Picosecond kinetic study of the dynamics for photoinduced homolysis and heterolysis in diphenylmethyl chloride

Article abstract of DOI:10.1021/jp952672v

The kinetics of both the ions and radicals formed upon photolysis of diphenylmethyl chloride, (4-methoxyphenyl)phenylmethyl chloride, and bis(4-methoxyphenyl)methyl chloride in acetonitrile and propionitrile are examined by picosecond pump - probe spectroscopy. Both radical pairs and ion pairs are formed directly from a common excited state. In addition, the geminate radical pair decays by electron transfer to form either the contact ion pair or a covalent bond, as well as undergoes diffusional separation to free radicals.

Full text of DOI:10.1021/jp952672v

3580
J. Phys. Chem. 1996, 100, 3580-3586
Picosecond Kinetic Study of the Dynamics for Photoinduced Homolysis and Heterolysis in
Diphenylmethyl Chloride
Matthew Lipson, Ashok A. Deniz, and Kevin S. Peters*
Department of Chemistry and Biochemistry, UniVersity of Colorado, Boulder, Colorado 80309-0215
ReceiVed: September 12, 1995X
The kinetics of both the ions and radicals formed upon photolysis of diphenylmethyl chloride, (4-
methoxyphenyl)phenylmethyl chloride, and bis(4-methoxyphenyl)methyl chloride in acetonitrile and propi-
onitrile are examined by picosecond pump-probe spectroscopy. Both radical pairs and ion pairs are formed
directly from a common excited state. In addition, the geminate radical pair decays by electron transfer to
form either the contact ion pair or a covalent bond, as well as undergoes diffusional separation to free radicals.
Introduction
Photoexcitation of many aromatic alkyl halides in polar
solvents yields products due to the formation of both ion pairs
and radical pairs, Figure 1.1-3 A central question raised by these
observations is, do both of these reaction pathways proceed
directly from a common excited state? Since the ion pairs are
energetically more stable than the radical pairs in polar solvents,
does electron transfer occur within the geminate radical pair to
form the contact ion pair? In this work, we report the results
of picosecond pump-probe spectroscopy experiments that
follow the kinetics of both the radicals and ions formed and
thereby address these issues.
We have previously reported picosecond pump-probe studies
of the ion pairs formed upon photolysis of diphenylmethyl
chloride (DPMC), Figure 2, in acetonitrile and propionitrile.4,5
The decay of the absorption of the diphenylmethyl cation, on
the sub-nanosecond time scale, is modeled well by a four state
Figure 1. Scheme of the photocreation of ion pairs and radical pairs
from alkyl chlorides in polar solvents. In polar solvents, the ion pair
is lower in energy than the radical pair.
model: the DPMC molecule, a contact ion pair (CIP), a solvent
separated ion pair (SSIP), and free ions (FI), where the CIP is
formed within the duration of the laser pulse.
SCHEME 1
With this model in place, we used an Arrhenius analysis to
measure the activation barriers that separate these states and
concluded that the conversion of CIP to DPMC occurs under
conditions of solvent polarization caging, in which the rate of
reaction is controlled by the rate of solvent reorganization.5 We
herein extend our previous studies of ion pair dynamics to
Figure 2. Diarylmethyl chlorides used in this study.
include two methoxy substituted derivatives of DPMC, Figure
2, as well as investigate the dynamics of the geminate radical
absorptions of both the cations and radicals formed have been
pairs.
