Time-resolved infrared absorption studies of the solvent-dependent vibrational relaxation dynamics of chlorine dioxide

Article abstract of DOI:10.1063/1.2000234

We report a series of time-resolved infrared absorption studies on chlorine dioxide (OClO) dissolved in H2O, D2O, and acetonitrile. Following the photoexcitation at 401 nm, the evolution in optical density for frequencies corresponding to asymmetric stretch of OClO is measured with a time resolution of 120±50 fs. The experimentally determined optical-density evolution is compared with theoretical models of OClO vibrational relaxation derived from collisional models as well as classical molecular-dynamics (MD) studies. The vibrational relaxation rates in D2O are reduced by a factor of 3 relative to H2O consistent with the predictions of MD. This difference reflects modification of the frequency-dependent solvent-solute coupling accompanying isotopic substitution of the solvent. Also, the geminate-recombination quantum yield for the primary photofragments resulting in the reformation of ground-state OClO is reduced in D2O relative to H2O. It is proposed that this reduction reflects enhancement of the dissociation rate accompanying vibrational excitation along the asymmetric-stretch coordinate. In contrast to H2O and D2O, the vibrational-relaxation dynamics in acetonitrile are not well described by the theoretical models. Reproduction of the optical-density evolution in acetonitrile requires significant modification of the frequency-dependent solvent-solute coupling derived from MD. It is proposed that this modification reflects vibrational-energy transfer from the asymmetric stretch of OClO to the methyl rock of acetonitrile. In total, the results presented here provide a detailed description of the solvent-dependent geminate-recombination and vibrational-relaxation dynamics of OClO in solution.

Full text of DOI:10.1063/1.2000234

Time-resolved infrared absorption studies of the solvent-dependent vibrational
relaxation dynamics of chlorine dioxide
Joshua C. Bolinger, Teresa J. Bixby, and Philip J. Reid
Citation: The Journal of Chemical Physics 123, 084503 (2005); doi: 10.1063/1.2000234
View online: http://dx.doi.org/10.1063/1.2000234
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/123/8?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Gas-phase photodissociation of CH3COCN at 308 nm by time-resolved Fourier-transform infrared emission
spectroscopy
J. Chem. Phys. 136, 044302 (2012); 10.1063/1.3674166
Time resolved infrared absorption studies of geminate recombination and vibrational relaxation in OClO
photochemistry
J. Chem. Phys. 121, 4795 (2004); 10.1063/1.1778373
Vibrational energy relaxation of azulene in the S 2 state. I. Solvent species dependence
J. Chem. Phys. 113, 2772 (2000); 10.1063/1.1305822
Geminate recombination and vibrational relaxation dynamics of aqueous chlorine dioxide: A time-resolved
resonance Raman study
J. Chem. Phys. 109, 2596 (1998); 10.1063/1.476873
Vibrational relaxation of ClO 2 in water
J. Chem. Phys. 108, 8461 (1998); 10.1063/1.476274
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

THE JOURNAL OF CHEMICAL PHYSICS 123, 084503 ͑2005͒
Time-resolved infrared absorption studies of the solvent-dependent
vibrational relaxation dynamics of chlorine dioxide
a͒
Joshua C. Bolinger, Teresa J. Bixby, and Philip J. Reid
Box 351700, Department of Chemistry, University of Washington, Seattle, Washington 98195
͑
Received 12 May 2005; accepted 20 June 2005; published online 29 August 2005͒
We report a series of time-resolved infrared absorption studies on chlorine dioxide ͑OClO͒ dissolved
in H O, D O, and acetonitrile. Following the photoexcitation at 401 nm, the evolution in optical
2
2
density for frequencies corresponding to asymmetric stretch of OClO is measured with a time
resolution of 120±50 fs. The experimentally determined optical-density evolution is compared with
theoretical models of OClO vibrational relaxation derived from collisional models as well as
classical molecular-dynamics ͑MD͒ studies. The vibrational relaxation rates in D O are reduced by
2
a factor of 3 relative to H O consistent with the predictions of MD. This difference reflects
2
modification of the frequency-dependent solvent-solute coupling accompanying isotopic
substitution of the solvent. Also, the geminate-recombination quantum yield for the primary
photofragments resulting in the reformation of ground-state OClO is reduced in D O relative to
2
H O. It is proposed that this reduction reflects enhancement of the dissociation rate accompanying
2
vibrational excitation along the asymmetric-stretch coordinate. In contrast to H O and D O, the
2
2
vibrational-relaxation dynamics in acetonitrile are not well described by the theoretical models.
Reproduction of the optical-density evolution in acetonitrile requires significant modification of the
frequency-dependent solvent-solute coupling derived from MD. It is proposed that this modification
reflects vibrational-energy transfer from the asymmetric stretch of OClO to the methyl rock of
acetonitrile. In total, the results presented here provide a detailed description of the
solvent-dependent geminate-recombination and vibrational-relaxation dynamics of OClO in
solution. © 2005 American Institute of Physics. ͓DOI: 10.1063/1.2000234͔
I. INTRODUCTION
modified relative to the gas phase by two phenomena:
intermolecular-vibrational relaxation and geminate recombi-
nation. Insight into both phenomena has been provided by
The photochemistry of chlorine dioxide ͑OClO͒ is of
interest due to the participation of this species in atmospheric
chemistry, and, in particular, as a reservoir species for atomic
time-resolved absorption and resonance Raman ͑TRRR͒
10,12,13,21–24
spectroscopy.
These studies have established that
1
–4
chlorine. The following reaction pathways are available to
in aqueous solution the quantum yield for OClO reformation
by OCl and O recombination is 0.9, and that this yield is
solvent dependent. Recombination of the primary photofrag-
ments occurs on the subpicosecond time scale. The TRRR
studies have shown that following OClO photoexcitation, the
photoinduced depletion in Stokes scattering intensity for
transitions involving the OClO symmetric stretch recovers
with two-time constants: a subpicosecond component corre-
sponding to recombination and a slower, 9-ps component
2
2
OClO following photoexcitation resonant with the B - A
1
2
electronic transition:
h␷
2
3
͑
͑
͑
i͒
OClO→ClO͑ ⌸͒+O͑ P ͒,
g
h␷
2
1
3
−
ii͒
OClO→Cl͑ P ͒+O ͑ ⌬ , ͚ ͒,
u
2
g
g
h␷ ⌬
2
3
−
1
g
iii͒ OClO→ClOO→Cl͑ P ͒+O ͑ ͚ , ⌬ ͒.
u
2
g
2
4,25
corresponding to intermolecular vibrational relaxation.
For gaseous OClO, the predominant photochemical pathway
The appearance of anti-Stokes intensity for transitions in-
volving the symmetric stretch was delayed ϳ5 ps relative to
the formation of OClO suggesting that excess vibrational
energy is initially deposited along another coordinate, pre-
sumably the asymmetric stretch. The 5-ps delay in the ap-
pearance of anti-Stokes intensity was taken as evidence that
intramolecular-vibrational-energy reorganization occurs on
is the production of ClO and O ͓pathway ͑i͔͒, with the quan-
tum yield for atomic chlorine production ͑⌽ ͒ having a
Cl
5
–8
maximum value of ϳ0.04.
In aqueous solution, ⌽Cl in-
1
,9–18
creases to 0.1
matrices.
and approaches unity in low-temperature
Understanding the role of the solvent in
1
5,16,18–20
defining ⌽Cl is a central issue in OClO photochemistry in
condensed environments.
