Adiabatic and diabatic dynamics in the photodissociation of CH2BrCl

Article abstract of DOI:10.1039/b004349o

The photodissociation dynamics of chlorobromomethane (CBM) were investigated between 193 and 242 nm by resonance-enhanced multiphoton ionization (REMPI) with time-of-flight mass spectrometry (TOFMS). Translational energy distributions, anisotropy parameters, and Br: Br(*) branching ratios were determined at 193 and 235 nm to explore the non-adiabatic dynamics near the avoided crossing. Additional measurements were made at intermediate wavelengths to characterize the wavelength dependence of the Br and Br(*) anisotropy parameters. The non-adiabatic crossing probabilities calculated by applying a one-dimensional Landau-Zener model were relatively insensitive to the excitation wavelength, indicating that the avoided crossing between 3A' and 4A' potentials lies in the exit channel. Additionally, we have determined the partial absorption cross sections for the excited states that contribute to the ultraviolet absorption spectrum of CBM.

Full text of DOI:10.1039/b004349o

View Article Online / Journal Homepage / Table of Contents for this issue
Adiabatic and diabatic dynamics in the photodissociation of CH BrCl
2
Peng Zou, W. Sean McGivern and Simon W. North*
Department of Chemistry, T exas A&M University, P.O. Box 30012, College Station, T X 77842,
USA. E-mail: north=mail.chem.tamu.edu
Received 30th May 2000, Accepted 6th July 2000
Published on the Web 4th August 2000
The photodissociation dynamics of chlorobromomethane (CBM) were investigated between 193 and 242 nm
by resonance-enhanced multiphoton ionization (REMPI) with time-of-Ñight mass spectrometry (TOFMS).
Translational energy distributions, anisotropy parameters, and Br : Br* branching ratios were determined at
193 and 235 nm to explore the non-adiabatic dynamics near the avoided crossing. Additional measurements
were made at intermediate wavelengths to characterize the wavelength dependence of the Br and Br*
anisotropy parameters. The non-adiabatic crossing probabilities calculated by applying a one-dimensional
LandauÈZener model were relatively insensitive to the excitation wavelength, indicating that the avoided
crossing between 3A@ and 4A@ potentials lies in the exit channel. Additionally, we have determined the partial
absorption cross sections for the excited states that contribute to the ultraviolet absorption spectrum of CBM.
and Das concluded that the change in the branching ratio on
Introduction
Ñuorination could be due to low-frequency A@ vibrations that
The photodissociation dynamics of alkyl bromides have
become the subject of signiÐcant study in recent years pri-
marily because bromine atoms are estimated to be 40 times as
efficient as chlorine atoms at destroying stratospheric ozone
on an atom-for-atom basis.1,2 Investigations of monohalogen
alkyl photodissociation are numerous and have revealed that
non-adiabatic curve crossing dynamics play an important role
in the ultraviolet photodissociation of these compounds.
There are three electronic excited states that contribute to the
couple the 1Q and 3Q surfaces. However, anisotropy
1
0
parameters were not measured, again calling the nature of the
initial excited states into question.
In the present work, we have examined the photo-
dissociation of chlorobromomethane (CBM) and have com-
pared these results with those of CH Br to identify the e†ects
3
of halogen substitution. There are four energetically allowed
dissociation channels thought to contribute in this wavelength
region for CBM:
Ðrst UV absorption band of CH X (X \ Cl, Br and I) systems
3
hl
labeled 3Q , 3Q and 1Q .3 Among these states, only the 3Q
CH BrCl ÈÈÈ
Õ
CH Cl ] Br(2P
3@2)
(1)
(2)
(3)
(4)
1
0
1
0
1
2
2
state diabatically correlates to X(2P ) ] CH , while the 3Q
1@2
3@2
3
hl
CH BrCl ÈÈÈ
Õ
CH Cl ] Br(2P
1@2)
and 1Q states correlate to X(2P ) ] CH products. Ama-
2
2
1
3
tatsu et al. have shown theoretically that for CH I photo-
hv
3
CH BrCl ÈÈÈ
Õ
CH Br ] Cl(2P
3@2)
1@2)
dissociation the molecular geometry in the region of the
2
2
conical intersection has a strong e†ect on the observed
product distribution,4 which is consistent with experimental
studies.5 Similar behavior has been observed in the ultraviolet
CH BrCl È Èh vÈ
Õ
CH Br ] Cl(2P
2
2
Throughout this paper, we adopt the notation X and X* to
photodissociation of CH Br.6h8
describe the X(2P ) and X(2P ) products.