fully characterized, simplifying our experiment.7 Second,
product quantum yields for both the parent compound and a
number of photolabile systems have been studied which appear
Although we limit these studies to diarylmethyl chlorides, a
number of its derivatives have been reported in the literature.7
From our previous work, we found the photolysis of DPMC
in acetonitrile yields ion pairs at a rate greater than the =20 ps
response time of our instrument.4,5 Since in acetonitrile the
radical pair formed is higher in energy (64 kcal/mol) than the
ion pair (11 kcal/mol), it is possible that the radical pair is
formed more rapidly than the ion pair and that the ion pair is
then formed via electron transfer within the radical pair, Figure
1.7,8 Such a reaction pathway has been proposed by Pincock
and co-workers for benzyl acetates and 1-naphthylmethyl
esters.3,9,10
to yield both ion pairs and radical pairs.1-3,6 Much of the
literature has concentrated on aryl and diarylmethyl systems,
but substrates in which the chromophore is separated from the
leaving group by one or more saturated carbons have also been
studied. As leaving groups, bromides, acetates, phenoxides,
hydroxides, and trimethylammonia salts show qualitatively
equivalent photochemical behavior.
We have chosen to concentrate our studies on DPMC and
its substituted derivatives for two reasons. One, the UV-vis
X Abstract published in AdVance ACS Abstracts, February 1, 1996.
0022-3654/96/20100-3580$12.00/0 © 1996 American Chemical Society

Homolysis and Heterolysis in DMPC
J. Phys. Chem., Vol. 100, No. 9, 1996 3581
Steenken and co-workers used a combination of nanosecond-
microsecond laser flash photolysis and product studies to
measure the quantum yields of ion and radical pair production
for DPMC and a number of substituted derivatives.7 They report
that photolysis of DPMC in acetonitrile leads to ion pairs and
radical pairs in comparable yields, 13% ion and 23% radical.
For MeO, those quantum yields are 0.32 and 0.26, and for
DiMeO, 0.31 and 0.24. We will attempt to explain these ratios
with this work, but, of equal importance, we wish to understand
why these quantum yields do not add to 1. It is unlikely that
such deactivation pathways as fluorescence or intersystem
crossing to the molecule’s triplet manifold or internal conversion
to ground state could compete efficiently with such rapid
dissociation pathways; therefore, the most reasonable explana-
tion for the low quantum yields that Steenken and co-workers
measure in the photolysis of DPMC at nanosecond time scales
is efficient cage recombination within the initially formed ion
pair and within the initially formed radical pair on the
picosecond time scale. The present study directly monitors the
dynamics of cage recombination for ion pairs and radical pairs.
Further, although extensive literature documents cage recom-
bination reactions within carbon-centered radical pairs, the
geminate reaction, in solution, between a carbon-centered radical
and a halogen atom has yet to be studied at ambient tempera-
tures.11,12
Figure 3. Transient absorption of the diarylmethyl radicals formed
upon photolysis of the corresponding diarylmethyl chlorides in aceto-
nitrile at 23 °C: laser excitation at 266 nm; radicals monitored at 330
nm; A ) DPMC, B ) MeO, and C ) DiMeO.
Experimental Section
Figure 4. Transient absorption of the diarylmethyl radical formed upon
photolysis of DiMeO in acetonitrile at 9.3 and 47.4 °C: laser excitation
at 266 nm; radicals monitored at 330 nm.
Diphenylmethyl chloride (DPMC) and all substituted deriva-
tives were produced from the corresponding alcohols, (Aldrich)
with thionyl chloride and purified by vacuum distillation or
sublimation to >95% purity as measured by NMR, IR, and gas
chromatography (HP5890, FID, DB17 column). Acetonitrile,
(Burdick and Jackson, UV grade) was distilled from calcium
hydride. Tetrahydrofuran (Aldrich) was distilled from sodium.
Cyclohexane, (Fisher Scientific, ACS Grade) was used as
received. Propionitrile (Aldrich, 97%) was refluxed over sodium
methoxide prior to fractional distillation. This purity of solvent
proved usable for our experiments on DPMC and MeO, but
DiMeO proved thermally unstable in it.
convolution of the instrument response function I(t) with the
transient signal, F(t).
t
A(t) ) -∞I(τ) F(t-τ) dτ
(1)
∫
The instrument response function, I(t), results from the convolu-
tion of the pump and probe pulse and is assumed to have the
analytical form of a Gaussian
A detailed description of our picosecond pump-probe
experiment has been reported elsewhere.13 The laser is a
Continuum (PY61C-10) Nd:YAG (=30 ps fwhm). All samples
were pumped at 266 nm (=200 µJ focused to 2 mm at the
sample). Notch filters were used to select the probe light
frequency out of a white light continuum generated by focusing
either 532 nm (good down to =350 nm) or 355 nm (good down
to =300 nm) laser pulses to a point in D2O/H2O. Probe intensity
is measured by Oriel 71902 UV-enhanced photodiodes inter-
faced to a Stanford (SRS 250) boxcar integrator.