1
,24,25
Since ⌽ =0.1 in aqueous solution, the production of
this time scale.
This interpretation of the TRRR results
Cl
ClO and O remains the dominant photochemical pathway in
this phase, but the reaction dynamics along this pathway are
is consistent with transient UV-visible absorption studies of
aqueous OClO where the temporal evolution in optical den-
sity could be reproduced assuming that the asymmetric-
stretch coordinate was primarily responsible for the loss of
a͒Author to whom correspondence should be addressed. Electronic mail:
preid@chem.washington.edu
1
0,12
excess vibrational energy to the solvent.
Since the asym-
0021-9606/2005/123͑8͒/084503/11/$22.50
123, 084503-1
© 2005 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-2
Bolinger, Bixby, and Reid
J. Chem. Phys. 123, 084503 ͑2005͒
metric stretch is not strongly Franck-Condon coupled to the
optical transition of interest, transient absorption experiments
provide only limited information regarding vibrational-
relaxation dynamics along this coordinate. In addition, this
coordinate is silent in the resonance Raman spectra of
solution-phase OClO. Therefore, spectroscopic techniques
that provide a direct measure of vibrational-energy deposi-
tion and relaxation along the asymmetric-stretch coordinate
are required to ascertain the role of this coordinate in the
geminate-recombination and vibrational-relaxation dynamics
of OClO.
asymmetric-stretch coordinate of OClO in both H O and
2
D O occurs on the subpicosecond time scale demonstrating
2
that geminate recombination of the primary photofragments
occurs on this time scale. The vibrational relaxation rates in
D O are approximately three times smaller than those in
2
H O, in qualitative agreement with the predictions of MD. In
2
acetonitrile, higher vibrational states are available for study
given the increased transparency of this solvent, and popula-
tion of these higher-energy states following geminate recom-
bination is observed. An ϳ10-ps delay in optical-density
evolution relative to photoexcitation is evident, and is pro-
posed to reflect a longer excited-state lifetime in acetonitrile
relative to H2O and D2O. In H2O and D2O, the vibrational-
relaxation dynamics are modeled assuming that the relax-
ation rates reflect the frequency dependence of solute/solvent
force-force correlation function consistent with the earlier
MD results. However, the relaxation dynamics in acetonitrile
are not well reproduced by this approach, potentially reflect-
ing the complex nature of the coupling between the OClO
asymmetric-stretch coordinate and specific modes of the sol-
vent. In total, the studies presented here provide a detailed
picture of the role of the asymmetric-stretch coordinate in the
condensed-phase reaction dynamics of OClO.
Theoretical techniques have also been employed to re-
fine our understanding of OClO vibrational-relaxation dy-
2
6–30
namics in solution.
Using classical molecular dynamics
͑
MD͒, Poulsen et al. were able to reproduce the lifetime of
the first excited vibrational state of the asymmetric stretch
through analysis of the solvent-solute force-force correlation
function. Benjamin and co-workers have provided a de-
tailed look into the solvent-dependent OClO vibrational-
relaxation dynamics by performing the MD simulations in
3
1
3
2,33
H O, D O, acetonitrile, and ethanol.
In this work, relax-
2
2
ation rates from the first excited state along the asymmetric-
stretch coordinate to all other lower-energy vibrational states
were determined, and good agreement with experiment was
observed. However, detailed information regarding the
vibrational-relaxation dynamics along the asymmetric-
stretch coordinate required to further test the predictions of
Benjamin and co-workers was not available at the time of
their study. To address this issue, we performed a time-
resolved infrared ͑TRIR͒ absorption study of OClO in
II. EXPERIMENT
The laser system employed in these studies has been
discussed in detail elsewhere; therefore, only a brief sum-
3
4
mary is presented here. A regenerative Ti:sapphire ampli-
fier ͑Spectra Physics͒ was seeded with a Ti:sapphire oscilla-
tor ͑KM Labs͒ pumped by the frequency-doubled output of a
3
4
water. In this study, relaxation rates along the asymmetric-
stretch vibrational manifold were measured through the
n=8 vibrational level. The MD results of Benjamin and co-
workers were compared to the experimentally determined
vibrational-relaxation rates, and qualitative agreement was
observed. However, the absolute rates predicted by MD were
Nd:VO laser ͑Coherent͒. The amplifier output consisted of
4
60-fs pulses centered at 802 nm with a pulse energy of 960
µJ at a repetition rate of 1 kHz. The amplifier output was
split into two beams using a 60/40 beamsplitter. The higher-
energy beam was used to pump an optical parametric am-
pliflier ͑OPA, Quantronix TOPAS͒, and the signal and idler
beams from the OPA were used for difference frequency gen-
3
4
five times less than those determined by experiment. This
difference raises the following question, does this discrep-
ancy in rates exist only for H O, or are similar differences
eration ͑DFG͒ in AgGaS ͑type I͒, to produce the probe field
2
2
observed in other solvents? In our previous TRIR study, the
population of levels above n=8 along the asymmetric-stretch
coordinate was not observed. However, the ϳ17 000 cm− of
excess energy available to OClO following geminate recom-
bination dictates that levels as high as n=15 could be popu-
lated if all the excess vibrational energy available to OClO
upon recombination is deposited along the asymmetric-
stretch coordinate. However, probe frequencies resonant
with the upper levels of the asymmetric-stretch vibrational
manifold overlap with the librational band of water, making
measurements in this frequency region quite difficult. By
measuring the relaxation dynamics in more transparent sol-
vents, information regarding the relaxation dynamics of
higher-energy vibrational states along the asymmetric-stretch
coordinate can be obtained.
centered at 9.5 µm. The lower-energy beam derived from the
amplifier output was frequency doubled using a 100-µm-
thick ␤-BBO crystal ͑type I͒ to provide the 401-nm pump
field.
1
The pump field was delivered to a retroreflector mounted
on a motorized delay stage to allow for temporal delay of the
pump relative to the probe. To minimize the contribution of
rotational dynamics to the measured optical-density evolu-
tion, the polarization of the pump was oriented to 54.7° rela-
tive to the probe using a zero-order half-wave plate. The
20-µJ pump field was weakly focused to a 1.5-mm spot size
in the sample. The change in sample optical density was
found to vary linearly with pump power. The 9.5-µm probe
field was collimated using two parabolic off-axis reflectors,
then passed through a ZnSe 50/50 beamsplitter. The trans-
mitted beam used as a reference while the reflected beam
was used as the probe. The probe ͑0.15-µJ per pulse͒ was
focused on the sample using a parabolic off-axis reflector.
The spatial and temporal overlap of the pump and probe
fields was measured using the change in optical density in a
2
5
We present here a series of TRIR absorption studies in
which the geminate-recombination and vibrational-relaxation
dynamics of OClO are investigated in three solvents: H O,
2
D O, and acetonitrile. The central results of this study are as
2
follows. The appearance of vibrational energy along the
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-3
Solvent-dependent vibrational relaxation of OClO
J. Chem. Phys. 123, 084503 ͑2005͒
Si wafer following a 401-nm photoexcitation. The Si re-
sponse was used to determine the instrument response
͑
120±50-fs full width half maximum͒.