3
3@2
1@2
In contrast to monohalogen alkyl compounds, however,
Multihalogenated alkanes with two di†erent halogen atoms
have lower intrinsic symmetry than their monohalogen ana-
logues, resulting in a splitting of the degenerate 3Q and 1Q
there have been relatively few investigations of multihalogen
alkyl halides. Attempts to understand the e†ects of the substi-
tution of halogen atoms have focused on the measurement of
quantum yields for the ground state and spinÈorbit excited
fragments. Godwin et al. measured I(2P ) : I(2P ) ratios for
a number of halogenated compounds and, including other
results, characterized the alkyl halides into ““familiesÏÏ with
similar dynamic features.9 Based on the observed trends, the
authors proposed that the substitution of an electronegative
functionality shifted the location of the avoided crossing
towards the exit channel. However, in the absence of mea-
sured product anisotropy parameters, it was assumed that
1
1
excited states into [2A@, 1AA] and [4A@, 2AA] states, respec-
tively. In the present paper, we use the notation [X, Y] to
represent the formerly degenerate set of excited states resulting
1
@2
3@2
from the decrease in symmetry. In addition, the 3Q excited
0
state becomes 3A@, and the ground state becomes 1A@. In C
s
symmetry, the 3A@ and 4A@ components interact to form an
avoided crossing. Measurement of the anisotropy parameters
and X : X* branching ratios provides a means to examine and
quantify the non-adiabatic dynamics in this region of the
potential energy surfaces.
only the 3Q state was initially excited, making it unclear if
Photofragment translational spectroscopy (PTS) experi-
ments by Tzeng et al. on the photodissociation of CBM at 193
and 248 nm suggested that the dissociation occurs on a repul-
sive surface following a predominately parallel transition at
both wavelengths.11 Although the PTS experiments were
unable to resolve the individual ground state and excited state
products, the observed marginal dependence of the anisotropy
0
the e†ects of the substitution were due to a shift in the loca-
tion of the avoided crossing or a change in the nature of the
initially excited electronic state(s). A more recent study on per-
Ñuorinated alkyl iodides by Kavita and Das10 found that the
branching ratio of the perÑuorinated species was too low to
be explained by the model proposed by Godwin et al. Kavita
DOI: 10.1039/b004349o
Phys. Chem. Chem. Phys., 2000, 2, 3785È3790
This journal is ( The Owner Societies 2000
3785

View Article Online
parameters with respect to wavelength implied that the nature
of the initially excited electronic states remained relatively
constant. However, more recent work by our group on the
photodissociation of CBM from 248 to 267 nm revealed that
anisotropy parameters for the Br loss channel had a distinct
wavelength dependence.12 The calculated curve crossing prob-
abilities in this wavelength region were found to be similar to
The CBM (99%) sample was purchased from Aldrich, and
used without additional puriÐcation.
Results and analysis
In order to determine the quantum yields of Br and Br*, the
individual REMPI detection efficiencies must be known. In
those observed for CH Br6,7 after accounting for the di†er-
3
the present work, the photodissociation of CH Br was
ences in mass and geometry between CBM and CH Br.
3
3
studied, and the known Br : Br* branching ratio6 was used to
We have reinvestigated the photodissociation of CBM at
calibrate the observed peak areas from a REMPI scan. The
REMPI detection efficiencies of Br : Br* were previously
determined to be 0.33 : 1 for the transitions near 267 nm.12 In
the present study, we Ðnd that the relative detection efficiency
for Br : Br* is 0.46 : 1 for the transitions near 235 nm.