Samples were prepared so as to exhibit ground state optical
density of 1.5 at 266 nm and were stirred in a 1 cm quartz
cuvette throughout the experiment; flowing the samples through
the cuvette did not change the experimental results. Laser pump
power was in the region where signal intensity increased linearly
with laser power and kinetics showed no sensitivity to laser
power. The experiment was not sensitive to the polarization
of the pump beam relative to that of the probe.
The cations formed by photolysis of DPMC and MeO were
probed at 440 nm. The cation formed by photolysis of DiMeO
was probed at 480 nm. All radicals were probed at 330 nm.
For ions, typically 200 data points were collected per run. Each
point is the average of 25 laser shots, and the decay between
points is 5-15 ps. For radicals, 100 data points were collected
per run. All reported decays are the average of 3-6 runs. Traces
shown in this report are point grouped for display purposes.
The method of deconvolution of the kinetic data has been
presented.13 The observed transient signal A(t) results from the
I(t) ) (2πσ)-0.5 exp(-(t - t0)2/2σ2)
(2)
where σ is the width and t0 the position of the peak of the
Gaussian.
To measure the response time of our instrument, we use the
instantaneous excited state absorption of pyrene at 440 and 480
nm and of naphthalene at 330 nm. For naphthalene, it is
necessary to subtract fluorescence from the data before optical
density is calculated. The response function of our instrument
at both wavelengths of interest is 20 ( 3 ps, as measured by
repeated experiments.
Results
The Radicals. Acetonitrile solutions of DPMC, MeO, and
DiMeO at 23 °C were irradiated at 266 nm, and the dynamics
of the resulting transient species were monitored at 330 nm.
Figure 3 shows the normalized transient absorption decays we
assign, in agreement with literature reports, to the diarylmethyl
radicals formed upon photolysis of the corresponding diaryl-
methyl chlorides in acetonitrile.7 Figure 4 shows the decays
of the radical of DiMeO in acetonitrile at two temperatures.
In acetontrile, all radicals decay exponentially to a constant
absorbance by 600 ps. The DPMC radical reaches this plateau
more rapidly than do the methoxy substituted radicals. An
increase in temperature leads to an increase in both the rate of
decay and in the constant absorbance, Figure 4, but the effect
is subtle. In contrast to the behavior of the radicals in

3582 J. Phys. Chem., Vol. 100, No. 9, 1996
Lipson et al.
Figure 5. Transient absorption of the diarylmethyl radical formed upon
photolysis of DiMeO in acetonitrile at 23 °C, fit to the model shown
in Scheme 2: laser excitation at 266 nm; radical monitored at 330 nm;
rate constants listed in Table 1; σ ) 24 ps; t0 ) 153 ps.
acetonitrile, no decay is observed for the radicals formed upon
photolysis of DPMC in either cyclohexane or tetrahydrofuran
(data not shown).