After passing through the sample, the probe field was
recollimated and delivered to a 0.25-m monochromator
Jarrell Ash͒ equipped with a 100-groove/mm grating blazed
at 5.6 µm to isolate individual probe frequency components
͑
−
1
with a resolution of 30 cm . The reference field also passed
through the monochromator to ensure that identical spectrum
components of the probe and reference were measured. De-
tection of the probe and reference intensities was accom-
plished using two LN -cooled MCT detectors ͑Infrared Prod-
2
ucts͒. The detector outputs were delivered to separate gated
integrators, and the integrator outputs were subtracted on a
shot-to-shot basis. A mechanical chopper operating at half
the frequency of the amplifier was used to block every other
pump pulse, and successive probe signals were subtracted to
determine the pump-induced change in the sample optical
density. The CaF cell windows demonstrated a nonlinear
2
response to the 400-nm actinic field employed. To remove
the contribution of the CaF2 response from the optical-
density evolution assignable to OClO, a biphasic exponential
decay with time constants of 0.61 and 2.34 ps was con-
structed and subtracted from the data. The robustness of this
approach was established by performing this procedure for a
variety of pump powers, and the resulting evolution in opti-
cal density assignable to OClO was not affected. In addition,
FIG. 1. Evolution in optical density for aqueous OClO following photoex-
citation at 401 nm measured at probe frequencies ranging from 1094 to
−
1
1
014 cm . The probe wavelength at which the data were obtained is indi-
cated. The data are presented as the circles, and the solid lines represent the
best fit to the data by a sum of exponentials with fit parameters provided in
Table I.
2
stretch coordinate of aqueous OClO in the ground B state
this process was compared to removal of the CaF response
1
2
−
1
has a harmonic frequency of 1105 cm and anharmonicity
of 11 cm− such that fundamental transition for this coordi-
nate is at 1094 cm . Fourier transform infrared ͑FTIR͒ mea-
by direct subtraction of the response measured in the absence
of OClO, and the resulting evolution in optical density as-
signable to OClO was in agreement with that obtained using
the exponential-decay subtraction procedure.
1
−
1
surements demonstrated that frequency of this transition did
−
1
Samples were prepared by heating potassium chlorate
and oxalic acid dihydrate in a 2.4M H SO ͑aq͒ at 60 °C to
not change within experimental error ͑±2 cm ͒ for the sol-
vents studied here ͑data not shown͒. The evolution in optical
density following OClO photoexcitation was measured at
2
4
produce gaseous OClO. The resulting yellow gas passed
through a drying tube filled with P O , and was then bubbled
14-cm− intervals starting from 1094 cm and ending at
1
−1
2
5
through the solvent of interest. This stock solution was di-
luted until a concentration of 50 mM was obtained, with the
diluted solution used for the experiments reported here. The
kinetics reported here was invariant with a twofold reduction
in concentration. All reagents and solvents ͑Fisher Scientific͒
014 cm−1 in H O and D O, or ending at 958 cm in aceto-
−1
1
2 2
nitrile. As the probe frequency is reduced, the evolution in
optical density provides a measure of the vibrational-
relaxation dynamics involving higher-energy states along the
asymmetric-stretch vibrational manifold. Probe frequencies
were of the highest purity available except for D O ͑99%,
2
−1
−1
at 1094 cm through 1080 cm demonstrate reductions in
optical density consistent with pump-induced depletion of
the ground state; therefore, the optical-density evolution at
these frequencies provides information involving the lowest-
energy levels of the asymmetric-stretch vibrational manifold.
Cambridge Isotopes͒. The sample was delivered to a Teflon
flow cell ͑Harrick Scientific͒ equipped with CaF windows.
2
A 56-µm cell path length was used for H O and D O, and a
2
2
150-µm pathlength was employed in acetonitrile. The time-
dependent optical-density evolution was analyzed by fitting
the data to a sum of exponentials using the Levenberg-
Marquardt algorithm. Goodness of fit was judged using the
−
1
Intermediate probe frequencies ͑1066 and 1053 cm ͒ pro-
vide information on the dynamics associated with mid-lying
levels of the asymmetric-stretch manifold. Finally, the lowest
2
reduced ␹ values and by visual inspection of the residuals.
−
1
Errors reported here represent one standard deviation from
the mean as determined from at least four measurements.
probe frequencies investigated ͑1040–958 cm ͒ provide a
measure of the dynamics of mid and upper vibrational levels
along the asymmetric-stretch coordinate. The description of
optical-density evolution presented in Figs. 1–3 will be par-
titioned into three frequency regions consistent with the
above discussion. Time constants for the optical-density evo-
lution as a function of probe frequency are summarized for
each solvent in Tables I ͑H2O͒, II ͑D2O͒, and III ͑acetoni-
trile͒.
III. RESULTS
A. Experimental results and kinetic analysis
Figures 1–3 present the measured evolution in optical
density following the photoexcitation of OClO dissolved in
H O, D O, and acetonitrile, respectively. The asymmetric-
2
2
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-4
Bolinger, Bixby, and Reid
J. Chem. Phys. 123, 084503 ͑2005͒
FIG. 3. Evolution in optical density for OClO dissolved in acetnotritile
following photoexcitation at 401 nm measured at probe frequencies ranging
from 1094 to 952 cm . The probe wavelength at which the data were ob-
tained is indicated. The data are presented as the circles, and the solid lines
represent the best fit to the data by a sum of exponentials with fit parameters
provided in Table III.
FIG. 2. Evolution in optical density for OClO dissolved in D O following
2
photoexcitation at 401 nm measured at probe frequencies ranging from 1094
−
1
−
1
to 1014 cm . The probe wavelength at which the data were obtained is
indicated. The data are presented as the circles, and the solid lines represent
the best fit to the data by a sum of exponentials with fit parameters provided
in Table II.
1
. Probe frequencies of 1094–1080 cm−1
that escapes the solvent cage, or that results in the production
of Cl and O2. In D2O ͑Fig. 2͒, the pump-induced depletion in
optical density recovers with a time constant of 26.3±4.1 ps,
roughly three times slower in comparison with H2O. A per-
The 1094-cm−1 probe is resonant with the n=0 to n=1
or fundamental transition of the asymmetric stretch; there-
fore, this frequency provides a measure of depletion and sub-
sequent recovery of ground-state OClO. An initial
instrument-response-limited reduction in optical density is
observed in all solvents, consistent with the photoinduced
sistent depletion in optical density is also observed in D O;
2
however, the residual depletion is 23±6% of the initial
depletion demonstrating that cage escape and/or atomic chlo-
rine production is enhanced in D O relative to H O. In ac-
depletion of ground-state OClO. In H O ͑Fig. 1͒, the initial
depletion in optical density recovers with a time constant of
2
2
2
etonitrile ͑Fig. 3͒, the pump-induced depletion in optical
density is followed by more complex behavior in comparison
with H O or D O. Specifically, the magnitude of the initial
9
.6±2.0 ps to reveal a small residual depletion of 10±2%,
that remains unchanged out to the longest delays investigated
100 ps͒. This residual depletion provides a measure of pho-
toexcited OClO that results in the formation of OCl and O
2
2
͑
depletion remains constant out to 31.7±2.9 ps, then recovers
with a time constant of 40±10 ps. The long-time optical-
TABLE I. Kinetic parameters determined from the TRIR study of OClO in H O.
