Photochemistry of phosgene in the solid phase: Dissociation, ejection, and thermal desorption

Article abstract of DOI:10.1063/1.471505

Understanding photochemistry and energy transfer mechanisms in molecular solid films is of interest to many scientific issues, ranging from matrix-assisted laser desorption ionization mass spectrometry to photochemical processes on polar stratospheric cloud particles. We present a study of a model system: the photochemistry (hv= 1.2-6.4 eV) of a molecular Cl₂CO solid film at low laser power density, 10 μJ-1 mJ/cm² for ～10 ns pulses. At hv≥3.5 eV, photon absorption by Cl₂CO leads to a major photodissociation channel resulting in CO (g) and Cl (g) and a minor molecular Cl₂CO ejection channel. Both photodissociation and molecular ejection are observed at the lowest laser power density and their yields depend linearly on pulse energy. This result establishes a single photon photoexcitation mechanism. The electronically excited Cl₂CO in the surface region of the solid film can either dissociate or convert its electronic energy to translational motion in Cl₂CO. The translational energy distribution of CO (g) from the photodissociation channel is bimodal: the flux-weighted mean translational energy of the fast channel is photon energy dependent (〈E_trans〉=210, 135, and ～90 meV at hv=6.4, 5.0, and 3.5 eV, respectively), while the slow channel is independent of photon energy and corresponds to completely thermalized CO molecules (〈E_trans/2k〉=84±3 K). The mean translational energy of photoejected Cl₂CO is 〈E_trans〉=220±20 meV. In addition to photoejection, there is also a distinctively different thermal desorption channel due to transient laser heating.

Full text of DOI:10.1063/1.471505

Photochemistry of phosgene in the solid phase: Dissociation, ejection, and thermal
desorption
Q.S. Xin and X.Y. Zhu
Citation: The Journal of Chemical Physics 104, 7895 (1996); doi: 10.1063/1.471505
View online: http://dx.doi.org/10.1063/1.471505
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/104/20?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
The importance of the anisotropic energy term for the structure of the solid phases of nitrogen
J. Chem. Phys. 105, 3235 (1996); 10.1063/1.471839
Photochemistry of phosgene in the solid phase: Stateresolved dissociation dynamics
J. Chem. Phys. 104, 7904 (1996); 10.1063/1.471506
Interrogation and control of condensed phase chemical dynamics with linearly chirped pulses: I2 in solid Kr
J. Chem. Phys. 104, 6497 (1996); 10.1063/1.471370
The 3ω technique for measuring dynamic specific heat and thermal conductivity of a liquid or solid
Rev. Sci. Instrum. 67, 29 (1996); 10.1063/1.1146545
Orientational order in the solid hydrogens: The boomerang phase line in HD
AIP Conf. Proc. 309, 841 (1994); 10.1063/1.46230
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

Photochemistry of phosgene in the solid phase: Dissociation, ejection,
and thermal desorption
Q.-S. Xin and X.-Y. Zhua)
Department of Chemistry & Biochemistry, Southern Illinois University, Carbondale, Illinois 62901
͑Received 12 January 1996; accepted 22 February 1996͒
Understanding photochemistry and energy transfer mechanisms in molecular solid films is of
interest to many scientific issues, ranging from matrix-assisted laser desorption ionization mass
spectrometry to photochemical processes on polar stratospheric cloud particles. We present a study
of a model system: the photochemistry ͑h␯ϭ1.2–6.4 eV͒ of a molecular Cl2CO solid film at low
laser power density, 10 ␮J–1 mJ/cm2 for ϳ10 ns pulses. At h␯у3.5 eV, photon absorption by
Cl2CO leads to a major photodissociation channel resulting in CO (g) and Cl (g) and a minor
molecular Cl2CO ejection channel. Both photodissociation and molecular ejection are observed at
the lowest laser power density and their yields depend linearly on pulse energy. This result
establishes a single photon photoexcitation mechanism. The electronically excited Cl2CO in the
surface region of the solid film can either dissociate or convert its electronic energy to translational
motion in Cl2CO. The translational energy distribution of CO (g) from the photodissociation
channel is bimodal: the flux-weighted mean translational energy of the fast channel is photon energy
dependent ͑͗Etrans͘ϭ210, 135, and ϳ90 meV at h␯ϭ6.4, 5.0, and 3.5 eV, respectively͒, while the
slow channel is independent of photon energy and corresponds to completely thermalized CO
molecules ͑͗Etrans/2k͘ϭ84Ϯ3 K͒. The mean translational energy of photoejected Cl2CO is
͗Etrans͘ϭ220Ϯ20 meV. In addition to photoejection, there is also a distinctively different thermal
desorption channel due to transient laser heating. © 1996 American Institute of Physics.
͓S0021-9606͑96͒01920-7͔
I. INTRODUCTION
motion; transient heating induced by the short laser pulse can
lead to thermal desorption ͑vaporization͒; ͑iv͒ initial photo-
excitation in the substrate can be transferred to the adsorbed
film, leading to chemical effects, such as molecular desorp-
tion induced by acoustic waves generated at color centers in
ionic substrates or by transient heating of the surface region
of the substrate; ͑v͒ dissociation or desorption induced by the
attachment of photoexcited substrate carriers to condensed
molecules.
Mechanisms ͑iv͒ and ͑v͒ are attributed to excitation in
the substrate and are closely related to surface photochemis-
try of adsorbates in the monolayer region. A number of re-
cent reviews are available on this subject.8,9 We will not
discuss these mechanisms further in the present paper.