Forward convolution methods were developed to Ðt both
non-cored and core-sampled data, and the non-cored TOF
proÐle Ðtting procedure has been described in detail else-
where.12 The core-sampled TOF proÐle Ðtting procedure is
similar, though restrictions are placed on the detected veloci-
ties due to the presence of the coring aperture. In core-
sampling experiments, only fragments satisfying the following
equation are detected:
shorter wavelengths in order to characterize the dissociation
dynamics over the entire UV absorption band corresponding
to the n ] p* transition of the CÈBr bond. The wavelength-
dependent anisotropy parameters and product quantum yields
have provided additional insight into the non-adiabatic curve
crossing dynamics. Based on an analysis using a one-
dimensional LandauÈZener description, the avoided crossing
appears to occur far into the exit channel. In addition, using
the wavelength-dependent quantum yields and anisotropy
parameters for each channel, we have decomposed the UV
absorption spectrum into individual contributions from the
[
2A@, 1AA], 3A@ and [4A@, 2AA] excited electronic states. The
deconvolution reveals a small contribution from the [2A@,
1
tions from the 3A@ and [4A@, 2AA] states centered near the
AA] states at long wavelengths and nearly equal contribu-
(v2 [ v2)1@2t O a,
(5)
z
where v is the center-of-mass speed, v is the speed along the
absorption maximum.
z
detection axis, t is the time required for a fragment to reach
the coring aperture, and a is the radius of the coring aperture.
Since the Ñight time is dependent on the velocity along the
Experimental
detection axis, v , this equation must be solved iteratively to
The experimental set-up has been described in detail else-
where,12 and only the general features are described here. A
pulsed molecular beam of 5% CBM in D1 atm He was colli-
mated and intersected at right angles with one or two laser
beams. In the one-color experiments, the dissociation/probe
beam was the doubled output of a pulsed dye laser operating
with Coumarin 480 dye. The dissociation wavelengths were
chosen to coincide with closely spaced 2 ] 1 resonance-
enhanced multiphoton ionization (REMPI) transitions of Br
at 233.68 nm (4p 2P ] 6p 2P ) and Br* at 235.21 nm (4p
z
obtain v
z, max
coring aperture for a speciÐc center-of-mass velocity. In prac-
, the maximum v that can pass through the
z
tice, a table of v values for each set of ion optic voltages
z, max
was calculated for an entire center-of-mass velocity grid and
archived. In order to Ðt the core-sampling TOF proÐle, a trial
center-of-mass speed distribution, g(v), of the form19
A
v
B
a
A
v
B
b
g(v) \
1 [
,
v
v
max
max
3@2
3@2
2
P
] 6p 2S ), and the product quantum yields were
where v
is the thermodynamic maximum velocity and a
1@2
1@2
max
assumed to remain constant over the small wavelength di†er-
ence. In the two-color experiments, the output of an excimer
laser operating on the ArF transition was counterpropagated
with the pulsed dye laser beam running Coumarin 500 dye. A
pile-of-plates polarizer was used to polarize the 193 nm
excimer beam, and the reported anisotropy parameters have
been corrected for the incomplete polarization. Br and Br*
photofragments were detected in the two-color experiments by
and b are adjustable parameters, was used to construct a trial
velocity distribution, f (v ), subject to the previously deter-
z
mined v
values. A Jacobian is applied to the TOF proÐle
z, max
to account for the distortion of the ion packet in the Ðeld-free
region. The f (v ) distributions are subsequently transformed to
z
time-of-Ñight space, convolved with a Gaussian instrument
response function (20 ns full width at half maximum
(FWHM)), and compared with the experimental TOF spectra.
The values of a and b are then iteratively adjusted to achieve
the best Ðt.
Fig. 1 shows a core-sampled experimental TOF proÐle and
forward convolution Ðt of the Br product at 224 nm. The
broken lines represent the contributions to the TOF spectrum
from the individual 79Br and 81Br isotopes, and the solid line
is the sum of these spectra to be compared with the experi-
2
2
] 1 REMPI transitions at 266.64 (4p 2P
@2 ] 5p 4P3@2) and
3
66.70 nm (4p 2P ] 5p 2D ), respectively.13
1@2
3@2
The resulting ions were detected using a traditional two-
stage WileyÈMcLaren14 time-of-Ñight (TOF) mass spectro-
meter coupled with a set of dual Chevron microchannel plates
(
MCP). In order to increase the velocity resolution, a copper
plate with a 3 mm diameter pinhole was placed in the Ñight
tube, allowing only those ions with small velocities perpen-
dicular to the detection axis to strike the MCP.15h17 The
core-sampling set-up has been described previously,18 but the
present work represents the Ðrst application of the core-
sampling technique on this apparatus to determine trans-
lational energy distributions for polyatomic molecules. The
anisotropy parameters were determined in the absence of the
coring aperture so that the anisotropy of the resulting spectra
was dependent only on the shape of the TOF spectra and not
on the relative intensities, which can be sensitive to slight
changes in the overlap of the two beams. The MCP output
was collected on a 400 MHz digital oscilloscope or a multi-
channel scaler and was analyzed and archived on a PC. The
REMPI scans were performed by recording boxcar-averaged
MCP signals as a function of wavelength.