The radical decays are fit to the model depicted in Scheme
2,
SCHEME 2
where RP denotes a geminate radical pair; RP is formed
instantaneously. This model predicts a single exponential decay
(with a rate constant kRP equal to the sum of kd and kesc) to a
constant absorbance given by
Figure 6. Transient absorption of the diphenylmethyl radical formed
upon photolysis of DPMC in acetonitrile at 23 °C, fit to the model
shown in Scheme 2: laser excitation at 266 nm; radicals montitor at
330 nm; A, fit with σ ) 20 ps, kd ) 2.0 × 1010 s-1, kesc ) 1.7 × 1010
s-1, t0 ) 123 ps; B, fit with σ ) 15 ps; rate constants listed in Table
1; t0 ) 118 ps.
kesc
[RP]∞ )
[RP]0
(3)
TABLE 1
kesc + kd
ion kineticsb
radical kineticsb
(×109/s)
(×109/s)
On the time scale of our experiment, radicals that escape their
cage to become free radicals do not return.
temp
(C)
compound solventa
k1
k2
Rc
kd
kesc
Figure 5 shows the fit of the Scheme 2 model to the data for
DiMeO in acetonitrile at 23 °C. Although the residuals to the
fits are, in general, rather good, the values we obtain for the
rate constants are not very precise, (20% (estimated). One
parameter employed in fitting the data is the response function
of our instrument whose value 20 ( 3 ps is measured by fitting
the instantaneous rise of the transeint absorption of naphthalene
at 330 nm. The uncertainty in this response function leads to
uncertainty in deconvoluting the intensity of the signal at early
times and therefore leads to uncertainty in fitting the two rate
constants kd and kesc. We further note that the data for DPMC,
MeO, and DiMeO are only fit acceptably if we set the instrument
response function at the somewhat low value of 15, 13.8, and
24.0 ps, respectively. Figure 6 shows the fits we obtain to the
radical kinetics of DPMC when we fix the response function at
20 and 15 ps. A possible explanation for the failure of the model
depicted in Scheme 2 to fit the experimental data with σ ) 20
ps is that the rate constants associated with processes kd and
kesc are time dependent and not time independent, as assumed.14-16
Presently we do not have the time resolution that would allow
for a detailed analysis of the time dependence of the rate constant
and therefore assume for the present study that kd and kesc are
time independent and their value reflects an “average” value of
the time dependent processes. The data reported in Table 1
are that derived from the fitting procedure which gives the best
residuals which necessitates a variation in σ. As shown in Table
2, the value of the rate constants kesc and kd are not very sensitive
to the value of σ.
DPMC
Acn
10
23
55
23
10
23
55
23
10
23
48
3.1 2.8 0.0
4.1 3.6 0.0
4.5 4.0 0.0
11.6 8.5 0.4
2.4 2.1 0.3
3.5 3.6 0.3
4.5 3.3 0.3
4.1 3.5 0.4
0.8 0.7 0.4
1.7 2.0 0.4
2.4 3.0 0.4
7.1
8.2
6.8
5.6
6.2
7.0
6.2
7.3
4.3
5.1
6.4
11.0
13.0
13.0
16.0
3.9
5.0
5.9
7.5
3.2
4.0
5.2
Prop
Acn
MeO
Prop
Acn
DiMeO
a Acn ) acetonitrile; Prop ) propionitrile. b Uncertainties in the fits
are (20%; the values for k3 and k4, Scheme 3, are not given. c R )
[RP]0/([RP]0 + [CIP]0); where [RP]0 is the concentration of radical
pairs that feed into the CIP.
TABLE 2: Dependence of the Rate Constants from Scheme
2 on the Instrument Response Function, σ, at 23 °C
compound
DPMC
σ
kesc (s-1
)
kd (s-1
)
20.0
15.1
20.0
13.8
1.7 × 1010
1.3 × 1010
5.6 × 109
5.0 × 109
2 × 1010
8 × 109
MeO
9.2 × 109
7.0 × 109
Ions. Acetonitrile solutions of DPMC, MeO, and DiMeO at
23 °C were irradiated at 266 nm, and the dynamics of the
resulting transient species were monitored at 440 nm for DPMC
and MeO and at 480 nm for DiMeO. Figure 7 shows the
normalized transient absorption decays we assign, in agreement
with literature reports, to the diarylmethyl cations formed upon

Homolysis and Heterolysis in DMPC
J. Phys. Chem., Vol. 100, No. 9, 1996 3583
Figure 7. Transient absorption of the diarylmethyl cations formed upon
photolysis of the corresponding diarylmethyl chlorides in acetonitrile
at 23 °C: laser excitation at 266 nm; A (DPMC) and B (MeO)
monitored at 440 nm; C (DiMeO) monitored at 480 nm.
