future
publication.21
For the oxalyl chloride photodissociation studies, experi-
ments were performed using a single molecular beam source
and a single laser providing both dissociation and probe light
͑one-color experiments͒. The molecular beam was produced
by bubbling helium through oxalyl chloride held at Ϫ10 °C,
yielding a 3% beam. The mixture was expanded through a
Proch-Trickl piezoelectric pulsed valve22 operating at 30 Hz,
with 150 microsecond pulses. The laser light was generated
using a seeded Nd-Yag-pumped ͑Spectra-Physics GCR 290-
30͒ dye laser ͑Laser Analytic Systems LDL͒ with a Bethune
amplifier cell, operating with pyridine 1 around 700 nm. The
laser light was first doubled in BBO, then the doubled light
mixed with the fundamental to give the third harmonic of the
dye. Polarization was adjusted with a half-wave plate before
the first doubling crystal. Typical output power was 1-2 mJ/
pulse at 30 Hz, and relatively constant throughout the tuning
range from 225-235 nm. This configuration was found to
give both a longer dye lifetime and broader tuning range
when contrasted with doubling a 355 nm pumped dye. The
laser linewidth was narrower than 0.04 cmϪ1 on the dye
fundamental, so that it was necessary to scan across the Dop-
pler width of the lines while each image was recorded.
Mass spectra were recorded by placing a photomultiplier
tube to view the detector. REMPI scans of the CO were also
recorded using the photomultiplier, with the signal fed into a
boxcar averager gated on the mass of interest. REMPI ͑2ϩ1͒
FIG. 1. Schematic view of ‘‘universal imaging’’ machine. Left view is from
above, right view is into the primary beam source.
*
the Cl both fast and slow components were of comparable
magnitude. Based upon the translational energy distributions,
the authors inferred a concerted three-body dissociation to
ClϩClϩCO in the phosgene case. Measurements of the CO
internal state and translational energy distributions were not
reported for that system.
We have undertaken a detailed study of the photochem-
istry of oxalyl chloride near 235 nm, measuring complete
translational energy and angular distributions for all frag-
ments with quantum state resolution, with an eye to explor-
ing the detailed dynamics in this sytem capable of dissocia-
tion to four fragments following one-photon excitation.
These experiments were performed in a new ‘‘universal im-
aging’’ crossed beam apparatus that has been constructed as
a part of the Chemical Dynamics Beamline, a national user
facility recently commissioned at the Advanced Light
Source.18,19
II. EXPERIMENT
A. The ‘‘universal imaging’’ machine
As these represent the first experimental results obtained
using the new ‘‘universal imaging’’ machine ͑so-called ow-
ing to the planned use in conjunction with tunable undulator
VUV probe͒, some detail will be provided concerning the
apparatus. The design is based on that of Houston et al.20
and is shown schematically in Fig. 1. The main chamber is a
modified six-way 12Љ conflat cross, with four additional
2-3/4Љ conflat ports at 45 deg to the main ports in one plane.
Four keying rings have been welded just inside the four main
ports, on which may be bolted the source differential cones.
All components have been manufactured to key together so
that there is no adjustment of the molecular beam sources.
The differential cones are configured so that a mechanical
chopper motor can be mounted to chop the molecular beam
pulses to 5 microseconds or less, for future crossed-beam
experiments or studies using the undulator radiation. The
source regions have ports for access of lasers to the nozzle
regions for photolytic radical production. The molecular
beam sources, typically fixed at 90° but capable of operating
*
detection of Cl and Cl was performed on numerous lines in
the vicinity of 235 nm.23 CO was detected by 2ϩ1 REMPI
J. Chem. Phys., Vol. 106, No. 18, 8 May 1997
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45

Ahmed et al.: Photodissociation of oxalyl chloride
7619
on the (B←X) Q-branch transition at 230 nm.24 The scans
were conducted under conditions with sufficent averaging
that the laser power was constant (Ϯ5%), so that no correc-
tion for laser power was necessary. The relative populations
were obtained from the simulations discussed below follow-
ing correction for the appropriate line strength factor ͑2J
ϩ1͒/9.