2
␯pr͑cm−1
͒
a
A1
b
␶1 ͑ps͒
A2
␶2 ͑ps͒
A3
␶3 ͑ps͒
c
1094
1080
1066
1053
1040
1027
1014
−0.95±0.05
−0.90±0.15
−0.69±0.09
−0.59±0.08
−0.51±0.04
−0.50±0.02
−0.57±0.16
9.6±2.0
6.9±1.2
4.5±0.9
4.2±1.0
3.0±0.9
2.7±0.5
2.1±0.5
¯
¯
−0.05±0.05
offset
offset
¯
¯
¯
−0.10±0.15
0.31±0.09
0.41±0.08
0.49±0.04
0.50±0.02
0.43±0.16
8.6±2.0
7.2±1.2
5.1±0.4
4.7±1.1
4.3±1.2
¯
¯
¯
¯
¯
¯
¯
¯
¯
a
Probe frequencies.
b
Amplitudes are normalized such that ͚ ͉A ͉=1. Errors represent one standard deviation from the mean of all
i
i
measurements ͑ജ4͒ at a given probe wavelength.
c
Time constant is set to 10 000 ps to account for residual depletion.
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-5
Solvent-dependent vibrational relaxation of OClO
J. Chem. Phys. 123, 084503 ͑2005͒
TABLE II. Kinetic parameters determined from the TRIR study of OClO in D O.
2
␯pr͑cm−1
͒
a
A1
b
␶1͑ps͒
A2
␶2͑ps͒
A3
␶3͑ps͒
c
1094
1080
1066
1053
1040
1027
1014
−0.90±0.07
−0.84±0.06
−0.49±0.01
−0.49±0.01
−0.43±0.04
−0.39±0.06
−0.42±0.06
26.3±4.1
21.7±3.5
13.5±1.5
11.6±2
¯
¯
−0.10±0.08
−0.16±0.06
−0.05±0.03
−0.02±0.02
−0.06±0.03
−0.10±0.04
−0.08±0.06
offset
offset
offset
offset
offset
offset
offset
¯
¯
0.46±0.03
0.49±0.01
0.51±0.02
0.51±0.02
0.50±0.01
18.1±1.1
14.5±2
15.2±3.5
14.2±3.4
11.7±3.9
8.5±1.7
5.9±1.5
5.8±1.7
a
Probe frequencies.
b
Amplitudes are normalized such that ͚ ͉A ͉=1. Errors represent one standard deviation from the mean of all
i
i
measurements ͑ജ7͒ at a given probe wavelength.
c
Time constant is set to 10 000 ps to account for residual depletion.
density depletion in acetonitrile is 47±16% of the initial
depletion, and demonstrates that geminate-recombination ef-
ficiency is reduced relative to H O and D O. Similar behav-
followed by an increase and subsequent decay in optical den-
sity. This evolution in optical density is consistent with the
production and subsequent decay of intermediate vibrational
levels ͑n=2 to n=5͒ along the asymmetric-stretch coordinate
during the vibrational-relaxation process. The evolution in
optical density was well fit by a sum of two exponentials
corresponding to the initial appearance and subsequent decay
2
2
ior was observed in previous UV-visible transient absorption
and time-resolved resonance Raman studies.
−
1
At 1080 cm , both the n=0 to n=1, and n=1 to n=2
transitions contribute to the optical-density evolution. In all
solvents, an instrument-response-limited reduction in optical
density is observed, but the depletion amplitude is reduced
−
1
of optical density. In H O at 1066 cm , the optical density
2
appears with a time constant of 4.5±0.9 ps and decays with
−
1
relative to that observed at 1094 cm consistent with a re-
duction in the contribution of the fundamental transition to
the optical-density evolution as the probe frequency is re-
duced. The pump-induced depletion of optical density recov-
a time constant of 8.6±2.0 ps. In D O at this same fre-
2
quency, the appearance and decay time constants are
13.5±1.5 and 18.1±1.1 ps, respectively. Similar to the ob-
servation at higher probe frequencies, these time constants
demonstrate that the vibrational-relaxation rates are two- to
threefold reduced in D O compared to H O. In acetonitrile,
−
1
ers with an enhanced rate relative to 1094 cm , consistent
with probing higher vibrational states along the asymmetric-
stretch manifold. Specifically, in H O and D O the recovery
2
2
2
2
the initial pump-induced depletion in optical density persists
out to 30.0±1.0 ps followed by an appearance time constant
of 25.0±6.0 ps, but no perceptible decay is observed. At
in optical density occurs with time constants of 6.9±1.2 and
1.7±3.5 ps, respectively. In acetonitrile, the delay in recov-
ery as well as the rate of optical-density depletion recovery is
2
1053 cm− in H O, optical density appears with a time con-
1
2
−
1
shortened relative 1094 cm , with the recovery beginning at
7.7±2.6-ps following photoexcitation, and proceeding with
a time constant of 32±8 ps.
stant of 4.2±1.0 ps and decays with a time constant of
2
7.2±1.2 ps. The corresponding time constants in D O are
2
11.6±2.0 and 14.5±2.0. Finally, in acetonitrile the initial
optical-density depletion persists to 21.8±2.8 ps followed by
an appearance of optical density with a time constant of
25.0±6.0 ps and subsequent decay with a time constant of
46±16 ps.
2
. Probe frequencies of 1066–1053 cm−1
In this frequency region for OClO dissolved in H O and
2
D O, a small pump-induced reduction in optical density is
2
TABLE III. Kinetic parameters determined from the TRIR study of OClO in acetonitrile.
␯pr͑cm−1
͒
a
A1
b
␶1͑ps͒
A2
␶2͑ps͒
A3
␶3͑ps͒
c
Dwell
c
1
1
1
1
1
1
1
1
094
080
066
053
040
027
014
000
−0.53±0.15
−0.80±0.15
−0.91±0.09
0.87±0.09
40±10
32±8
−0.45±0.15
−0.20±0.15
0.05±0.11
0.11±0.12
0.46±0.02
0.45±0.03
0.52±0.18
0.47±0.01
0.47±0.02
0.47±0.02
0.49±0.02
0.52±0.05
offset
¯
¯
¯
¯
¯
31.7±2.9
27.7±2.6
30.0±0.1
21.8±2.8
16.3±1.7
15.5±1.5
15.2±1.8
15.4±1.2
13.3±0.5
12.8±0.7
11.2±1.7
11.9±1.6
offset
¯
25±6
offset
¯
25±6
46±16
¯
−0.49±0.01
−0.48±0.01
−0.43±0.16
−0.49±0.1
−0.49±0.01
−0.48±0.01
−0.50±0.02
−0.47±0.05
12.8±1
8.4±1.1
9.3±1.9
8.5±1.2
7.1±0.7
6.3±1.3
5.6±1.7
4.8±1.4
23.1±4.5
17.1±2.6
15.5±2.9
13.3±1.4
14.8±3.1
12.6±2.3
10.6±1.7
9.9±2.9
0.06±0.02
0.06±0.04
0.05±0.03
0.04±0.02
0.04±0.02
0.05±0.02
0.02±0.01
0.01±0.02
offset
offset
offset
offset
offset
offset
offset
offset
9
9
9
9
92
80
69
58
a
Probe frequencies
b
Amplitudes are normalized such ͚ ͉A ͉=1. Errors represent one standard deviation from the mean of all
i
1
measurements ͑ജ8͒ at a given probe wavelength.
c
Time constant is set to 10 000 ps to account for residual depletion.
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-6
Bolinger, Bixby, and Reid
J. Chem. Phys. 123, 084503 ͑2005͒
3
. Probe frequencies Ï1040 cm−1
vibrational-relaxation rate constants for the two models and
the comparison of the predicted evolution in optical density
derived from these models to experiment are described be-
low.
In this frequency region a pump-induced increase in op-
tical density is observed in all solvents followed by decay.