Mechanism ͑i͒ is common to photochemistry in other media,
such as gas phase, solution phase, or solid surfaces. Mecha-
nism ͑iii͒ is thermal in nature and is an inevitable conse-
quence at high laser power density, due to the poor thermal
conductivity of most molecular solids. Of particular interest
here is mechanism ͑ii͒, which is perhaps unique to photo-
chemistry in molecular solid films. Under various circum-
stances, this mechanism has been called ‘‘photodettach-
ment,’’ ‘‘photoejection,’’ and ‘‘photodesorption.’’ Nishi
et al. observed the ‘‘photodettachment’’ of energetic neutrals
from pure or mixed solids of H2O and NH3 and suggested
that the repulsion between an electronically excited molecule
and a ground state molecule could lead to translational mo-
tion and hence desorption.1 Polanyi and co-workers reported
‘‘photoejection’’ in a number of systems and suggested that
quenching of an electronically excited molecule by a partner
Photochemistry of molecular solid films is of interest in
several aspects. From a fundamental perspective, it provides
a rich test ground for understanding energy transfer mecha-
nisms in the condensed state and enables a valuable compari-
son to photochemical processes in the gas phase, in solu-
tions, and on solid surfaces. Laser interaction with molecular
thin films is employed in a number of important techniques
in analytical chemistry, such as laser ejection of biomol-
ecules for gas phase electronic spectroscopy or matrix as-
sisted laser desorption ionization ͑MALDI͒ mass spectrom-
etry for macromolecules ͑polymers, proteins, DNA, etc.͒.
Photochemistry in or on the surface of polar stratospheric
cloud particles, which are water containing molecular solids,
plays important roles in the chemical cycles for stratospheric
ozone depletion. In medical technology, laser removal of
biological tissues is enjoying increasing popularity.
There have been considerable efforts on the photochem-
istry of molecular thin films condensed on cryogenic
substrates.1–7 Our current understanding can be summarized
by the following mechanistic pictures: ͑i͒ electronic excita-
tion of molecules in the surface region of the solid film can
lead to dissociation or bimolecular reaction with products
ejected into the gas phase; ͑ii͒ electronic excitation energy in
the molecule can be converted to translational motion, result-
ing in the ejection ͑desorption͒ of the parent molecule; ͑iii͒
initial photon absorption in the film is converted into thermal
a͒
To whom correspondence should be addressed. Phone: ͑618͒ 453-3077;
Fax: ͑618͒ 453-6408; E-mail: zhu@qm.c-chem.siu.edu
J. Chem. Phys. 104 (20), 22 May 1996
0021-9606/96/104(20)/7895/9/$10.00
© 1996 American Institute of Physics
7895
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

7896
Q.-S. Xin and X.-Y. Zhu: Photochemistry of phosgene in solid phase
with concurrent E→TЈ,VЈ,RЈ energy transfer was
responsible.4 These mechanisms are related to a special case
of thermal desorption, termed ‘‘explosive’’ desorption by
Domen and Chuang.2 At high laser power densities,
E→TЈ,VЈ,RЈ energy transfer can lead to very intense local
heating in the molecular film, effectively resulting in an ‘‘ex-
plosion’’ to yield fast molecules. This process is also called
‘‘laser ablation’’ in other context and is usually accompanied
by supersonic expansion.5,6 Note that, in the majority of
studies, relatively high laser power densities ͑Ͼ0.1
MW/cm2͒ were employed. As a result, the E→TЈ mecha-
nism was usually accompanied by other molecular desorp-
tion channels, particularly collective processes, such as ther-
mal and/or explosive desorption.
FIG. 1. Schematics of the experimental setup.
In the present study, we investigate the photochemistry
of solid Cl2CO in the low laser power density region ͑1–100
kW/cm2͒ to firmly delineate and establish various photo-
chemical channels in the solid film. In particular, we estab-
lish the wavelength and power dependences of these chan-
nels. This is the second of three papers on the
photodissociation dynamics of phosgene in the condensed
phase. The first paper is focused on monolayer phosgene
adsorbed on Ag.10 The accompanied paper details the state-
resolved dynamics of the photodissociation channel for the
solid molecular film.11 A brief report of these results has
already been published elsewhere.12
The gas phase photochemistry of phosgene has been
studied in the past. Phosgene absorbs appreciably at ␭р305
nm. UV photodissociation of Cl2CO is believed to yield the
unstable ClCO radical, which promptly dissociates to CO
and Cl:13
PC-based multichannel scaler ͑MCS͒, which was synchro-
nized to the laser pulse. The flight time was corrected by the
QMS transit time ͑ionizer to channeltron͒, t0, which was
determined in separate experiments by directly irradiating
the QMS ionizer region with the excimer laser. The TOF
spectra of background species desorbing from materials in
the ionization region ͑metal grids, ceramic insulators, etc.͒
were recorded and the onset of each spectrum was taken as t0
for that particular mass. After t0 correction and baseline sub-
traction, each experimental TOF spectrum was fit with the
following equation:
2
a
t4
d
I t͒ϭ exp Ϫb
,
͑3͒
͑
ͫ
ͩ ͪ
ͬ
t
where I(t) is the signal intensity; a is a proportionality con-
stant; b(ϭm/2kT), relating to the mass (m) and temperature
(T) of the desorbed molecule, is a measure of the width of
the distribution; dϭ76 mm, is the distance form the sample
to the QMS ionizer. Equation ͑3͒ was originally derived to
describe the TOF distribution of a molecular beam from a
nozzle14 and has been successfully applied to photodesorp-
tion studies.3,9 Note that QMS is a density sensitive detector.
Transforming the signal to flux, a physically meaningful
quantity, requires velocity correction, i.e., dividing the signal
by flight time.
Cl2COϩh␯→ClCOϩCl,
ClCO→COϩCl.
͑1͒
͑2͒
The energy requirement for breaking the first Cl–CO bond,
reaction ͑1͒, is 3.3 eV and that for the second Cl–CO bond,
reaction ͑2͒, is only 0.26 eV.