Figure 8. Transient absorption of the diarylmethyl cation formed upon
photolysis of DiMeO in acetonitrile at 9.8 and 47.7 °C: laser excitation
at 266 nm; cation monitored at 480 nm.
Figure 10. Transient absorption of the diarylmethyl cation formed
upon photolysis of DiMeO in acetonitrile at room temperature: A fit
to the model shown in Scheme 1, k1 ) 6.5 × 108 s-1, k2 ) 4.6 × 108
s-1, σ ) 24 ps, t0 ) 72 ps; B fit to the model shown in Scheme 3, rate
constants listed in Table 1, σ ) 24 ps, t0 ) 162 ps; laser excitation at
266 nm; cation monitored at 480 nm.
SCHEME 3
Pincock and co-workers3,10 who proposed that CIP can be
formed from the geminate radical pair through electron transfer,
we expand the model to Scheme 3, where the RP (geminate
radical pair) has no transient absorption at the absorption
wavelength of the cation, the initial concentration of transients
is split between the CIP and the RP, and the RP feeds into the
CIP irreversibly at a rate kRP[RP]. kRP is fixed to the value kRP
) kesc + kd derived from the corresponding kinetics for the
radicals monitored at 330 nm. We vary the ratio
Figure 9. Transient absorption of the diphenylmethyl cation formed
upon photolysis of DPMC in acetonitrile at 23 °C, fit to the model
shown in Scheme 1: laser excitation at 266 nm; cation monitored at
440 nm; rate constants listed in Table 1; σ ) 24 ps; t0 ) 72 ps.
[RP]0
photolysis of the corresponding diarylmethyl chlorides in
acetonitrile7 Figure 8 shows the decay of the cation of DiMeO
in acetonitrile at 9.8 and 47.7 °C.
R )
[RP]0 + [CIP]0
Of note, the absorption of DPMC appears to decay more
rapidly than that of MeO which decays more rapidly than that
of DiMeO. Further, the actual decay appears to begin more
slowly for DiMeO than for MeO, which is slower than for
DPMC. This “delay” in the decay is seen strikingly for DiMeO
at low temperatures and less so at higher tmperatures, Figure
8. All transient absorptions decay to a constant absorbance that
is sensitive to temperature.
to find the best fit for the data, Figure 10B. Note that [RP]0
here is not the total number of radical pairs formed. Instead it
is the total number of radical pairs that decay into the CIP, and
[CIP]0 is the initial amount of CIP formed directly from the
excited singlet state.
Table 1 lists the kinetic parameters which yield the best fits
to the data. For the case of DPMC at 23 °C, changing the initial
concentration of RP relative to CIP has only a small effect on
the quality of the fits. Changing R from 0 to 0.9 increases the
residuals in the fit by 9%. That same change in R, however,
changes k1 from 4.1 × 109 to 3.7 × 1010 s-1. For DPMC, we
find the lowest residuals when we set R ) 0.
We previously reported4,5 that the kinetics of the ion pair
formed upon photolysis of DPMC in acetonitrile and propioni-
trile can be modeled well, Figure 9, by Scheme 1. In this model,
CIP, SSIP, and FI have a common extinction coefficient and
all ion pairs are formed instantaneously as CIPs. However, this
model fails to reproduce the kinetics of the ion pairs formed
from MeO and DiMeO, Figure 10A. In light of the studies of
The quality of the fits for DiMeO and MeO is a good deal
more sensitive to R. For the case of DiMeO at 23 °C, an
increase in R from 0 to 0.4 decreases the residuals in the fit by

3584 J. Phys. Chem., Vol. 100, No. 9, 1996
Lipson et al.
a factor greater than 3. A further increase in R from 0.4 to 0.9
increases the residuals by 8%. For DiMeO and MeO, we find
the best fit to the data at R = 0.4 and 0.3, respectively.