The images recorded are 2-dimensional projections of
the 3-dimensional recoiling product sphere. The translational
energy and angular distributions are reconstructed from the
projection using standard techniques.10 In all cases presented
here the translational energy distributions are reported for
each fragment separately, since there are 3-body processes
involved that do not conserve linear momentum between any
two fragments.
III. RESULTS
These experiments represent one-color studies of the
photodissociation of oxalyl chloride at numerous wave-
lengths in a narrow range. As such, the photolysis wave-
length changes slightly with the species probed. Since the
absorption spectrum is not discrete in this region it is as-
sumed, as is commonly done,17,25 that small changes in the
available energy do not dramatically influence the dissocia-
tion dynamics. This is not always the case, even for systems
lacking a discrete absorption spectrum.26 However, the avail-
*
ability of several lines for Cl and Cl provided a valuable
check on this assumption. Both translational energy and an-
gular distributions for the chlorine atom products were insen-
sitive to the particular state probed, confirming both that the
dynamics do not change rapidly with wavelength in this re-
gion, and that orbital alignment of the chlorine atom, if
present, is not influencing the measured distributions. Multi-
photon dissociation processes may be ruled out since the
additional 500 kJ/mol would appear in the translational and
rotational energy distributions shown below.
2
*
FIG. 2. ͑a͒. Raw image of Cl ( P1/2) recorded with dissociation and probe
at 235.205 nm on the 4p2P1/2←←3p2P1/2 transition. ͑b͒ Raw image of
Cl(2P3/2) recorded with dissociation and probe at 235.336 nm on the
4p2D3/2←←3p2P1/2 transition. The laser polarization direction is vertical
in the plane of the images.
A. Chlorine atom translational energy distributions
kJ/mol. The fit yields 22% of the fast component for the
ground state Cl, though the fraction could vary considerably
with the choice for the description of the slow component.
The P(ET) for the ground state Cl atom is shown in Fig. 3.
2
*
Images recorded for the spin-orbit excited ͑Cl ( P1/2))
and ground state ͑Cl (2P3/2)) chlorine atoms are shown in
Figs. 2͑a͒ and ͑b͒, respectively. The corresponding transla-
tional energy distributions (P(ET)’s͒ extracted from the im-
*
In contrast to the Cl , the Cl P(ET) is dominated by the
*
ages are shown in Fig. 3. The Cl P(ET) shows a dominant
slow component. The ground state Cl also shows a contribu-
tion from the ‘‘fast’’ component, shown as a dot-dash line in
Fig. 3. The bimodal nature of the distribution is more appar-
ent in the speed distributions rather than the translational
energy distributions, but the justification for fitting it as two
components is further shown in the angular distributions be-
low. These ‘‘slow’’ and ‘‘fast’’ components used to fit the
fast ͑i.e., high ET) component peaking at 48 kJ/mol, with an
average of 47 and a full width at half-maximum of 38 kJ/
mol. A small ͑8%͒ contribution from a ‘‘slow’’ component is
also present. The translational energy distributions were fit-
ted as two components, shown as dashed lines in the figure.
The slow component peaks at 10 and extends to 80 kJ/mol
with an average of 22 kJ/mol. This component was fitted
*
experimental results are identical for both the Cl and Cl,
P
with the expression P(ET)ϭ(EcolϪEexo
)
(EexoϪB)Q, of-
*
only the relative contributions have been varied in each case.
ten used to describe the statistical decomposition of a colli-
sion complex, using the parameters Ecolϭ 28.9 kJ/mol,
Pϭ0.4, Qϭ5.4 and BϭϪ2 kJ/mol and Eexoϭ86 kJ/mol. No
direct physical significance is ascribed to these parameters,
however. A Gaussian was employed to fit the ‘‘fast’’ com-
ponent. The latter peaks at 48 kJ/mol with an average of 47
B. Chlorine atom angular distributions
The chlorine atom angular distributions derived from the
*
images are shown in Fig. 4. The Cl angular distribution is
clearly anisotropic, yielding an anisotropy parameter ␤ϭ
J. Chem. Phys., Vol. 106, No. 18, 8 May 1997
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45

7620
Ahmed et al.: Photodissociation of oxalyl chloride
FIG. 4. Chlorine atom angular distributions obtained from the images in
2
*
Fig. 2. ͑a͒ Cl ( P1/2). Circles are experimental points and the line is the fit
yielding ␤ϭ0.7 ͑b͒ Cl(2P3/2). Circles are experimental points below 1000
m/s ͑18 kJ/mol͒, ‘‘slow component’’, and line is fit yielding ␤ϭ0.07. Tri-
angles are experimental points above 1000 m/s ͑18 kJ/mol͒, ‘‘fast compo-
nent’’, and solid line is fit yielding ␤ ϭ 0.9.