mental data. The translational energy distribution, P(E ), used
T
to generate the Ðts to the present data are near-Gaussian in
shape with FWHM of D10 kcal mol~1, consistent with a
prompt dissociation process. Fig. 2 shows a TOF spectrum for
Br* at 243 nm. As an example of the intensity changes on
rotation of the polarization, we have included spectra taken at
two experimental geometries, corresponding to the laser
polarization taken vertical (s \ 0¡) and horizontal (s \ 90¡) to
the detector axis. The solid line in Fig. 2 is a forward convolu-
tion Ðt of the experimental data using the method described
above. The relative intensities of the Ðts give an anisotropy
parameter, b, of 1.6 ^ 0.2. Shown in Fig. 3 are non-core-
sampled TOF spectra for Br at 193 nm and their forward con-
volution Ðts (solid lines). As shown in Table 1, we obtain an
average translational energy of 32 kcal mol~1 with a state-
3786
Phys. Chem. Chem. Phys., 2000, 2, 3785È3790

View Article Online
Fig. 1 Core-sampled Br TOF spectrum at 224 nm. Circles represent
experimental data, and the solid line is the forward convolution Ðt.
The broken lines show the individual contributions from the 79Br and
81Br isotopes.
averaged value of 31.5 kcal mol~1, in good agreement with
the results of Tzeng et al.11 The FWHM of the P(E ) distribu-
T
tions for both Br and Br* are similar and are also found to be
in good agreement with the translational energy distribution
reported by Tzeng et al. As discussed in the Experimental
section, we believe that a Ðt to the non-core-sampled TOF
spectra provides a more accurate determination of the anisot-
Fig. 3 Non-core-sampled Br TOF spectra at 193 nm. The angle
between the photolysis laser polarization and detector axis, s, is indi-
cated for each spectrum. The solid lines represent a forward convolu-
tion Ðt with an anisotropy parameter of 0.53. The broken lines show a
Ðt using an anisotropy parameter of 1.2, as suggested in ref. 11.
ropy parameter, because it depends only on the shape of the
horizontal and vertical proÐles and not on their relative inten-
sities. However, anisotropy parameters determined using the
relative intensities from core-sampled TOF spectra were fully
consistent with the non-core-sampled results and within the
cited error bounds. The values of the anisotropy parameters at
193 nm for both Br and Br* are inconsistent with the pre-
viously measured state-averaged anisotropy parameter of
Tzeng et al. of 1.2.11 The broken lines in Fig. 3 show Ðts at
both geometries assuming a P(E ) distribution identical with
T
that determined from the data with an anisotropy parameter
of 1.2. For this anisotropy, the vertical TOF spectrum would
show a signiÐcantly larger decrease at the slowest center-of-
mass velocities than was observed. The origin of this discrep-
ancy is unclear. The measured anisotropy parameters as a
function of wavelength are shown in Fig. 4 and include data
from our previous CBM photodissociation study.12
Fig. 2 Core-sampled Br* TOF spectra at 243 nm. Circles represent
experimental data, and the solid lines are the forward convolution Ðts.
The angle between the photolysis laser polarization and detector axis,
s, is indicated for each spectrum.
Table 1 Total translational energies, anisotropy parameters, and quantum yields in the ultraviolet photodissociation of CBM at several pho-
tolysis wavelengths
E
/
SE T/
trans
E
E
Quantum
yield
avail
trans
avail
j/nm
kcal mol~1
kcal mol~1
b
Br
1
2
2
2
Br*
1
2
2
93
34
48a
66.7a
80.3
54.6
47.7
39.6
32 ^ 1
29 ^ 1
22 ^ 1
24 ^ 1
0.40
0.53
0.46
0.60
0.53 ^ 0.10
0.73 ^ 0.10
1.1 ^ 0.1
0.82 ^ 0.10
0.80 ^ 0.10
0.86 ^ 0.10
0.91 ^ 0.10
0.7 ^ 0.1
93
35
48a
66.7a
69.8
44.1
37.2
29.1
29 ^ 1
24 ^ 1
20 ^ 1
20 ^ 1
0.42
0.55
0.53
0.68
0.49 ^ 0.20
1.3 ^ 0.2
1.5 ^ 0.2
1.4 ^ 0.2
0.18 ^ 0.10
0.20 ^ 0.10
0.14 ^ 0.10
0.09 ^ 0.10
2
a Ref. 12.