II. EXPERIMENT
All experiments were performed in a two-level ultra-
high vacuum ͑UHV͒ chamber, with a base pressure of
1ϫ10Ϫ10 torr.10 The upper level was used mainly for surface
preparation and analysis. Most experiments were done on the
lower level ͑Fig. 1͒. A quadrupole mass spectrometer ͑QMS͒
was mounted vertically from the bottom of the chamber. It
was equipped with a cross beam ionizer and a channeltron
detector, and used for residual gas analysis ͑RGA͒, thermal
desorption spectroscopy ͑TDS͒, and time-of-flight ͑TOF͒
distribution measurements. An electron energy of 70 eV was
used in electron bombardment ionization. The TOFMS was
used for mass resolved ion detection in resonance-enhanced
multiphoton ionization ͑REMPI͒ spectroscopy, as detailed in
the companion paper.
In most experiments, phosgene was condensed on a
Ag͑110͒ sample ͑␾10 mm͒ mounted at the center of the
chamber. It was cooled by liquid nitrogen and heated resis-
tively. Cleaning of the surface was achieved by standard
procedures.10 In some experiments, a GaAs͑100͒ substrate
was also used. All results on condensed phosgene were
found to be independent of the substrates, including, clean
Ag͑110͒, Cl covered Ag͑110͒, and GaAs͑100͒. Phosgene
was purified by freeze–pump–thaw cycles and its purity
confirmed by RGA. Dosing was carried out through a posi-
tive shut-off valve ͑VTI͒ with calibrated leak rates, at a sub-
strate temperature (Ts) of ϳ90 K. This calibrated leak valve
was connected to a ␾9 mm tube terminated at ϳ1.5 cm from
the sample surface. All coverages are presented on a relative
scale as determined by TDS peak areas, with one monolayer
͑1 ML͒ corresponding to the saturation of the monolayer
peak in thermal desorption spectrum.10
The setup for TOF measurement is shown in Fig. 1. The
cross-beam ionizer of the QMS was at surface normal and 76
mm away from the sample. Ion signal from the channeltron
detector was amplified in a fast-preamp and recorded by a
J. Chem. Phys., Vol. 104, No. 20, 22 May 1996
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

Q.-S. Xin and X.-Y. Zhu: Photochemistry of phosgene in solid phase
7897
Three laser systems were employed in the experiments:
an excimer laser ͑LamdaPhysik Lextra-100͒, a Q-switched
Nd:YAG laser ͑Continuum Powerlite-7010͒, and a tunable
dye laser ͑Continuum ND6000͒. The excimer laser generated
193 nm ͑17 ns pulse width͒ and 248 nm ͑23 ns pulse width͒
light when filled with Ar–F2 and Kr–F2 mixtures, respec-
tively. It could be operated at repetition rates of 50 Hz or less
and was used as the main photolysis laser source. The fun-
damental ͑1064 nm͒, second harmonic ͑532 nm͒, and third
harmonic ͑355 nm͒ outputs of the YAG laser ͑6 ns pulse
width, 10 Hz͒, as well as the YAG pumped dye output at 650
nm, were used as photolysis source in some experiments.
The laser output was expanded to ensure uniformity on the
sample surface. The laser pulse energy varied from 0.01 to 2
mJ/cm2. All measurements were carried out at a substrate
temperature of ϳ90 K with the photolysis laser at an angle of
incidence of 60° with respect to the surface normal.
tribution of CO is independent of laser pulse energy ͑0.01–
1.0 mJ/cm2͒ and can be deconvoluted into two components,
as shown by the dashed curves, with 60% of the total CO
flux in the slow channel. The flux-weighted mean transla-
tional energies of the two channels are: ͗Etrans1͘ϭ135Ϯ5
meV or ͗Etrans1/2k͘ϭ790Ϯ30 K, ͗Etrans2͘ϭ14.5Ϯ0.5 meV or
͗Etrans2/2k͘ϭ84Ϯ3 K. The translational temperature of the
slow channel is nearly identical to the substrate temperature.
As detailed in the companion paper, internal state character-
ization of CO molecules also shows a completely thermal-
ized origin at the substrate temperature for this slow CO
channel. Note that, in the high laser pulse energy region
͑у0.2 mJ/cm2͒, significant amount of molecular Cl2CO de-
sorption is also observed ͑see below͒. The contribution of
Cl2CO QMS cracking signal to the CO spectra must be neg-
ligible, due to the low relative intensity of Cl2CO in this
pulse energy region ͑see Fig. 4͒. As a result, these TOF
distributions should be exclusively assigned to CO (g).
The situation is different for the Cl TOF spectra ͑m/e
ϭ35͒, middle panel in Fig. 2. At low pulse energies, only a
fast Cl peak was observed. With increasing pulse energy, a
slow peak sets in and becomes dominating at pulse energies
у0.4 mJ/cm2. Note the scaling factors for the upper two
spectra. The slow peak results from the cracking of molecu-
lar Cl2CO in the QMS ionizer, as evidenced by the observa-
tion of other masses ͑see Fig. 3͒. The fast channel should be
assigned to atomic Cl, since contributions from QMS crack-
ing signals of other Cl-containing species, including fast
Cl2CO and fast Cl2, should be negligibly small due to their
low relative intensities ͑see Fig. 4͒. The flux-weighted mean
translational energy of the fast channel is ͗Etrans͘ϭ280Ϯ20
meV.