FIG. 3. Chlorine atom center-of-mass translational energy distributions ob-
tained from the images in Fig. 2. Circles are experimental points and lines
2
are results of fit as described in text. ͑a͒ Cl ( P1/2). ͑b͒ Cl(2P3/2).
*
ther subdivided according to the relative amounts of fast and
slow components, obtained via the fit to the ground state
distribution discussed above. The result gives very close to
half-and-half fast and slow Cl atom product, with the bulk of
0.89, where the ␤ parameter describes the angular distribu-
tion via the well-known expression: I(␪)ϭ(1/4␲)
1ϩ␤(P (cos(␪)) .27 The reconstruction technique accumu-
͓
͔
2
lates the noise in the image into the center strip, so that the
results near the poles are always associated with larger error
bars. The angular distribution for the ground state chlorine
atom are shown in Fig. 4͑b͒. Two different angular distribu-
tions are shown: one corresponding to the total contribution
below 18 kJ/mol translational energy, the other represents
the total above 18 kJ/mol. The faster component shows a
substantially more polarized angular distribution, in fact
*
the fast Cl appearing as Cl . These branching fractions are
summarized in Table I, with the overall results for the aver-
age energy release and ␤ parameters summarized in Table II.
D. REMPI spectrum of CO
The REMPI spectrum for the CO vϭ0 product is shown
in Fig. 5͑a͒, along with a simulation. The isolated peak near
the Q-branch bandhead is well-fitted with a rotational tem-
perature of 300 K, and represents background CO. The vϭ1
distribution ͑not shown͒ is not rotationally resolved and is
not quantified in this study. However, the total integrated
intensity of the ͑1-1͒ band was less than 5% that of the ͑0-0͒
band, implying the predominance of the vϭ0 component de-
spite the uncertainty in the Franck-Condon factors and other
*
quite close to that of the Cl . The ␤ values obtained are 0.07
for the slow fraction and 0.72 for the fast component.
C. Branching fractions
The branching into the different distributions for
*
the Cl and Cl are summarized in Table I. The relative
*
contributions for the Cl and Cl were obtained by scaling
the integrated REMPI signals using the relative line
*
strengths for detection of Cl
at 237.808 nm
TABLE I. Branching in Cl photofragments. See text for discussion of
‘‘fast’’ and ‘‘slow’’ components. Values in percent. Values in parentheses
are obtained based on a revised value for the spin-orbit branching ratio in
HCl at 193 nm ͑Ref. 32͒.
͑‘‘a,’’ 4p2P1/2←←3p2P1/2
)
and 235.205 nm ͑‘‘b,’’
4p2D3/2←←3p2P1/2) with that for Cl at 235.336 nm ͑‘‘c,’’
4p2D3/2←←3p2P3/2). Published values for the relative line
strengths are 2.5Ϯ0.1 for c/a and 0.8Ϯ0.1 for c/b.28–31 Al-
ternative values of 1.8 and 0.6 are obtained for c/a and c/b
Slow
Fast
Total
Cl
Cl
Total
50͑56͒Ϯ7
2͑2͒
52͑58͒Ϯ10
14͑15͒Ϯ7
34͑27͒Ϯ10
48͑42͒Ϯ10
64͑71͒Ϯ10
36͑29͒Ϯ10
*
based on recent measurements of the Cl/Cl branching ratio
*
in HCl by Wittig and co-workers.32 In addition to the spin-
*
orbit branching, yielding 36% Cl , the distributions are fur-
J. Chem. Phys., Vol. 106, No. 18, 8 May 1997
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45

Ahmed et al.: Photodissociation of oxalyl chloride
7621
TABLE II. Energy release and ␤ parameters.