This behavior is consistent with population of higher-energy
levels of the asymmetric-stretch vibrational manifold during
the geminate-recombination process followed by vibrational
relaxation. As the probe frequency is reduced, the time con-
stants associated with the optical-density appearance and de-
cay become smaller, reflecting the increased vibrational-
1. Collisional model
In the collisional model, the level-dependent vibrational-
10,12
relaxation rate constants are given by
relaxation rates for higher-energy states. In H O, the
kn→n−1 = nk1→0.
͑1͒
2
evolution in optical density is well modeled by a sum of two
exponentials corresponding to the appearance and decay in
optical density, with time constants ranging from 3.0±0.9
In the above expression, k
ated with relaxation between levels n and n−1, and k
is the rate constant associ-
n →n−1
Ј
is
1
→0
the relaxation rate constant from level n=1 to the n=0
ground state. Geminate recombination is assumed to result in
initial population of the n=15 level along the asymmetric-
stretch coordinate. The level populations of the asymmetric-
stretch manifold were then calculated using the following
master equation:
−
1
and 5.1±0.4 ps, respectively, at 1040 cm to 2.1±0.5 and
−
1
4
.3±1.2 ps, respectively, at 1014 cm . In D O, the appear-
2
ance and decay time constants are again increased relative to
water, with time constants of 8.5±1.7 and 15.2±3.5 ps, re-
−
1
spectively, at 1040 cm and 5.8±1.7 and 11.7±3.9 ps, re-
−
1
spectively, at 1014 cm . In acetonitrile, lower probe fre-
quencies could be employed due to the increased
transparency of the solvent relative to H O and D O. In ac-
d␳n͑t͒
dt
=
͑n + 1͒k1 ␳n+1͑t͒ − nk ␳n͑t͒.
͑2͒
→0
1→0
2
2
etonitrile, a prominent long-time increase in optical density
is observed in addition to the optical-density evolution asso-
ciated with OClO vibrational relaxation. To accurately fit the
data, a sum of three exponentials was required with one of
the time constants set equal to 10 000 ps to account for the
long-time increase in optical density. Consistent with the be-
havior observed at higher probe frequencies in this solvent,
an initial delay before the onset of optical-density evolution
The master equation was solved utilizing Euler’s method
with a time step of 0.1 ps, and the results of this calculation
are the populations of each vibrational level as a function of
time. Detailed balance was initially included in this analysis,
but it was determined that the reverse process is negligible
due to the large energy-level spacings of interest relative to
the available thermal energy. The contributions of the indi-
vidual vibrational-level populations were weighted by the
monochromator bandwidth ͑Gaussian with a 30-cm full
width at half maximum͒, and the weighted contributions
were summed at a given time delay to produce the predicted
evolution in optical density. The predicted evolution was also
convolved with the instrument response; however, due to the
time scales of interest the effect of convolution was negli-
gible.
−
1
is observed. At 1040 cm , after a 16.3±1.7-ps delay the
optical density increases with a time constant of 12.8±1.0 ps
−
1
−
1
and decays with a 23.1±4.5-ps time constant. At 1014 cm ,
the delay in optical-density evolution is reduced to
1
5.2±1.8 ps followed by rise and decay times of 9.3±1.9
−
1
and 15.5±2.9 ps, respectively. Finally, at 958 cm , centered
roughly on the n=12 to n=13 transition, a 11.9±1.6-ps delay
in spectrum evolution is followed by optical-density appear-
ance and decay with time constants of 4.8±1.4 and
9.9±2.9 ps, respectively.
2. Perturbative model
Previous MD studies of OClO suggested an alternative
description of the vibrational-relaxation dynamics from that
B. Modeling of the optical-density evolution
3
1–33
of the collisional model.
In this approach, the linear de-
Two models for the vibrational-relaxation dynamics of
OClO were employed to analyze the solvent-dependent evo-
lution in optical density described above. The first model
employed was used in earlier UV-visible transient absorption
pendence of the relaxation rate constants on vibrational level
is augmented by an additional term corresponding to the fre-
quency dependence of the solute-solvent coupling. In this
approach, the level-dependent relaxation rate constants are
given by
1
0,12
studies of OClO dissolved in water and acetonitrile.
This
approach borrows from Landau-Teller theory in which the
vibrational-relaxation rate constants are predicted to scale
˜
−
2
k
= nk
F
͑␻͒ប .
͑3͒
n→n−1
1→0 q,q
Ј
3
5,36
linearly with vibrational level.
This model assumes that
˜
relaxation occurs primarily through collisions with nearby
solvent molecules; therefore, we will refer to this model as
the “collisional model.” The second model investigated also
predicts that the vibrational-relaxation rate constants will in-
crease with vibrational level; however, the linear level de-
pendence of the collisional model is augmented by an addi-
tional term that describes the frequency dependence of the
solvent-solute coupling. We will refer to this model as the
In this equation, F ͑␻͒ is the real part of the Fourier trans-
q,q
Ј
form of the solvent-solute force-force correlation function. In
Eq. ͑3͒, the linear dependence of the relaxation rate constants
on n inherent in the collisional model is retained; however,
an additional term is included that takes into account the
frequency dependence of the solvent-solute coupling. This
additional term has the effect of increasing the rate constants
relative to the values predicted by the collisional model. The
frequency-dependent coupling was derived using the MD re-
“
perturbative model.” The level dependence of the
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-7
Solvent-dependent vibrational relaxation of OClO
J. Chem. Phys. 123, 084503 ͑2005͒
FIG. 5. ͑A͒ Comparison of the experimental evolution in optical density of
aqueous OClO ͑circles͒ to that predicted using the collisional model for
vibrational relaxation along the asymmetric-stretch coordinate ͑solid line͒.
͑
B͒ Comparison of the experimental evolution in optical density of aqueous
OClO ͑circles͒ to that predicted using the perturbation model for describing
the vibrational-relaxation dynamics along the asymmetric-stretch coordinate
͑
1
solid line͒. Comparisons are presented for probe wavelengths of 1094,
−
1
040, and 1014 cm to represent the entire vibrational dynamics across the
FIG. 4. ͑A͒ The variation in the power spectrum of the solvent-solute force-
force correlation functions F͑␻͒ as a function of frequency as reported by
Benjamin and co-workers. Spectra are with reference to the asymmetric
probed asymmetric-stretch manifold.
˜
and 1014 cm−1 are reproduced; however, the rate of relax-
ation at higher levels is slightly overestimated as evidenced
stretch of OClO in water ͑dotted line͒, ethanol ͑thin line͒, and acetonitrile
͑
thick line͒. The figure inset provides an expanded view of the low-
˜
frequency region of the power spectrum. ͑B͒ The variation in F͑␻͒ as a
function of frequency for H O ͑solid line͒ and D O ͑dashed line͒.
by the modest discrepancy between experiment and theory at
2
2
−1
1
014 cm . Although the perturbation model is in qualitative
agreement with the measured evolution in optical density, it
should be noted that quantitative agreement is not achieved.
sults of Benjamin and co-workers, with their results for the
solvents of interest reproduced in Fig. 4.
−
1
Specifically, the MD results suggest that k1 =0.019 ps , a
→0
significant reduction from the value employed here and in
previous work.