The right panel in Fig. 2 shows a set of TOF spectra of
ClCOϩ, which is the major QMS signal for molecular
Cl2CO. Each spectrum is normalized to the photon exposure
and scaled by the indicated factor. Molecular Cl2CO desorp-
tion is bimodal, as evidenced by the spectra at low pulse
energies. The flux-weighted mean translational energy of the
fast channel is ͗Etrans͘ϭ210Ϯ20 meV and that of the slow
channel is ͗Etrans͘ϭ41Ϯ3 meV. The photon-normalized flux
of the fast channel is constant at all pulse energies. In other
words, the flux per laser pulse in the fast channel increases
linearly with pulse energy, establishing a single photon
mechanism. On the other hand, the intensity of the slow
channel increases exponentially with pulse energy. In fact,
this slow channel overwhelms the fast channel at pulse en-
ergies у0.2 mJ/cm2. Note that the fast channel is nearly
invisible in the upper two spectra, due to the use of small
scaling factors. In the following, we will call the fast mo-
lecular desorption channel ‘‘photoejection’’ and the slow
channel ‘‘thermal desorption,’’ as discussed in detail later.
Figure 3 compares the spectra of major QMS signals for
molecular Cl2CO ͑except for CO, at a pulse energy of 0.47
mJ/cm2. The QMS cracking pattern predicts that the COϩ
signal from Cl2CO should be ϳ30% of that for ClCOϩ. Due
to the dominance of CO (g) from the photodissociation
channel, the actual COϩ signal is seven times higher than
that of ClCOϩ. As shown by the fitting curves, the slow TOF
III. RESULTS
The experimental results are organized into two sections.
In Sec. III A, we present results at 248 nm, including TOF
distributions and laser power dependences. Section III B ex-
amines the dependence of photochemistry on laser wave-
length, ranging from deep UV to IR ͑␭ϭ193, 248, 355, 532,
650, and 1064 nm͒. No photochemical effect was observed at
␭ϭ650 and 1064 nm. Note that, at the low laser pulse ener-
gies employed here ͑10 ␮J–1 mJ/cm2͒, the fluxes of photo-
chemical products are much lower than that necessary for
significant post-desorption collision in the gas phase. There-
fore, the TOF distributions reported here should not be dis-
torted upon leaving the surface of the molecular film.
In most experiments, we used a 50 ML thick phosgene
film at 90 K, which was below the melting temperature for
Cl2CO, Tmϭ145.3 K. Since deposition was also done at 90
K, we assume the molecular film was in the form of amor-
phous solid. The effects of film thickness and surface prepa-
ration were also investigated. We found that the observed
photochemistry was independent of thickness ͑10–200 ML͒
or surface preparation ͓clean Ag͑110͒, Cl covered Ag͑110͒,
or GaAs͑100͔͒. This indicates that all photochemical results
presented below were due to photon absorption by the mo-
lecular film. Photo-excited substrate electrons play no role in
the film thickness region investigated. In the following, we
only present results for 50 ML Cl2CO on Ag͑110͒. No effort
was made to quantitatively account for effects due to the
thickness-dependent optical interference in the molecular
film.
A. 248 nm
The photochemistry at 248 nm was investigated in a
broad laser pulse energy region, from 0.01 to 1.0 mJ/cm2.
The observed gas phase products include CO, Cl, Cl2, and
Cl2CO. These species lead to the following QMS signals:
COϩ, Clϩ, COClϩ, Cl2ϩ, and Cl2COϩ.
The left panel in Fig. 2 shows a set of TOF spectra of
CO at the indicated laser pulse energies. Each spectrum has
been normalized to the total photon exposure. The TOF dis-
J. Chem. Phys., Vol. 104, No. 20, 22 May 1996
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

7898
Q.-S. Xin and X.-Y. Zhu: Photochemistry of phosgene in solid phase
FIG. 2. TOF spectra of m/eϭ28 ͑left͒, 35 ͑middle͒, and 63 ͑right͒ from 50 ML Cl2CO irradiated by 248 nm light at the indicated laser pulse energies. All
spectra were normalized to the number of incident photons. Note the scaling factors for some of the spectra. The dashed curves are deconvolutions based on
modified MB distributions ͑see text for details͒. The solid curve on the bottom spectrum in the right panel is the sum of dashed curves.
FIG. 3. TOF spectra of, from top to bottom, m/eϭ35, 63, 70, and 98,
obtained from the irradiation of 50 ML Cl2CO at 248 nm and a laser pulse
energy of 0.47 mJ/cm2. Each spectrum was normalized to the number of
incident photons and scaled by the indicated factor. The dashed curves are
fits based on modified MB distributions ͑see text for details͒. The solid
curves are the sums of dashed curves.
FIG. 4. The dependences of TOF fluxes as a function of laser pulse energy
at 248 nm. The data points are: COϩ ͑circles͒, Clϩ fast channel ͑crosses͒,
ClCOϩ fast channel ͑triangles͒, Clϩ slow channel ͑squares͒, ClCOϩ slow
channel ͑diamonds͒. The solid lines in the left panel are least square linear
fits to the data and those in the right channel are smoothed version of the
data.
J. Chem. Phys., Vol. 104, No. 20, 22 May 1996
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

Q.-S. Xin and X.-Y. Zhu: Photochemistry of phosgene in solid phase
7899
channel of Cl is identical to the TOF spectra at m/eϭ63 and
98 ͑ClCOϩ and Cl2COϩ͒, as well as the slow channel at
m/eϭ70 ͑Clϩ2 ͒. The relative intensities of these spectra are
also in excellent agreement with the QMS cracking pattern
determined in separate RGA measurements. The flux
weighted mean translational energy for this slow molecular
desorption channel is ͗Etrans͘ϭ41Ϯ3 meV or ͗Etrans/2k͘
ϭ230Ϯ20 K.