E
͑kJ/mol͒
͘
␤
͗
Cl ͑slow͒
Cl ͑fast͒
Cl ͑total͒
22.1
47.2
27.4
22.1
47.2
45.0
33.8
12.3
29.9
0.07
0.72
*
Cl ͑slow͒
*
Cl ͑fast͒
0.89
0.89
*
Cl ͑total͒
*
ClϩCl
CO ͑vϭ0, Jϭ22͒
CO ͑vϭ0, Jϭ30͒
0.04
Ϫ0.03 ͑slow͒
0.41 ͑fast͒
0.45
CO ͑vϭ0, Jϭ55͒
30.3
aspects of the relative transition strength. The simulation was
used to extract the rotational distribution of the CO vϭ0,
shown in Fig. 5͑b͒. The distribution peaks at Jϭ28, with
levels as low as 15 and beyond 50 populated. It was neces-
sary to include a J-dependent width of the lines in the simu-
lation owing to the broadening of the Doppler width at high
J ͑see below͒.
E. State-resolved CO images
Images corresponding to three different representative J
levels of the CO ground vibrational level are shown in Fig. 6,
and the energy distributions extracted from the images are
shown in Fig. 7. These images were obtained by scanning
across the corresponding rotational lines of Fig. 5 at the in-
dicated wavelength. As can be clearly seen in both figures,
the energy distributions change dramatically with J: for low J
͑Jϭ22͒, the translational energy release is relatively low,
FIG. 6. Photofragment images for indicated rotational levels of CO vϭ0.
FIG. 5. ͑a͒ REMPI scan of CO vϭ0 on the B←←X transition obtained as
described in text ͑ϩ͒, along with simulation ͑solid line͒. ͑b͒ Rotational
population obtained from the simulation above.
FIG. 7. Translational energy distributions for CO ͑vϭ0,J͒ obtained from the
preceding 3 images. Solid line, Jϭ22; dot-dash, Jϭ30; dashed, Jϭ55.
J. Chem. Phys., Vol. 106, No. 18, 8 May 1997
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45

7622
Ahmed et al.: Photodissociation of oxalyl chloride
and probe lasers are parallel, the translational energy distri-
butions are not affected even if unknown vector correlations
are present, so long as the angular and translational energy
distributions are uncoupled. This is because cylindrical sym-
metry is preserved, so the inverse Abel transformation used
to reconstruct the distribution remains valid. It is not
straightforward to investigate these phenomena in one-laser
experiments, since the polarization of the probe and disso-
ciation laser cannot be varied independently. However, Sato
et al. have shown that the ͑B-X͒ Q-branch transition in CO is
insensitive to vector correlations.25 v-J correlations may in-
deed be present in the CO product, but the present experi-
ment is not sensitive to them.
Alignment of the chlorine atom atomic orbital is pos-
sible, however,36,35 and could impact the observed angular
distributions. The strongest evidence that unknown orbital
alignment is not dramatically influencing the observed distri-
butions is that the same angular distributions are seen on
numerous different transitions in this vicinity. Given atomic
orbital alignment in the product, differing transitions would,
in general, show clearly different behavior.
IV. DISCUSSION
These distributions suggest a simple picture of the dis-
sociation dynamics. Because we find Cl and CO fragments
that are near momentum-matching, it seems clear that they
are dynamically linked. However, the only way to obtain free
Cl and CO in a single step is to break both the C-C and the
Cl-CO bonds simultaneously. Furthermore, at the same time
we must end up with another population of Cl and CO that
possesses little translational energy, since there is a similar
momentum-matching pair that shows little recoil speed and
an isotropic angular distribution. For this to occur, the C–C
bond scission must take place with little concomitant mo-
mentum transfer to the remaining chloroformyl radical, yet
with transfer of sufficient internal energy to break the Cl–CO
bond in the radical. This process is schematically illustrated
in Fig. 9. The initial step breaking the C–C bond and one of
the Cl–CO bonds takes place with an impulse more-or-less
along the Cl–CO bond axis. The C–C bond must break si-
multaneously to preclude coupling of a substantial portion of
this impulse to the remaining fragments. The initial impulse
yields fast, rotationally excited CO and Cl that is predomi-
nantly (72%) in the spin-orbit excited state. However, a por-
tion of the fast Cl is formed in the ground state.