Phys. Chem. Chem. Phys., 2000, 2, 3785È3790
3787

View Article Online
dynamics associated with the individual halogen loss channels
may be studied independently.
In order to understand the excited states involved in the
ultraviolet photodissociation of CBM, the absorption spec-
trum was deconvoluted into contributions from the [2A@,
1AA], 3A@ and [4A@, 2AA] states. Individual Gaussian functions
in energy were used to model the absorption from the three
individual excited electronic states.22 Chlorine atoms were
observed only at 193 nm, and the n ] p* contribution of the
CÈCl bond is very small even at this wavelength. A fourth
Gaussian contribution to the absorption spectrum at wave-
lengths shorter than D200 nm was included to account for
this small contribution. By considering both the wavelength-
dependent quantum yields of Br and Br* and the relative
parallel and perpendicular contributions to each product
channel, the excitation probabilities for each excited electronic
state may be determined. The shapes and locations of the
three Gaussian functions were adjusted until the relative con-
tributions from each state reproduced the experimental
absorption spectrum and gave the correct ratio of parallel to
perpendicular contributions at each wavelength studied. Since
both the [2A@, 1AA], and [4A@, 2AA] states may result in per-
pendicular bromine, we assume that the only perpendicular
contribution at wavelengths longer than 248 nm comes from
the [2A@, 1AA] state. As shown in Fig. 4, the Br* anisotropy
parameters show little change within the experimental error
from 267 to 248 nm, indicating that the electronic state from
which the Br* product is derived is not changing. However,
the Br component shows a clear increase with decreasing
wavelength, indicating that the 3A@ state is preferentially
populated with increasing photolysis energy. A signiÐcant
contribution from the [4A@, 2AA] state would tend to decrease
the Br* anisotropy parameter with decreasing wavelength,
which is not the observed trend. Therefore, the assumption
that only the [2A@, 1AA] state contributes to the red-wing
absorption appears to be reasonable. The resulting decompo-
sition is shown in Fig. 5. The [2A@, 1AA] state is seen to make
a very small contribution to the overall absorption band,
though it plays a prominent role in the far red-wing disso-
ciation dynamics ([250 nm). The 3A@ and [4A@, 2AA] states
are very close in energy and have nearly equal widths, though
the 3A@ absorption is slightly broader and makes a signiÐcant
contribution at longer wavelengths.
Fig. 4 Wavelength-dependent anisotropy parameters for the pho-
tolysis of CBM from 193 to 267 nm. The open circles represent Br
anisotropy parameters and the closed circles are Br*. Photolysis in the
shorter wavelength region is dominated by the 3A@ (3Q ) and [4A@,
0
2
AA] (1Q ) states, but the [2A@, 1AA] (3Q ) state becomes much more
1
1
signiÐcant at wavelengths longer than 248 nm.
The observed anisotropy parameter, b, is a measure of the
orientation of the transition dipole moment with respect to
the recoil velocity in the limit of prompt dissociation. In the
axial recoil limit, b \ 2 when the transition dipole is aligned
parallel to the recoil direction and, b \ [1 when the tran-
sition dipole is perpendicular to the recoil direction. The
observed anisotropy parameter may then be considered a
linear combination of these two limiting cases. For the ultra-
violet photodissociation of CBM, there are Ðve excited elec-
tronic states, 2A@, 1AA, 3A@, 4A@ and 2AA, that are optically
coupled to the ground state. Among them, only the 1A@ ] 3A@
transition has a transition dipole moment parallel to the CÈBr
bond.12 The 3A@ state diabatically correlates to Br* products,
but an avoided crossing between the 3A@ and 4A@ states is
present in C symmetry which results in an adiabatic corre-
s
lation of the 3A@ state to Br and the 4A@ state to Br* products.