The above results establish the following photochemical
processes: a photodissociation channel yielding CO (g) and
Cl (g); a photoejection channel resulting in fast Cl2CO (g);
and a slow molecular desorption channel at high pulse ener-
gies. In addition, there is a minor photoreaction product, Cl2,
which is evidenced by the fast TOF channel in the Cl2 spec-
trum ͑Fig. 3͒. The flux-weighted mean translational energy
of this Cl2 product is ͗Etrans͘ϭ360Ϯ60 meV. This Cl2 prod-
uct probably comes from the recombination of atomic Cl
resulting from photodissociation. Note that the contribution
of the QMS cracking signal of Cl2 to the Cl TOF spectrum in
Fig. 3 must be negligible, due to the significant difference in
the relative intensities.
The dependence of TOF fluxes on laser pulse energy is
shown in Fig. 4. The fluxes of photodissociation products,
CO and Cl ͑fast TOF channel͒, as well as photoejected
Cl2CO, increase linearly with laser pulse energy. The linear
power dependence establishes a single photon excitation
mechanism. The slow molecular desorption channel has, on
the other hand, a supralinear or nearly exponential depen-
dence on laser pulse energy, as shown by the major QMS
signals for molecular Cl2CO in the right panel. This channel
can be assigned to thermal desorption due to transient laser
heating.
FIG. 5. TOF spectra of m/eϭ28 ͑top͒, 35 ͑middle͒, and 63 ͑bottom͒ from
the irradiation of 50 ML Cl2CO at 193 nm and a pulse energy of 0.02
mJ/cm2. Each spectrum was normalized to the number of incident photons
and scaled by the indicated factor. The top spectrum was fit with two chan-
nels while the other two spectra with a single MB distribution. The inset
shows the laser pulse energy dependence for the COϩ flux.
weighted mean translational energy of ͗Etrans͘ϭ220Ϯ10
meV or ͗Etrans/2k͘ϭ1280Ϯ60 K. Within experimental uncer-
tainty, this energy is the same as that of the fast channel at
248 nm, pointing to the same origin at the two wavelengths.
The slow molecular desorption channel sets in at higher
pulse energies ͑not shown͒.
B. Other wavelengths
The photochemical pathways at 193 nm are qualitatively
similar to those at 248 nm: a single photon photodissociation
channel to yield CO (g) and Cl (g) and molecular desorp-
tion channels which give fast and slow Cl2CO (g). Figure 5
shows the TOF spectra of major QMS signals at a laser pulse
energy of 0.022 mJ/cm2, which is below the threshold for the
slow molecular desorption channel. The TOF distribution of
CO is again bimodal and independent of laser pulse energy
͑0.01–0.35 mJ/cm2͒, while the total flux increases linearly
with pulse energy ͑inset͒. This result establishes a single
photon excitation mechanism for the photodissociation chan-
nel, same as that at 248 nm. Each CO spectrum can be de-
convoluted into two components, as shown by the dashed
curves on the topmost spectrum, with 60% of the total CO
flux in the slow channel. The flux-weighted mean transla-
tional energies of the two channels are: ͗Etrans1͘ϭ210Ϯ20
meV or ͗Etrans1/2k͘ϭ1200Ϯ100 K, which is higher than that
at 248 nm; ͗Etrans2͘ϭ14.5Ϯ0.5 meV or ͗Etrans2/2k͘ϭ84Ϯ3
K, corresponding to thermalized CO.
The middle spectrum in Fig. 5 shows a Clϩ TOF distri-
bution. Due to the overlap of Cl from the photodissociation
channel and cracking of Cl2CO from the fast molecular chan-
nel, we are not able to unambiguously determine the trans-
lational energy distribution of Cl. The solid curve on the
spectrum is an approximate fit to the data. Details are pre-
sented below.
Figure 6 compares the TOF distributions of photodisso-
ciation products at two photon energies: 6.4 eV ͑193 nm͒ and
5.0 eV ͑248 nm͒. All spectra have been converted to the flux
domain, which is more physically meaningful than the den-
sity domain in original data. For CO, while the slow TOF
channel is independent of h␯ ͑͗Etrans͘ϭ14.5 meV͒, the fast
channel is positively related to h␯: ͗Etrans͘ϭ210 and 135
meV at h␯ϭ6.4 and 5.0 eV, respectively. Note that fitting to
the modified MB distribution at 193 nm is not satisfactory;
the actual TOF distribution in the fast channel is somewhat
broader than the MB distribution.
For Cl (g) at 248 nm, lower spectrum in the right panel
of Fig. 6, the TOF distribution is broader than a single MB
distribution. It can be deconvoluted, with equal population,
into a fast and a slow component ͑dashed curve͒. The flux-
weighted mean translational energies are ͗Etrans͘ϭ450 and
The intensity of the fast molecular Cl2CO desorption
channel at 193 nm ͑bottom spectrum͒ is much higher than
that at 248 nm. The TOF distribution can be well described
by a single MB distribution ͑solid curve͒ with a flux-
J. Chem. Phys., Vol. 104, No. 20, 22 May 1996
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

7900
Q.-S. Xin and X.-Y. Zhu: Photochemistry of phosgene in solid phase
FIG. 7. TOF spectra of m/eϭ28 ͑top͒, 35 ͑middle͒, and 63 ͑bottom͒ from
the irradiation of 50 ML Cl2CO at 355 nm and a laser pulse energy of 0.18
mJ/cm2. The top spectrum was fit with two channels while the other two
spectra with single MB distribution.
FIG. 6. A comparison of the TOF distributions of CO ͑left͒ and Cl ͑right͒ at
two laser wavelengths: 193 and 248 nm. The pulse energy for the 193 nm
spectra was 0.02 mJ/cm2 and that for 248 nm was 0.2 mJ/cm2. Each spec-
trum was normalized to the number of incident photons and scaled by the
indicated factor. Note the contributions of molecular Cl2CO to the Cl spec-
tra. See text for details on the fitting curves.