This simple model has several implications that can
readily provide a rough check on its validity. The first point
is that since we have state-resolved translational energy dis-
tributions for representative examples of all fragments, we
can roughly verify the energy balance. Although we have not
obtained full translational energy distributions for all CO ro-
tational levels, we do have the rotational distribution for
vϭ0, clearly the principal vibrational level. Furthermore, if
we use the P(ET) for the dominant rotational levels to rep-
resent the total distribution we will not be led too far astray,
since they contain contributions from the fast and slow com-
ponents that dominate the higher and lower J levels, respec-
FIG. 8. Center-of-mass angular distributions for the indicated rotational
levels of CO vϭ0, obtained from the images in Fig. 6. For Jϭ30, the two
curves represent those CO molecules recoiling less than 1000 m/s ͑‘‘slow’’͒
or faster than 1000 m/s ͑‘‘fast’’͒.
peaking at about 8 and extending beyond 60 kJ/mol. At in-
termediate values, the translational energy distributions are
clearly bimodal, further suggesting a relation to the chlorine
atom distributions. At high J, the fast component is domi-
nant, with the translational energy release peaking at about
40 kJ/mol. The nature of the distributions appeared to change
fairly smoothly as a function of J, with no sudden variations
in the relative amounts of the slow and fast components.
Angular distributions for these three rotational levels are
shown in Fig. 8. A single beta value is extracted from the
Jϭ22 (␤ϭ0.04͒ and Jϭ55 (␤ϭ0.45͒ data, although the
Jϭ22 may include some contribution from a distinct fast
component. The Jϭ30 result clearly consists of two compo-
nents, and distinct beta values are obtained for the slow and
fast portions of the distribution, somewhat arbitrarily set at
34 kJ/mol. Just as in the Cl case, the slow component is
nearly isotropic (␤ϭϪ0.03͒ while the fast component shows
significant anisotropy (␤ϭ0.41͒.
F. Vector correlations
Photofragment imaging can be a sensitive probe of vec-
tor correlations in the dissociation process, both for correla-
tion between the recoil direction and the rotational angular
momentum,33,34 and between the recoil direction and atomic
orbital alignment.35 Such vector correlations can impact the
inferred angular distributions if they are present yet unac-
counted for. It should be emphasized that if the dissociation
J. Chem. Phys., Vol. 106, No. 18, 8 May 1997
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45

Ahmed et al.: Photodissociation of oxalyl chloride
7623
state to yield the slower product. Although the arguments are
somewhat indirect, we find this picture unsatisfying for sev-
eral reasons. The most important reason is that for both the
slow and the fast atomic chlorine product, we find near-
momentum-matching CO. As noted above, this requires si-
multaneous breaking of both the Cl–C and C–C bonds in the
parent molecule. This leads directly to the picture we have
outlined, unless it is argued that some events give exclu-
sively fast products, while others give exlusively slow prod-
ucts. But any model that would give four fast fragments or
four slow fragments, even for a portion of the dissociation
events, would not conserve energy.
The nature of the electronic transition is not known here,
and it may be related to the somewhat puzzling case of acetyl
halides.37,38,7,39 All of the acetyl halides show a parallel dis-
sociation, with ␤ values for the Cl product similar to those
reported here. This is surprising since it is generally believed
*
that the transition is an n→␲ transition, with the transition
dipole expected to be perpendicular to the molecular plane. It
has been argued that the upper state for acetyl chloride may
be pyramidal, so that the out-of-plane transition moment
might still yield a positive ␤ owing to geometry changes in
the decaying molecule. The more commonly invoked expla-
FIG. 9. Cartoon representation of the dissociation model ͑see Discussion͒.
*
nation for these positive ␤ values is that the n→␲ transi-
tively. We can thus obtain the average energy in translation
for all fragments. These are summarized in Table III. We
obtain the average energy in rotation from the rotational dis-
tribution in Fig. 5. We neglect what we suspect are minor
contributions from vibrationally excited CO. A small contri-
bution from the electronic excitation is also included. In ad-
dition, the C–C bond energy and twice the Cl–CO bond in
chloroformyl radical are also included. The total is compared
to the photon energy at 235 nm. The rough agreement indi-
cates that the mechanism is at least reasonable, and none of
the products likely arise from processes involving another
photon. The fact that our estimated energy release is a bit
larger ͑531 vs. 509 kJ/mol͒ could be accounted for if some of
the ClCO remains bound. Attempts to find the radical
through ionization in this wavelength region were unsuccess-
ful, but future studies will seek to quantify it as a primary
product using alternative detection methods.
tion may be coupled to an n→␴* transition that is repulsive
along the Cl–C axis. We are more comfortable with the latter
picture, particularly since this behavior seems so widespread,
and is relatively insensitive to the identity of the molecule
and the dissociation wavelength.