The remaining states correlate adiabatically to ground state
products. The diabatic and adiabatic dynamics may be e†ec-
tively examined because any parallel contribution to the Br*
anisotropy parameter must be derived from a non-adiabatic
crossing from the 3A@ state. The non-adiabatic curve crossing
probability, P, at a given dissociation wavelength may be
determined by
The deconvolution of the CBM absorption spectrum in this
wavelength region shows a number of di†erences from pre-
viously reported partial absorption cross sections of CH Br.6
3
Both CBM and CH Br show relatively small contributions
3
aBArRU
BrR
A
P \
,
(6)
aBArRU ] aBrU
BrR
Br
where ai Ïs and U Ïs are the normalized parallel contributions
A
i
and quantum yields for the product channel containing
halogen i, respectively. The measured anisotropy parameters
and branching ratios are shown in Table 1 in addition to the
crossing probabilities derived from these quantities.
Discussion
Ultraviolet absorption spectrum
The ultraviolet absorption spectrum of CBM20 shows a broad
peak with a maximum at D202 nm, corresponding to an
n ] p* transition on the CÈBr bond. A second peak begins to
emerge at D195 nm which is presumably due to an n ] p*
transition of the CÈCl bond. The absorption spectra of CH Br
3
and CBM in the wavelength region between 195 and 270 nm
Fig. 5 Decomposition of ultraviolet absorption spectrum into three
components which represent the [2A@, 1AA] (3Q ), 3A@ (3Q ) and [4A@,
are very similar, and the absorption onset for the CÈCl com-
ponent of CBM corresponds closely to that of CH Cl.21 This
suggests that the interaction between these two chromophores
1
0
2
AA] (1Q ) states, respectively. A contribution from the n ] p*(CÈCl)
3
1
excitation was included at large photolysis energies and closely corre-
is probably minimal, and the electronic transitions and the
sponds to the CH Cl absorption onset.
3
3788
Phys. Chem. Chem. Phys., 2000, 2, 3785È3790

View Article Online
from the lowest energy electronic states at long wavelengths,
but the [2A@, 1AA] state in CBM appears to be shifted to sig-
niÐcantly longer wavelengths and makes a much smaller con-
tribution. The 3Q and 1Q states also show a much larger
energy di†erence than the corresponding 3A@ and [4A@, 2AA]
states in CBM. However, the CÈBr absorption spectra appear
to be very similar, and the CÈBr dissociation dynamics appear
to be uncoupled from the CÈCl bond. Such a large change in
a†ects the curve crossing probability. If the crossing point
were located at a signiÐcantly higher energy, a dramatic
change in the crossing probability would be expected as a
function of the photolysis wavelength for energies near the
crossing point.
Godwin et al. have measured I* : I branching ratios for the
photodissociation of a large number of alkyl and halogenated
alkyl iodides at 248 nm.9 The molecules were categorized into
““familiesÏÏ based on the assumption that the locations of the
avoided crossings, the non-adiabatic couplings, and the di†er-
ences in the slopes of the potentials for each molecule were
dependent only on the chemical environment surrounding the
0
1
the nature of the initially excited electronic states from CH Br
3
is somewhat unexpected. In the decomposition of the CH Br
3
absorption spectrum, Gougousi et al. do not make the
assumption that the 3Q state contributes only at the longest
1
wavelengths and Ðnd a relatively large contribution from the
a-carbon. The authors assumed that only the 3Q state is
0
3
Q
state in the red-wing of the absorption spectrum.6
populated during the photolysis, and based on observed
1
Gedanken and Rowe have performed magnetic circular
dichroism (MCD) experiments on methyl halides and note
changes in the I : I* branching ratios, proposed that the sub-
stitution of an electronegative moiety on the a-carbon shifts
the curve crossing towards the exit channel. Though the
that the contribution from the 3Q state of methyl iodide is
1
extremely small.23 A re-analysis of the MCD data based on a
assumption that the 3Q state is exclusively populated at 248
0
3
more realistic numerical model24 found results consistent with
the MCD experiments, showing only a 5% contribution from
the 3Q state of CH I. Amatatsu et al.4 have identiÐed a 17%
nm is not severe in CH I,27 anisotropy parameters were not
measured for any of the molecules studied. Therefore, it is
unclear whether the approximation holds for the entire class
of molecules studied.28 More recently, Kavita and Das mea-
sured I* quantum yields for the photodissociation of per-
Ñuorinated alkyl iodides at 266, 280 and 305 nm.10 The
quantum yields measured by Kavita and Das were signiÐ-
cantly lower than the values expected from a LandauÈZener
analysis using the f values measured by Godwin et al. The
authors proposed a di†erent explanation in which the low-
1
3
contribution from the 3Q state using trajectory calculations
1
on a nine-dimensional potential energy surface, which is con-
sistent with previous experimental results.25,26 Since the spinÈ
orbit coupling of methyl bromide is signiÐcantly smaller than
in methyl iodide, both the 3Q and 3Q contributions in
1
0
methyl bromide would be expected to decrease relative to the
1
Q state. The decomposition shown in Fig. 5 shows a very
1
small [2A@, 1AA] (equivalent to the 3Q state in C symmetry)
frequency components were believed to couple the 1Q and
1
3v
1
3Q surfaces more e†ectively in the Ñuorinated compound.