Also listed in Table 1 are the kinetic parameters we measure
for the behavior of DPMC and MeO in propionitrile.
the CIP irreversibly at a rate kRP[RP]. kRP is measured from
the corresponding radical decays.
The kinetics predicted by Scheme 1 are biexponential. By
expanding Scheme 1, we are adding a new adjustable parameter,
namely, the initial ratio of [RP]0/([RP]0 + [CIP]0) which we
have labeled R. However, since Winstein and co-workers first
proposed19 the basic model of an equilibrium between a CIP
and an SSIP in 1954, numerous experimental and theoretical
studies have supported it for both ion pairs and radical-ion
pairs.4,13,18,20,21 We therefore feel it reasonable that the basic
model of Scheme 1 should apply here as well as for the analysis
of the MeO and DiMeO data.
Discussion
Geminate Radical Pair Dynamics. The model in Scheme
2 we propose to rationalize the kinetic behavior of the radicals
formed upon photoinduced homolysis of diarylmethyl chlorides
in acetonitrile would appear to be oversimplified: photolysis
generates a solvent caged singlet radical pair which may either
escape from the cage or collapse in a time independent fashion.
A typical rate for radical pair diffusional escape from a solvent
cage is of the order of 5 × 109 s-1, in reasonable agreement
with the values of kesc we report in Table 1.17 The escape
process we observe, however, is not likely to be a simple
diffusional separation of two independent particles. If it were,
one would predict that the rate of this process would be
controlled by the rate of diffusion of the smaller radical, in our
case the chlorine atom, and thus should be the same for the
three molecules. It is clear, though, in Table 1, that kesc is
dependent upon substituents on the phenyl rings of the aryl
radical as kesc increases from 4 × 109 s-1 for DiMeO to 1.3 ×
1010 s-1 DMPC. Further, if this escape process were purely
diffusional, it should slow as solvent viscosity is increased. As
shown in Table 1, kesc (for DPMC and MeO) increases from
acetonitrile to propionitrile even though viscosity increases as
well.18 It would appear the barrier to geminate radical separation
depends not only on forces associated with solvent caging but
also upon the electronic interaction between the two radicals,
as evidenced by the effect of substituents upon the rate of this
process.
Another aspect of this model for a caged radical pair is
perplexing. Although singlet radical pairs normally couple
readily,11 such does not appear to be the case for carbon-
chlorine radical pairs. We observe no decay in the absorbance
of diarylmethyl radicals in cyclohexane or tetrahydrofuran. The
obvious property difference between these two solvents and
acetonitrile is that cyclohexane and tetrahydrofuran have mark-
edly lower dielectric constants. However, the rates of radical
coupling reactions are normally not strongly sensitive to solvent
polarity.11 We therefore propose that direct adibatic coupling
of chlorine atoms with diarylmethyl radicals does not occur.
At this stage one can only speculate as to the nature of the
electronic process or processes associated with kd. As will be
discussed, direct coupling of the two radicals to form the
carbon-chlorine bond must be considered a nonadibatic process
as the potential energy surface associated with ground state
dissociation correlates with the ion pair state in polar solvents
and it is the excited state surface that correlates with the radical
pair. An alternative pathway for the decay of the radical pair
is through an electron transfer process, as has been proposed
by Pincock and co-workers.3,10 If a charge transfer process is
involved in the decay of the radical pair, the rate of this process
should depend upon the polarity of the solvent, as is observed.
Ion Pair Dynamics. The kinetic behavior of the ions formed
upon photolysis of DPMC in acetonitrile can be an acceptably
fit by the model shown in Scheme 1 in which all of the CIP are
formed instantaneously. The CIP can either collapse to starting
material or diffusionally separate into the SSIP. The SSIP can
either return to the CIP or escape to FI.