110 meV, respectively, with an overall energy of ͗Etrans
͘
ϭ280 meV. While this deconvolution can only be taken as
speculation, it is consistent with the two step reaction
mechanism presented in the introduction section and dis-
cussed later. Note that, at this wavelength and pulse energy,
contribution of molecular Cl2CO cracking signal to the Cl
TOF spectrum is negligibly small ͑see the dashed curve for
Cl2CO͒. At 193 nm, interference from the overlapping
Cl2CO ejection channel prohibits a reliable assessment of the
translational distribution of Cl (g). The dashed curves on the
upper right spectrum in Fig. 6 shows a possible deconvolu-
tion which yields an overall flux-weighted mean translational
energy of ͗Etrans͘ϭϳ350 meV, higher than that at 248 nm.
Figure 7 shows a set of TOF spectra at 355 nm, with a
pulse energy of 0.18 mJ/cm2. This wavelength is near the
threshold for photodissociation, since the slow molecular
Cl2CO desorption channel ͑thermal desorption͒ now domi-
nates the spectra. The photodissociation process is shown by
the fast CO TOF channel, with a flux-weighted mean trans-
lational energy of ͗Etrans͘ϭ90Ϯ20 meV or ͗Etrans/2k͘ϭ500
Ϯ100 K, lower than those at 193 and 248 nm. The relative
intensity of the slow CO channel is higher than the corre-
sponding value at shorter wavelengths, due to the contribu-
tion from Cl2CO thermal desorption, as shown by the Clϩ
and ClCOϩ spectra. The flux-weighted mean translational
energy for Cl2CO desorption is ͗Etrans͘ϭ43Ϯ3 meV.
FIG. 8. Left panel: TOF spectra of, from bottom to top, m/eϭ28, 35, 63, 70,
and 98 obtained from the irradiation of 50 ML Cl2CO at 532 nm and a laser
pulse energy of 0.90 mJ/cm2. Note the scaling factors for some spectra. Each
spectrum was fit with a single modified MB distribution ͑solid curve͒. The
right panel compares the TOF distributions of m/eϭ63 at different laser
pulse energies. The inset shows the dependence of flux-weighted mean
translational energy ͑obtained from the fitting curves͒ as a function of laser
pulse energy.
At 532 nm, only thermal desorption of slow Cl2CO was
J. Chem. Phys., Vol. 104, No. 20, 22 May 1996
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

Q.-S. Xin and X.-Y. Zhu: Photochemistry of phosgene in solid phase
7901
TABLE I. Flux-weighted mean translational energies ͑meV͒. N/Aϭnot available; N/Rϭnot resolved; N/Oϭnot observed.
Photodissociation
␭
͑nm͒
Photoejection
Cl2CO
Photoreaction
Cl2
Thermal des.
h␯
͑eV͒
CO
Cl
Cl2CO
193
248
355
532
6.4
5.0
3.5
2.3
210Ϯ20; 14.5Ϯ0.5
135Ϯ5; 14.5Ϯ0.5
ϳ90; N/R
ϳ350
280Ϯ20 ͑450, 110͒
N/R
220Ϯ10
210Ϯ20
N/R
N/A
360Ϯ60
N/R
N/A
41b
43
a
N/O
N/O
N/O
N/O
42–62
aNumbers in parenthesis correspond to possible deconvolution.
bDepending on laser pulse energy; see text for detail.
observed, as shown by TOF spectra in the left panel of Fig.
8. Within experimental uncertainty, the TOF distributions of
all signals are identical. The relative intensities are also in
excellent agreement with the QMS cracking pattern of
Cl2CO. The right panel compares the TOF distribution of
ClCOϩ at different laser pulse energies. There is an observ-
able shift to shorter TOF at higher laser pulse energy. The
inset shows the pulse energy dependence of ͗Etrans͘, which
increases from 42 meV at 0.18 mJ/cm2 to 62 meV at 1.8
mJ/cm2. This observation is consistent with a thermal mecha-
nism due to transient laser heating. At longer laser wave-
lengths ͑650 and 1064 nm͒, neither dissociation nor desorp-
tion was observed.
energetic information from the photodissociation reaction, as
evidenced by the positive correlation between ͗Etrans͘ and h␯
͑͗Etrans͘ϭϳ90, 135, and 210 meV at h␯ϭ3.5, 5.0, and 6.4
eV, respectively͒. We believe these fast CO molecules are
from photodissociation in the topmost surface layer of the
molecular film. They desorb promptly into the gas phase,
without suffering significant energy loss to the environment.
The same can be said about Cl whose translational energy is
also positively related to h␯ ͑͗Etrans͘ϭ280 and ϳ350 meV at
h␯ϭ5.0 and 6.4 eV, respectively͒.
The gas phase photodissociation mechanism predicts
two different Cl atoms, from reactions ͑1͒ and ͑2͒, respec-
tively. However, the limited resolution of the TOF spectrum
does not allow us to unambiguously resolve the two chan-
nels. As shown in the right panel of Fig. 6, the TOF distri-
bution of Cl at 248 nm is broader than a single modified MB
distribution. It can be deconvoluted into two components,
with ͗Etrans͘ϭ110 and 450 meV, respectively. While this de-
convolution should in no way be interpreted as evidence for
the existence of two channels, it is not unreasonable since
͗Etrans͘ of the slow component ͑110 meV͒ is in agreement
with expectation from ͗Etrans͘ of CO ͑135 meV͒ based on
momentum conservation in reaction ͑2͒. Accepting this as-
sumption, the fast Cl component ͑͗Etrans͘ϭ450 meV͒ can be
attributed to reaction ͑1͒.
Table I summarizes the flux-weighted mean translational
energies of all gas products from the laser irradiation of the
solid Cl2CO film at different laser photon energies.