0
component, and the 3A@ (3Q ) component has become nearly
0
1
equal to the [4A@, 2AA] (1Q ) component. Though Gougousi
However, like the Godwin et al. study, anisotropy parameters
et al.6 have shown a relatively large 3Q component at long
were not measured, and the nature of the initially excited elec-
tronic states for each molecule remains unclear. Studies mea-
1
wavelengths for CH Br photodissociation, the qualitative fea-
3
tures of the deconvolution shown in Fig. 5 are consistent with
suring
anisotropy
parameters,
translational
energy
the expected behavior of the CH I partial absorption crossing
sections with decreasing spinÈorbit coupling.
distributions, and product branching ratios as a function of
wavelength are necessary to obtain a Ðrm understanding of
the e†ects of halogen substitution on a photodissociation
process.
3
Non-adiabatic dynamics
The non-adiabatic curve crossing probability, P, may be esti-
mated by using a one-dimensional LandauÈZener expression
Conclusions
The photodissociation of CBM between 193 and 242 nm has
been studied using resonance enhanced multiphoton ioniza-
tion coupled with a TOF mass spectrometer. We have pre-
sented a method of analysis for the Ðtting of core-sampled
TOF data and used it to derive translational energy distribu-
tions for each photodissociation channel. Non-adiabatic curve
crossing probabilities were calculated by measuring anisot-
ropy parameters and product quantum yields at both 193 and
A
f
v
B
P \ exp [
(7)
where f is an adjustable parameter dependent on the splitting
between adiabatic surfaces and the di†erence in the slopes of
the potential surfaces at the avoided crossing, and v is the
radial velocity at the avoided crossing. Although the LandauÈ
Zener model neglects several degrees of freedom that may
play an important role in photodissociation processes, its
application can provide a qualitative understanding of the
non-adiabatic dynamics. As shown in Table 1, the average
translational energy increases approximately linearly with
increasing photolysis energy. Though the intercept is clearly
non-zero, the behavior of the translational energy for the
wavelengths relevant to the present study is approximately in
accord with the impulsive model. Therefore, we have chosen
to estimate the radial velocity in the region of the avoided
crossing by assuming that the asymptotic translational energy
may be determined from a linear Ðt. The location of the
avoided crossing may then be estimated by Ðtting the
observed curve crossing probabilities to eqn. (7) as a function
of the radial velocity and the adjustable parameter f. The best
Ðt to the wavelength-dependent non-adiabatic curve crossing
probability strongly suggests that the energy of the avoided
crossing is located in the exit channel, though the simpliÐca-
tions present in the LandauÈZener model preclude a more
quantitative analysis. At such a curve crossing, the radial
velocity at the point of crossing is relatively insensitive to the
photolysis wavelength, and, as a consequence, only weakly
235 nm. Using these results, we have decomposed the UV
absorption spectrum into the sum of four Gaussian functions,
representing the [2A@, 1AA], 3A@, and [4A@, 2AA] electronic
states, which correspond to CÈBr bond dissociation, and a
small contribution from the n ] p* transition of the CÈCl
bond. A one-dimensional LandauÈZener analysis based on the
measured translational energy distributions strongly suggests
that the 3A@ ] 4A@ avoided crossing is located in the exit
channel.