3
. H O
2
Figure 5 presents a comparison between the predicted
4
. D O
2
evolution in optical density derived using the collisional and
perturbation models to experiment. Three probe frequencies,
Comparison of the vibrational-relaxation dynamics ob-
served in D O relative to H O provides a good test of the
−
1
1
094, 1040, and 1014 cm , were chosen for presentation
2
2
since they provide a measure of the dynamics that occur in
the lower, middle, and upper regions of the asymmetric-
stretch vibrational manifold, respectively. In this simulation,
perturbation model since these solvents are structurally simi-
lar, the mode frequencies of these solvents are significantly
different, and the frequency dependence of the solvent-solute
coupling is altered. Reflecting the change in solvent-solute
coupling in D O relative to H O near the frequency of the
−
1
k1
work.
was set equal to 0.105 ps consistent with previous
→0
1
0,12,24,25
The collisional model ͓Fig. 5͑A͔͒ demon-
2
2
−
1
strates reasonable agreement with experiment at 1094 cm ;
however, the experimental data do not demonstrate the delay
in recovery relative to zero time as predicted by this model.
At 1040 and 1014 cm , the maximum increase in optical
density predicted by the collisional model is delayed relative
to experiment. In contrast, the perturbation model ͓Fig. 5͑B͔͒
better reproduces the evolution in optical density. Specifi-
OClO asymmetric stretch, the MD studies predict that the
vibrational-relaxation rate constants should be reduced in
D O relative to H O. Figure 6 presents a comparison of the
2
2
−
1
predicted evolution in optical density derived using the col-
lisional and perturbative models with experiment. The com-
parisons are performed using a rate constant of k
1
→0
=0.0385 ps− determined using the time constant for optical-
1
−
1
−1
cally, the evolution at 1094 cm is more accurately repro-
duced due to the faster relaxation rates for higher vibrational
levels predicted by this model. In addition the appearance
and decay kinetics of the optical-density evolution at 1040
density recovery measured at 1094 cm ͑Table II͒. Similar
to the observation in H O, the collisional model ͓Fig. 6͑A͔͒
2
fails to accurately predict the observed evolution in optical
density. For example, at 1094 cm− the collisional model
1
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-8
Bolinger, Bixby, and Reid
J. Chem. Phys. 123, 084503 ͑2005͒
FIG. 7. ͑A͒ Comparison of the experimental evolution in optical density of
OClO dissolved in acetonitrile ͑circles͒ to that predicted using the colli-
sional model for describing the vibrational-relaxation dynamics along the
asymmetric stretch ͑solid line͒. ͑B͒ Comparison of the experimental evolu-
tion in optical density of OClO dissolved in acetonitrile ͑circles͒ to that
predicted using the perturbation model for describing the vibrations-
relaxation dynamics along the asymmetric stretch ͑solid line͒. Comparisons
FIG. 6. ͑A͒ Comparison of the experimental evolution in optical density of
OClO dissolved in D O ͑circles͒ to that predicted using the collisional
2
model for describing the vibrational-relaxation dynamics along the asym-
metric stretch ͑solid line͒. ͑B͒ Comparison of the experimental evolution in
optical density of OClO dissolved in D O ͑circles͒ to that predicted using
2
the perturbation model for describing the vibrational-relaxation dynamics
along the asymmetric stretch ͑solid line͒. Comparisons are presented for
probe wavelengths of 1094, 1040, and 1014 cm−1 to represent the entire
vibrational dynamics across the probed asymmetric-stretch manifold.
are presented for probe wavelengths of 1094, 1040, 1014, 992, and
−1
9
69 cm to represent the entire vibrational dynamics across the probed
asymmetric-stretch manifold.
predicts a delay in optical-density recovery that is not ob-
served experimentally. Also, the predicted appearance in op-
tion is in good agreement with experiment, except for the
persistent long-time offset in optical density. It will be ar-
gued below that this offset arises from vibrational-energy
transfer from the asymmetric-stretch coordinate to the me-
−
1
tical density at 1040 and 1014 cm is significantly delayed
relative to experiment. These discrepancies indicate that the
relaxation rate constants for higher-energy vibrational levels
are greater than those predicted by the collisional model.
Consistent with this observation, the perturbative model
more accurately predicts the experimental optical-density
−1
thyl rock of acetonitrile. Finally, at 969 cm the experimen-
tal optical-density evolution occurs on a significantly longer
time scale relative to the prediction of the collisional model.
In contrast to the case in H O and D O, the perturbative
2
2
−
1
evolution ͓Fig. 6͑B͔͒. Specifically, at 1094 cm a slight de-
lay in optical-density recovery is predicted, but it is not as
long as that predicted by the collisional model, and the
optical-density recovery is well reproduced. At 1040 and
model does not accurately reproduce the optical-density evo-
−1
lution in acetonitrile ͓Fig. 7͑B͔͒. At 1094 cm , the experi-
mental and predicted evolution in optical density is in good
−1
agreement; however, at the 1040, 1014, and 969 cm the
−
1
1
014 cm , both the location of the maximum increase in
predicted dynamics occur on a much shorter time scale rela-
tive to experiment.
optical density and the kinetics associated with the optical-
density evolution are well reproduced by the perturbation
model.
IV. DISCUSSION
The TRIR studies outlined here provide a detailed look
into the solvent dependence of the vibrational-energy relax-
ation dynamics along the asymmetric-stretch coordinate of
OClO. The results above demonstrate that following the ref-
ormation of ground-state OClO by geminate recombination
of the primary O and ClO photofragments, vibrational en-
ergy is deposited along the asymmetric-stretch coordinate,
and that vibrational energy is lost from this coordinate to the
solvent. In H O and D O, excess vibrational energy appears
5
. Acetonitrile
Figure 7 presents a comparison of the optical-density
evolution predicted using the collisional and perturbative
models to experiment. In addition to evolution in optical
density observed at 1094, 1040, and 1014 cm , a compari-
son is also presented at 969 cm since these data provide a
measure of relaxation dynamics for the highest-energy levels
of the asymmetric-stretch vibrational manifold. In the mod-
eling, k1 was set equal to 0.027 ps , consistent with pre-
vious TRRR and UV-visible transient absorption studies.
At 1094 cm , the collisional model ͓Fig. 7͑A͔͒ reproduces
the observed delay in optical-density recovery, but the pre-
dicted recovery is slow in comparison with the experiment.
−
1
−
1
2
2
−
1
along the asymmetric-stretch coordinate in ϳ1-ps following
photoexcitation demonstrating that reformation of ground-
state OClO occurs on this time scale in these solvents. This is
in agreement with the previous transient absorption and
TRRR studies which deduced that ground-state reformation
→0
1
2,33
−
1
−
1
10,12,24,25
At 1040 and 1014 cm , the predicted optical-density evolu-
occurs on the subpicosecond time scale.