IV. DISCUSSION
A. Photodissociation
The major chemical channel for the solid Cl2CO film
under UV laser irradiation is photodissociation to produce
CO (g) and Cl (g). The power dependence in Fig. 4 firmly
establishes a single photon electronic excitation mechanism
for this channel. Much about the dynamics of this process
can be learned from the product TOF distributions. The TOF
distribution of CO (g) is bimodal. As summarized in Table I,
the ͗Etrans͘ of the slow channel is independent of photon
energy and is completely thermalized to the substrate tem-
perature. This conclusion is further supported by the obser-
vation that the rotational state distribution of slow CO is also
completely thermalized to the substrate temperature.11 Evi-
dently, these CO molecules have lost all nascent energetic
information from the photodissociation reaction. Since irra-
diation of the solid film should lead to electronic excitation
of not only the topmost layer but also molecules beneath the
surface, some of these electronic excitation events could re-
sult in photodissociation within the solid film. The product
CO is not condensable at 90 K. On their way out, these CO
molecules can achieve efficient thermal equilibration with
the surrounding and give rise to the slow TOF peak. The
other photodissociation product, Cl, if formed below the sur-
face, can be easily trapped within the film. Post-irradiation
analysis reveals the accumulation of Cl on the substrate.
The fast CO channel is very energetic and not equili-
brated with the substrate. These CO molecules retain nascent
In addition to the major photodissociation channel, we
also observed
a
minor photoreaction product, Cl2
͑͗Etrans͘ϭ360 meV͒. One possible source for Cl2 (g) is a
secondary reaction:
ClϩCl2CO→ClCOϩCl2.
͑4͒
However, in the gas phase, this reaction is endothermic by
20 kcal molϪ1 or 0.87 eV.13 We believe reaction ͑4͒ is unlike
responsible for the observed energetic Cl2 (g), whose ͗Etrans
͘
is even higher than that of Cl (g). A more likely source for
fast Cl2 (g) is the recombination of two Cl atoms.
ClϩCl→Cl2.
͑5͒
During the photodissociation of Cl2CO on the surface of the
molecular film, the repulsive interaction between photodisso-
ciation products and neighboring Cl2CO molecules ͑in the
first and second layer͒ can force the recombination of two Cl
atoms. Similar processes are well known in the solution
phase, e.g., the photodissociation of I2.15 The solvent mol-
ecules could trap the initial atom pair, the geminate pair, and
force them to recombine, the so-called ‘‘cage effect.’’
J. Chem. Phys., Vol. 104, No. 20, 22 May 1996
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

7902
Q.-S. Xin and X.-Y. Zhu: Photochemistry of phosgene in solid phase
Having addressed the origins of photodissociation prod-
E→TЈ,RЈ,VЈ. The recombination of dissociation fragments
can lead to internally excited molecule on the ground elec-
tronic state. Before thermalization, the internal energy ͑RЈ
and VЈ͒ can be converted to translational motion and even-
tually lead to desorption/ejection.
ucts, we now discuss briefly the energetics of this process.
This issue is treated more comprehensively in the companion
paper. At h␯ϭ5.0 eV, the total translational energy is:
͗E
͘
trans totalϭ͗Etrans͘COϩ2͗Etrans͘Clϭ0.70 eV. As shown in the
companion paper, the flux-weighted mean internal energy of
CO at h␯ϭ5.0 eV is ͗Eint͘ϭ0.35 eV. In the gas phase, break-
ing the two Cl–C bonds in Cl2CO requires 3.56 eV.13 As-
suming the same energy requirement on the solid film, the
excess energy available is Eexϭ1.44 eV, of which 74% is
distributed in products. The remaining excitation energy
͑26%͒ is likely lost to surrounding molecules.
The threshold photon energy for the photodissociation of
Cl2CO in the solid film is near 3.5 eV, since the intensity of
product CO, although much less than those at shorter wave-
lengths, is still visible in the TOF spectrum ͑Fig. 7͒. This
threshold is lower than that in the gas phase ͑ϳ4.0 eV͒ and is
almost the same as the minimum energy required for break-
ing two CO–Cl bonds in gas phase Cl2CO.13 Such a red shift
in absorption threshold from the gas phase value may result
from the stabilization of the electronically excited state by
polarization effects of surrounding molecules.
C. Thermal desorption
Unlike fast Cl2CO from the photoejection channel, the
yield of slow Cl2CO ͑͗Etrans͘у40 meV͒ depends nearly ex-
ponentially on laser pulse energy ͑Fig. 4͒. This nonlinear
power dependence, along with the positive correlation be-
tween ͗Etrans͘ and laser pulse energy ͑Fig. 8͒, suggests a ther-
mal origin. Similar laser-induced thermal desorption has
been observed in a number of other systems.3,5,7 The initial
electronic excitation, if does not lead to the prompt desorp-
tion of photochemical products, should eventually be con-
verted in part to thermal energy. This can come from a num-
ber of pathways, e.g., the cage effect discussed above, the
trapping of photochemical products within the film, and the
inter-molecular quenching of electronic excitation. Due to
the poor thermal conductivity of the molecular film, tempera-
ture within the light penetration depth can be raised rapidly
during the laser pulse duration. This effectively results in the
thermal desorption/vaporization of molecular Cl2CO from
the solid film. Based on the flux-weighted mean translational
temperature, the transient temperature rise necessary for de-
sorption is of the order ⌬Tϳ140 K. At much higher pulse
energies, the rapid accumulation of thermal energy can lead
to the ‘‘explosion’’ or ablation phenomenon reported
earlier.2,5,6
B. Photoejection: E˜TЈ transfer
While the nonthermal photodesorption/ejection of parent
molecules from solid films has been reported in a number of
systems,1,4,7 the present study establishes unambiguously the
single photon nature for this mechanism. This channel is
observed at the lowest laser pulse energy ͑20 ␮J/cm2 in Fig.