Acknowledgements
The authors thank Professor Bruce Johnson for his helpful
comments. We also acknowledge the technical assistance of
Trieu Nguyen during these experiments. This work was sup-
ported by a Research Enhancement Grant from Texas A&M
University.
References
1
2
F. S. Rowland, Annu. Rev. Phys. Chem., 1991, 42, 731.
S. C. Wofsy, M. B. McElroy and Y. L. Yung, Geophys. Res. L ett.,
1975, 2, 215.
Phys. Chem. Chem. Phys., 2000, 2, 3785È3790
3789

View Article Online
3
4
R. S. Mulliken, Phys. Rev., 1942, 61, 277.
16 H. J. Hwang, J. Griffiths and M. A. El-Sayed, Int. J. Mass
Spectrom. Ion Processes, 1994, 131, 265.
Y. Amatatsu, S. Yabushita and K. Morokuma, J. Chem. Phys.,
1
996, 104, 9783.
17 J. A. Syage, J. Chem. Phys., 1996, 105, 1007.
5
6
7
8
9
E. Hirota, Annu. Rev. Phys. Chem., 1991, 42, 1 and references
18 W. S. McGivern, R. Li, P. Zou, T. Nguyen and S. W. North,
Chem. Phys., 1999, 249, 237.
cited therein.
T. Gougousi, P. C. Samartzis and T. N. Kitsopoulos, J. Chem.
Phys., 1998, 108, 5742.
J. G. Underwood and I. Powis, Phys. Chem. Chem. Phys., 2000, 2,
19 S. W. North, A. J. Marr, A. Furlan and G. E. Hall, J. Phys.
Chem., A, 1997, 101, 9224.
20 V. L. Orkin, V. G. Khamaganov, A. G. Guschin, R. E. Huie and
M. J. Kurylo, J. Phys. Chem. A, 1997, 101, 174.
7
47.
G. N. A. Van Veen, T. Baller and A. E. De Vries, Chem. Phys.,
21 W. B. DeMore, S. P. Sander, C. J. Howard, A. R. Ravishankara,
D. M. Golden, C. E. Kolb, R. F. Hampson, M. J. Kurylo and M.
J. Molina, Chemical Kinetics and Photochemical Data for Use in
Stratospheric Modeling, Evaluation 12, Jet Propulsion Labor-
atory, Pasadena, CA, 1997.
1
985, 92, 59.
F. G. Godwin, P. A. Gorry, P. M. Hughes, D. Raybone, T. M.
Watkinson and J. C. Whitehead, Chem. Phys. L ett., 1987, 135,
1
63.
1
1
0
1
K. Kavita and P. K. Das, J. Chem. Phys., 2000, 112, 8426.
22 E. U. Condon, Phys. Rev., 1928, 32, 858.
W. B. Tzeng, Y. R. Lee and S. M. Lin, Chem. Phys. L ett., 1994,
23 A. Gedanken and M. D. Rowe, Chem. Phys. L ett., 1974, 34, 39.
24 B. R. Johnson and J. L. Kinsey, J. Phys. Chem., 1996, 100, 18937.
25 R. A. Hertz and J. A. Syage, J. Chem. Phys., 1994, 100, 9625.
26 D. H. Fairbrother, K. A. Briggman, E. Weitz and P. C. Stair, J.
Chem. Phys., 1994, 101, 3787.
227, 467.
1
1
1
1
2
3
4
5
W. S. McGivern, R. Li, P. Zou and S. W. North, J. Chem. Phys.,
1
999, 111, 5771.
S. Arepalli, N. Presser, D. Robie and R. J. Gordon, Chem. Phys.
L ett., 1985, 117, 64.
27 A. T. J. B. Eppink and D. H. Parker, J. Chem. Phys., 1998, 109,
4758.
28 See, for example, W. E. Kang, K. W. Jung, K. H. Jung and H. J.
Hwang, J. Phys. Chem., 1994, 98, 1525.
W. C. Wiley and I. H. McLaren, Rev. Sci. Instrum., 1955, 26,
1
150.
R. Ogarzalek-Loo, J.-P. Haerri, G. E. Hall and P. L Houston, J.
Chem. Phys., 1989, 90, 4222.
3790
Phys. Chem. Chem. Phys., 2000, 2, 3785È3790