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-9
Solvent-dependent vibrational relaxation of OClO
J. Chem. Phys. 123, 084503 ͑2005͒
Comparison of the vibrational-relaxation dynamics ob-
served in H O and D O provides a good test of the theoret-
ical models of OClO vibrational relaxation given the limited
structural differences between these solvents yet substantial
differences in frequency associated with solvent degrees of
freedom. The success of the perturbative model in reproduc-
and the appearance of excess vibrational energy along the
asymmetric-stretch coordinate. Even when probe frequencies
resonant with the highest-energy levels along the
asymmetric-stretch coordinate are employed, the appearance
of vibrational energy is delayed 10-ps relative to photoexci-
tation. This observation stands in contrast to H O and D O
2
2
2
2
−
1
ing the experimental optical-density evolution in H O and
where at the lowest probe frequencies employed ͑1014 cm ͒
that provide a measure of the population dynamics for inter-
mediate vibrational-energy levels, optical-density increases
consistent with vibrational-energy deposition are observed in
ϳ1-ps following photoexcitation. This observation suggests
that there is a delay in population of the ground state in
acetonitrile relative to H O and D O. At this juncture, it is
2
D O demonstrates that frequency dependence of the solvent-
2
solute coupling is an important factor in defining the
vibrational-relaxation rates of OClO in aqueous solution. The
MD studies predict that the vibrational-relaxation rate in
D O is reduced relative to H O by a factor-of 1.3 due to
2
2
modification of the solvent-solute force-force correlation
function in the frequency region of the asymmetric-stretch
2
2
useful to revisit the proposed sequence of events that occur
3
3
fundamental transistion. Consistent with this expectation,
an approximately threefold reduction in rate in D O relative
following OClO photoexcitation. Photoexcitation results in
2
2
population of the A state which undergoes internal conver-
2
2
to H O is observed. Although qualitative agreement with
2
sion to the A excited state, and this state undergoes subse-
quent internal conversion to the lower-energy B excited
1
2
theory is observed, quantitative agreement is obtained only if
the relaxation rates in both solvents are increased by approxi-
mately fivefold relative to the theoretical values. This dis-
crepancy suggests that the intermolecular potentials used in
the MD simulations for OClO in aqueous solution underes-
timate the magnitude of the OClO-solvent coupling. This
discrepancy was observed in earlier MD studies in which
quantitative reproduction of the vibrational-relaxation kinet-
ics was achieved by increasing the charges on the Cl and O
atoms to values that were substantially larger than those sup-
ported by the dipole moment of OClO. The increase in
charge has the effect of enhancing the magnitude of the elec-
trostatic coupling between solvent and OClO thereby in-
2
state. Therefore, the delay in the appearance of vibrational
energy along the asymmetric-stretch coordinate in acetoni-
trile could reflect a reduced rate of internal conversion to the
2
ground state from the B state. Another possibility is that
2
upon reformation of ground-state OClO by geminate recom-
bination, vibrational energy is initially deposited into a local
mode of OClO, presumably a mode of substantial Cl–O
stretch chararcter, and that the ϳ10-ps appearance time for
energy along the asymmetric-stretch coordinate reflects
intramolecular-vibrational reorganization. This would imply
that energy localization only occurs at the very top of the
asymmetric-stretch manifold, an idea that is consistent with
3
3
3
8,39
creasing the relaxation rate.
mode-bifurcation theory.
However, the observation that
A second difference in the photochemical reaction dy-
namics observed in H O and D O is the reduction in the
vibrational levels up to n=12 are populated in acetonitrile
2
2
demonstrates that only a modest amount of energy
geminate-recombination quantum yield in D O. Previous
−1
2
͑ϳ3000 cm ͒ can be lost from the local mode to the solvent
studies of geminate recombination for other solutes dissolved
in H O and D O have not observed a difference in the re-
in the first 10-ps following geminate recombination.
2
2
A second unique aspect of the relaxation dynamics ob-
served in acetonitrile is the inability of the perturbative
model to adequately reproduce the level dependence of the
vibrational-relaxation rate constants. Given the success of
combination quantum yield. For example, visible pump
−
probe studies of O3 found no variation in the geminate-
3
7
recombination yield and dynamics between H O and D O.
2
2
One explanation for the reduction in the geminate-
this model in describing the relaxation dynamics in H O and
2
recombination quantum yield in D O is that the reduced
2
D O, this result is quite surprising. One reason for this dis-
2
vibrational-relaxation rate in this solvent provides for a cor-
responding increase in the rate of dissociation. Benjamin and
crepancy may be the substantial delay in internal conversion
to the ground state. To explore this possibility, the perturba-
tive model was modified to include a 10-ps delay before
population of the n=15 state. Even with this delay, the
vibrational-relaxation rate constants predicted by the pertur-
bation model were incapable of reproducing the observed
evolution in optical density in acetonitrile. Therefore, the
˜
co-workers have suggested that if the dissociation occurs
2
over an energy barrier along the optically prepared A sur-
2
face, then excess vibrational energy should promote
3
2
dissociation. Similar reasoning can be applied to dissocia-
tion on the ground 2B1 surface. Specifically, if initially
formed ground-state OClO is capable of proceeding along
the reaction coordinate towards dissociation to form ClO and
O, then excess vibrational energy could promote dissocia-
tion. In summary, the reduction of the geminate-
recombination quantum yield in D O relative to H O may
values of F ͑␻͒ ͓Eq. ͑3͔͒ were determined by fitting the
q,q
Ј
experimental evolution, and comparison of the experimental
and predicted evolution in optical density obtained from this
procedure is shown in Fig. 8. Figure 9 presents a comparison
˜
2
2
of the fit values for F ͑␻͒ to those obtained from MD. This
simply reflect the variation in the vibrational-relaxation rate
between these solvents.
q,qЈ
comparison demonstrates that for higher-energy vibrational
level ͑nϾ9͒, the MD results overestimate the relaxation
rates while the relaxation rates for mid- and lower-energy
levels ͑n=2 through n=7͒ are underestimated. Interestingly,
˜
In comparison with H O and D O, the geminate-
2
2
recombination and vibrational-relaxation dynamics of OClO
in acetonitrile are significantly more complicated. Most no-
tably, there is a substantial delay between photoexcitation
the frequency range over which the MD values for F ͑␻͒
q,q
Ј
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-10
Bolinger, Bixby, and Reid
J. Chem. Phys. 123, 084503 ͑2005͒
the asymmetric-stretch coordinate of OClO to the methyl
rock of acetonitrile. The fundamental transition of methyl
rock has a significant IR absorption cross section, and time-
resolved Raman studies have placed the vibrational lifetime
4
0
of the n=1 state of the methyl rock at Ͼ100 ps. Both the
significant absorption of the methyl rock fundamental transi-
tion as well as the literature of the n=1 state are consistent
with assignment of the long-time optical density increase to
the methyl rock.
V. CONCLUSIONS
We have performed a series of TRIR absorption studies
of OClO dissolved in H O, D O, and acetonitrile. Following
2
2
photoexcitation, the evolution in optical density was moni-
−
1
tored at frequencies ranging from 1094 to 956 cm to inves-
tigate the solvent dependence of geminate-recombination
and vibrational-relaxation dynamics involving the
asymmetric-stretch coordinate. In both H O and D O, popu-
2
2
lation of mid-lying energy levels ͑n=8͒ along the
asymmetric-stretch coordinate in ϳ1-ps following photoex-
citation demonstrating that geminate recombination occurs
on this time scale in aqueous solution. The level-dependent
vibrational-relaxation rate constants were approximately
FIG. 8. Comparison of the experimental evolution in optical density of
OClO dissolved in acetonitrile ͑circles͒ to that determined by fitting the rate
constants for the vibrational relaxation dynamics along the asymmetric
stretch ͑solid line͒. Comparisons are presented for probe wavelengths of
threefold reduced in D O relative to H O in qualitative
2
2
1
094, 1040, 1014, and 969 cm−1
.
agreement with the predictions of MD. In both H O and
2
D O, the vibrational-relaxation kinetics was well modeled by
2
considering the solvent-solute force-force correlation func-
tion derived from MD. The geminate-recombination quan-
are less than those determined by fitting correspond to the
frequency region where the acetonitrile methyl rock is local-
ized. This observation suggests that the MD simulations may
underestimate the coupling between the asymmetric stretch
of OClO and this solvent mode. This observation suggests
that quantum-mechanical effects not included in the classical
MD treatment are important in describing the relaxation dy-
namics of OClO in acetonitrile.