V. CONCLUSION
We have studied the detailed photodissociation dynam-
ics of oxalyl chloride near 235 nm using the photofragment
imaging technique. Observed products include both ground
state Cl (2P3/2) and spin-orbit excited Cl ( P1/2) chlorine
2
*
atoms and ground electronic state CO molecules. Quantum
state resolved translational energy and angular distributions
were obtained for representative examples of all fragments.
These distributions suggest a novel dissociation mechanism
in which the first step is an impulsive three-body dissociation
An alternate model to explain the bimodal distributions
would involve invoking two distinct electronic states, or at
least two different dissociation paths, with one showing
prompt dissociation dynamics and the other perhaps follow-
ing internal conversion and decomposition on the ground
*
yielding predominantly Cl , rotationally excited CO and
chloroformyl radical ClCO, with only modest momentum
transfer to the latter species. Most, if not all, of the remaining
chloroformyl radical subsequently dissociates, yielding low
rotational levels of CO and little translational energy release.
TABLE III. Energy balance ͑kJ/mol͒. Photon energy at 235 nm is 509
kJ/mol.
ACKNOWLEDGMENTS
C–C
Cl–CO
322
50
2ϫ25
The authors are grateful to J. Zhang, M. Dulligan and C.
Wittig for providing the revised value of the HCl spin-orbit
branching ratio prior to publication, and thank C. Knopf for
technical assistance. This work was supported by the Direc-
tor, Office of Energy Research, Office of Basic Energy Sci-
ences, Chemical Sciences Division of the U. S. Department
of Energy under contract No. DE-ACO3-76SF00098.
Etrans
Cl
CO
Erot
Eel
:
2ϫ33.8
2ϫ30
2ϫ14.2
2ϫ3.5
67.6
60
28.4
7
Total
535
J. Chem. Phys., Vol. 106, No. 18, 8 May 1997
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45

7624
Ahmed et al.: Photodissociation of oxalyl chloride
1 X. Zhao, W. B. Miller, E. J. Hintsa, and Y. T. Lee, J. Chem. Phys. 90,
5527 ͑1989͒.
21 D. A. Blunt and A. G. Suits ͑unpublished͒.
22 D. Proch and T. Trickl, Rev. Sci. Instrum. 60, 713 ͑1989͒.
23 S. Arepalli, N. Presser, D. Robie, and R. J. Gordon, Chem. Phys. Lett.
118, 88 ͑1985͒.
2 J. Hepburn, R. Buss, L. Butler, and T. Lee, J. Phys. Chem. 87, 3638
͑1983͒.
3 I. Burak, J. W. Hepburn, N. Sivakumar, G. E. Hall, G. Hall, and P. L.
Houston, J. Chem. Phys. 86, 1258 ͑1987͒.
24 G. Loge, J. Tiee, and F. Wampler, J. Chem. Phys. 79, 196 ͑1983͒.
25 Y. Sato, Y. Matsumi, M. Kawasaki, K. Tsukiyama, and R. Bersohn, J.
Phys. Chem. 99, 16307 ͑1995͒.
26 J. A. H. M. A. Thelen, T. Gejo, and J. R. Huber, J. Chem. Phys. 103, 7946
͑1995͒.
27 R. A. Zare and D. R. Herschbach, Proc. IEEE 51, 173 ͑1963͒.
28 Y. Matsumi, P. K. Das, and M. Kawasaki, J. Chem. Phys. 97, 5261
͑1992͒.
4 G. Scuseria and H. S. III, J. Am. Chem. Soc. 111, 7761 ͑1989͒.
5 K. A. Trentelman, S. H. Kable, D. B. Moss, and P. L. Houston, J. Chem.
Phys. 91, 7498 ͑1989͒.