5͒ and the yield depends linearly on pulse energy ͑Fig. 4͒.
These fast Cl2CO molecules cannot result from simple colli-
sion between Cl2CO and nascent photodissociation products
in the surface region of the film, since ͗Etrans͘ of neither CO
nor Cl can give rise to ͗Etrans͘ of fast Cl2CO. Unlike those of
photodissociation products, the translational distribution of
fast Cl2CO is also independent of h␯ ͑Table I͒. We believe
these fast Cl2CO molecules must result from an efficient
E→TЈ energy transfer within the surface region of the film.
For a process originates below the surface, the Cl2CO mol-
ecule should loss most of the excess energy on its way out.
How can the electronic excitation energy convert effi-
ciently to translational motion? One most likely source is the
‘‘cage effect’’ discussed earlier. If a Cl2CO molecule absorbs
a photon, it is promoted to the dissociative excited state,
which in the gas phase should lead exclusively to
dissociation.13 In the solid film, the separating fragments on
the excited potential should encounter surrounding mol-
ecules and may be forced to recombine. Meanwhile, this pro-
cess should exert force to surrounding molecules. If it occurs
in the second molecular layer, the molecule in the first layer
can be ejected into the gas phase. On the other hand, a partial
cage effect also exists for an excited molecule in the first
molecular layer. Upon quenching or recombination, this mol-
ecule can also be ejected into the gas phase.
It is interesting to note that the thermal desorption chan-
nel is observed even at photon energies ͑2.3 eV͒ much lower
than the threshold for molecular Cl2CO absorption. This can
be attributed to absorption by color centers ͑such as impurity
molecules͒ within the solid film. We do not know the origin
of these color centers. One possible candidate is impurity
Cl2. A small percentage of Cl2CO can dissociate in the dos-
ing tube or the gas handling system to form Cl2, whose ab-
sorption threshold in the solid state is near h␯ϭ2.3 eV.16
V. CONCLUSIONS
Single photon electronic excitation of Cl2CO in the sur-
face region of a solid film can either dissociate to form CO
(g) and Cl (g) or convert its electronic energy to transla-
tional motion in Cl2CO (g) ͑photoejection͒. The translational
energy distribution of CO (g) from the photodissociation
channel is bimodal: while the fast channel retains nascent
energetic information from the photodissociation reaction
and is photon energy dependent, the slow channel is com-
pletely thermalized to the substrate temperature and likely
originates below the topmost surface. In addition to photo-
ejection of parent Cl2CO, thermal desorption also occurs at
high laser pulse energies. This thermal desorption channel
can be attributed to transient laser heating due to initial pho-
ton absorption by Cl2CO and/or other impurity color centers
within the solid film.
Of course, the exact mechanism for photoejection can be
much more complex. For example, the caging dynamics can
also involve other energy transfer channels, such as
J. Chem. Phys., Vol. 104, No. 20, 22 May 1996
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38

Q.-S. Xin and X.-Y. Zhu: Photochemistry of phosgene in solid phase
7903
6 L. M. Cousins, R. J. Levis, and S. R. Leone, J. Chem. Phys. 91, 5731
͑1989͒.
7 Z. J. Sun and J. M. White, J. Phys. Chem. 98, 4641 ͑1994͒.
8 R. R. Cavanagh, D. S. King, J. C. Stephenson, and T. F. Heinz, J. Phys.
Chem. 97, 786 ͑1993͒.
9 X.-Y. Zhu, Annu. Rev. Phys. Chem. 45, 113 ͑1994͒.
10 Q.-S. Xin and X.-Y. Zhu, J. Chem. Phys. 104, 4274 ͑1996͒.
11 Q.-S. Xin and X.-Y. Zhu, J. Chem. Phys. 104, 7904 ͑1996͒.
12 Q.-S. Xin and X.-Y. Zhu, Phys. Rev. Lett. 75, 4776 ͑1995͒.
13 H. Okabe, Photochemistry of Small Molecules ͑Wiley, New York, 1978͒.
14 W. L. Gurthrie, T.-H. Lin, S. T. Ceyer, and G. A. Somorjai, J. Chem.
Phys. 76, 6398 ͑1982͒.
ACKNOWLEDGMENTS
This work was supported by the National Science Foun-
dation, Grant No. CHE-9415337. Partial support from the
Camille and Henry Dreyfus New Faculty Award and the Ma-
terials Technology Center at SIUC is also acknowledged.
1 N. Nishi, H. Shinohara, and T. Okuyama, J. Chem. Phys. 80, 3898 ͑1984͒.
2 K. Domen and T. J. Chuang, Phys. Rev. Lett. 59, 1484 ͑1987͒.
3 W. C. Natzle, D. Padowitz, and S. J. Sibener, J. Chem. Phys. 88, 7975
͑1988͒.
4 I. Harrison, J. C. Polanyi, and P. A. Young, J. Chem. Phys. 89, 1498
͑1988͒; K. Leggett, J. C. Polanyi, and P. A. Young, ibid. 93, 3645 ͑1990͒;
J. C. Polanyi and P. A. Young, ibid. 93, 3673 ͑1990͒.
5 J. Kutzner, G. Lindeke, K. H. Welge, and D. Feldmann, J. Chem. Phys.
90, 548 ͑1989͒.
15 A. L. Harris, J. K. Brown, and C. B. Harris, Annu. Rev. Phys. Chem. 39,
341 ͑1988͒.
16 L. M. Cousins and S. R. Leone, Chem. Phys. Lett. 155, 162 ͑1989͒.
J. Chem. Phys., Vol. 104, No. 20, 22 May 1996
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:
128.252.67.66 On: Wed, 24 Dec 2014 17:07:38