tum yield is reduced in D O relative to H O, potentially
2
2
reflecting the slower vibrational relaxation dynamics in D O
2
leading to an increased rate of dissociation. In acetonitrile,
the appearance of ground-state population along the
asymmetric-stretch coordinate is delayed by ϳ10-ps relative
to photoexcitation. This delay was assigned to a longer lived
excited state of OClO in acetonitrile relative to H O and
One final aspect of the optical-density evolution in ac-
etonitrile is the long-time offset in optical density observed
2
D O. In contrast to H O and D O, the vibrational-relaxation
2
2
2
−
1
rate constants predicted by MD did not reproduce the experi-
mental evolution in optical density. Modification of the
frequency-dependent solvent-solute coupling predicted by
MD was required to reproduce the vibrational-relaxation dy-
namics. This modification suggests that quantum-mechanical
effects not included in MD are important in describing the
relaxation dynamics of OClO in acetonitrile.
for the probe frequencies between 1040 and 1014 cm . We
attribute this offset to the transfer of vibrational energy from
ACKNOWLEDGMENTS
The National Science Foundation is acknowledged for
their support of this work ͑CHE-0350191͒. Acknowledgment
is also made to the donors of the Petroleum Research Fund,
administered by the American Chemical Society. One of the
authors ͑I.C.B.͒ gratefully acknowledges support from the
Rabinovich Fund administered by the Department of Chem-
istry at the University of Washington.
1
S. C. Hayes, C. C. Cooksey, P. M. Wallace, and P. J. Reid, J. Phys. Chem.
1
05, 9819 ͑2001͒.
˜
2
3
4
5
FIG. 9. Comparison of the values for F ͑␻͒. The solid line represents the
q,qЈ
F. S. Rowland, Annu. Rev. Phys. Chem. 42, 731 ͑1991͒.
V. Vaida and J. D. Simon, Science 268, 1443 ͑1995͒.
E. C. Richard and V. Vaida, J. Chem. Phys. 94, 163 ͑1991͒.
H. F. Davis and Y. T. Lee, J. Chem. Phys. 105, 8142 ͑1996͒.
values obtained from the MD study while the dashed line represents the
values obtained from the manual adjustment of vibrational relaxation rates
to obtain the best fit. The values are shown relative to F ͑1094 cm−1͒.
˜
q,qЈ
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41

084503-11
Solvent-dependent vibrational relaxation of OClO
J. Chem. Phys. 123, 084503 ͑2005͒
6
23
E. Bishenden and D. J. Donaldson, J. Chem. Phys. 101, 9565 ͑1994͒.
V. Vaida, G. J. Frost, L. A. Brown, R. Naaman, and Y. Hurwitz, Ber.
Bunsenges. Phys. Chem. 99, 371 ͑1995͒.
H. Fidder, F. Tschirschwitz, O. Duhr, and E. T. J. Nibbering, J. Chem.
Phys. 114, 6781 ͑2001͒.
4
7
8
9
0
2
C. C. Cooksey and P. J. Reid, Photochem. Photobiol. 80, 386 ͑2004͒.
S. C. Hayes, M. J. Philpott, and P. J. Reid, J. Phys. Chem. A 103, 5534
͑1999͒.
25
26
27
W. G. Lawrence, K. C. Clemitshaw, and V. A. Apkarian, J. Geophys. Res.
9
5, 18591 ͑1990͒.
C. L. Thomsen, M. P. Philpott, S. C. Hayes, and P. J. Reid, J. Chem.
S. Gnanakaran and R. M. Hochstrasser, J. Chem. Phys. 105, 3486
͑1996͒.
R. Rey and J. T. Hynes, J. Chem. Phys. 108, 142 ͑1998͒.
P. Hamm, M. Lim, and R. M. Hochstrasser, J. Chem. Phys. 107, 10523
͑1997͒.
D. A. V. Kliner, J. C. Alfano, and P. F. Barbara, J. Chem. Phys. 98, 5375
͑1993͒.
M. Lim, S. Gnanakaran, and R. M. Hochstrasser, J. Chem. Phys. 106,
3485 ͑1997͒.
J. Poulsen, T. M. Nymand, and S. R. Keiding, Chem. Phys. Lett. 343,
581 ͑2001͒.
I. Chorny, J. Vieceli, and I. Benjamin, J. Chem. Phys. 116, 8930 ͑2002͒.
I. Chorny, J. Vieceli, and I. Benjamin, J. Chem. Phys. 116, 8904 ͑2002͒.
J. C. Bolinger, S. C. Hayes, T. J. Bixby, and P. J. Reid, J. Chem. Phys.
100, 4795 ͑2004͒.
L. Landau and E. Teller, Phys. Z. Sowjetunion 10, 34 ͑1936͒.
D. J. Nesbitt and J. T. Hynes, J. Chem. Phys. 77, 2130 ͑1982͒.
P. K. Walhout, C. Silva, and P. F. Barbara, J. Phys. Chem. 100, 5188
͑1996͒.
Z. Lu and M. E. Kellman, J. Chem. Phys. 107, 1 ͑1997͒.
G. M. Schmid, S. L. Coy, R. W. Field, and R. J. Silbey, J. Chem. Phys.
101, 2869 ͑1994͒.
Phys. 112, 505 ͑2000͒.
1
J. A. Poulsen, C. L. Thomsen, S. R. Keiding, and J. Thogersen, J. Chem.
Phys. 108, 8461 ͑1998͒.
1
2
8
1
J. Thogersen, C. L. Thomsen, J. A. Poulsen, and S. R. Keiding, J. Phys.
Chem. A 102, 4186 ͑1998͒.
M. J. Philpott, S. C. Hayes, and P. J. Reid, Chem. Phys. 236, 207 ͑1998͒.
Y. J. Chang and J. D. Simon, J. Phys. Chem. 100, 6406 ͑1996͒.
R. C. Dunn, B. N. Flanders, and J. D. Simon, J. Phys. Chem. 99, 7360
2
9
1
1
1
2
3
4
30
3
1
͑
1995͒.
1
1
5
6
C. J. Pursell, J. Conyers, and C. Denison, J. Phys. Chem. 100, 15450
1996͒.
C. J. Pursell, J. Conyers, P. Alapat, and R. Parveen, J. Phys. Chem. 99,
0433 ͑1995͒.
3
2
͑
33
3
4
1
1
1
1
7
8
9
H. S. P. Mueller and H. Willner, J. Phys. Chem. 97, 10589 ͑1993͒.
A. Arkell and I. Schwager, J. Am. Chem. Soc. 89, 5999 ͑1967͒.
C.-P. Liu, L.-H. Lai, Y.-Y. Lee, S.-C. Hung, and Y.-P. Lee, J. Chem. Phys.
35
36
3
7
1
09, 978 ͑1998͒.
2
2
2
0
1
2
C.-P. Liu, L.-H. Lai, Y.-Y. Lee, S.-C. Hung, and Y.-P. Lee, J. Chem. Phys.
3
8
1
09, 978 ͑1998͒.
39
40
J. Thogersen, P. U. Jepsen, C. L. Thomsen, J. A. Poulsen, J. R. Byberg,
and S. R. Keiding, J. Phys. Chem. A 101, 3317 ͑1997͒.
M. J. Philpott, S. Charalambous, and P. J. Reid, Chem. Phys. Lett. 281, 1
J. C. Deak, L. K. Iwaki, and D. D. Dlott, J. Phys. Chem. A 102, 8193
͑1998͒.
͑1997͒.
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: 193.0.65.67
On: Tue, 25 Nov 2014 08:58:41