6 G. E. Hall, D. V. Bout, and T. J. Sears, J. Chem. Phys. 94, 4182 ͑1991͒.
7 S. W. North, D. A. Blank, J. D. Geseltzer, C. A. Longfellow, and Y. T.
Lee, J. Chem. Phys. 102, 4447 ͑1995͒.
8 L. Walker and H. Prophet, J. Trans. Faraday Soc. 63, 879 ͑1967͒.
9 D. W. Chandler and P. L. Houston, J. Chem. Phys. 87, 1445 ͑1987͒.
10 B. J. Whitaker, in Research in Chemical Kinetics Vol. 1, edited by R. G.
Compton and G. Hancock ͑Elsevier, Amsterdam, 1993͒.
11 B. Saksena and R. Kagarise, J. Chem. Phys. 19, 99 ͑1951͒.
12 D. Hassett, K. Hedberg, and C. Marsden, J. Phys. Chem. 97, 4670 ͑1993͒.
13 D. Danielson, L. Hedberg, K. Hedberg, K. Hagen, and M. Trae, J. Phys.
Chem. 99, 9374 ͑1995͒.
29 J. Cao, H.-P. Loock, and C. X. W. Qian, Can. J. Chem. 72, 758 ͑1993͒.
30 R. Liyanage, Y. Yang, S. Hashimoto, R. Gordon, and R. W. Field, J.
Chem. Phys. 103, 6811 ͑1995͒.
31 V. Skorokhodov, Y. Sato, K. Suto, Y. Matsumi, and M. Kawasaki, J.
Phys. Chem. 100, 12321 ͑1996͒.
32 J. Zhang, M. Dulligan, and C. Wittig ͑unpublished͒.
33 J. W. Thoman, D. W. Chandler, D. H. Parker, and M. H. M. Janssen,
Laser Chem. 9, 27 ͑1988͒.
14 J. Durig, J. Davis, and A. Wang, J. Mol. Struct. 375, 67 ͑1996͒.
15 J. Pola, Collec. Czech. Chem. Comm. 47, 3258 ͑1982͒.
16 W. Schroeder, M. Monnier, G. Davidovics, A. Allouche, P. Verlaque, J.
Pourcin, and H. Bodot, J. Mol. Struct. 197, 227 ͑1989͒.
17 C. Maul, T. Has, K.-H. Gericke, and F. Comes, J. Chem. Phys. 102, 3238
͑1995͒.
34 A. G. Suits, R. L. Miller, L. S. Bontuyan, and P. L. Houston, J. Chem.
Soc. Faraday Trans. 89, 1443 ͑1993͒.
35 Y. Mo, H. Katayanagi, M. C. Heaven, and T. Suzuki, Phys. Rev. Lett. 77,
830 ͑1996͒.
18 P. A. Heimann, M. Kolke, C. W. Hsu, M. Evans, C. Y. Ng, X. M. Yang,
C. Flaim, A. G. Suits, and Y. T. Lee, SPIE Proc. 2856, 90 ͑1996͒.
19 X. Yang, D. A. Blank, J. Lin, P. A. Heimann, A. M. Wodke, Y. T. Lee,
and A. G. Suits, in Synchrotron Radiation Techniques in Industrial,
Chemical, and Materials Science, edited by D’Amico et al. ͑Plenum, New
York, 1996͒.
36 Y. Wang, H.-P. Loock, J. Cao, and C. Qian, J. Chem. Phys. 102, 808
͑1995͒.
37 M. D. Person, P. W. Kash, and L. J. Butler, J. Phys. Chem. 96, 2021
͑1992͒.
38 M. D. Person, P. W. Kash, and L. J. Butler, J. Chem. Phys. 97, 355
͑1992͒.
20 L. S. Bontuyan, A. G. Suits, P. L. Houston, and B. J. Whitaker, J. Phys.
Chem. 97, 6342 ͑1993͒.
39 I. C. Lane, R. Meehan, and I. Powis, J. Phys. Chem. 99, 12371 ͑1995͒.
J. Chem. Phys., Vol. 106, No. 18, 8 May 1997
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:
131.111.164.128 On: Mon, 22 Dec 2014 06:45:45