For both MeO, the best fit to the data is when R is set to
=0.3, while, for DPMC, the best fit is when R ) 0. The value
of R ) 0.3 for MeO would suggest that partitioning between
the reaction pathways leading to the formation of the CIP
corresponds to 70% coming directly from the excited singlet
state and 30% coming from the radical pair via an electron
transfer process. It is surprising that the best fit of Scheme 3
to DPMC kinetics occurs for R ) 0, suggesting that all of the
CIP comes directly from the excited singlet state. If all the RP
that decays through the process kd were to form CIP, the value
of R would be 0.20 for DPMC. However, given the very small
variation in the residuals with the change in R, electron transfer
within the DPMC radical pair to produce the CIP may indeed
occur and our experiment is not able to resolve this process,
although the experiment should be sensitive to R values of 0.2-
0.3. One reason the process is observed in MeO and DiMeO
is that the geminate radical pair lifetimes (1/kRP) are 83 and
110 ps, and it is during this interval that electron transfer occurs
to produce 30% of the CIP. The DPMC geminate radical pair,
however, has a lifetime of only 47 ps, which may be too short
to allow observation of the electron transfer in the formation of
the DPMC CIP, particularly if the amount of the CIP derived
from the geminate pair is small, R < 0.1.
As a brief summary, we find that both radical pairs and ion
pairs are created directly from a common excited state. Radical
pairs are not simply solvent caged, as the rate of cage escape,
kesc, depends upon the electronic structure of the radical species.
Furthermore, radical pairs do not adiabatically couple to form
the carbon-chlorine bond but, for MeO and DiMeO, decay by
electron transfer to form CIP.
Quantum Yields for Initial Production of CIP and
Geminate Radical Pairs. The quantum yields for the produc-
tion of radicals and ions, measured at 70 ns, for DPMC, MeO,
and DiMeO in acetonitrile have been measured.7 For the series
DPMC, MeO, and DiMeO the radical quantum yields are 0.23,
0.26, and 0.24, while the ion quantum yields are 0.13, 0.32,
and 0.31. Assuming these transient species do not decay on
the 1-70 ns time scale, it is then possible to derived the initial
amounts of radicals and ions formed upon photolysis given the
kinetic data in Table 1. For the radicals derived from DPMC,
MeO, and DiMeO, the initial amount of geminate radical pair
is 0.37, 0.62, and 0.54, respectively.
The determination of the initial amount of CIP, derived from
both the excited singlet state and from the radical pair, is
somewhat more complicated as the ions do decay on the
nanosecond time scale as a result of the processes associated
with k3 and k4, Scheme 3. However, the amplitude for decay
on this time scale is very small relative to the decay of the ions
on the picosecond time scale, k1 and k2, and thus can be
neglected. With this assumption, the initial amount of CIP is
0.28, 0.62, and 0.57 for the three ion species. To determine
the initial amount of CIP that is produced directly from the
excited singlet state, the contribution derived from the radical
This model, though, fails to reproduce the dynamical behavior
of the ions formed upon photolysis of MeO or DiMeO. Their
behavior is only well-reproduced if the laser pulse initially
creates two species, the CIP and the RP, and the RP feeds into

Homolysis and Heterolysis in DMPC
J. Phys. Chem., Vol. 100, No. 9, 1996 3585
transition state for bond cleavage of DPMC is not known. If â
is of the order of 25 kcal/mol this would lead to a local minimum
on the excited singlet state surface below the corresponding to
homolytic dissociation limit.23 If â is substantially greater, such
a minimum on the S1 surface should not occur.
It is necessary to point out that the free energy diagram for
the excited singlet state is only qualitative. In particular, it is
not known whether the initial portion of the curve to which
light induced excitation takes place corresponds to a repulsive
potential; the surface could certainly be more complex, perhaps
possessing a small electronic barrier along the dissociation
curves whose magnitude is solvent dependent.
Figure 11. Potential of mean force for the solvent equilibrated surfaces
for diphenylmethyl chloride in acetonitrile. RP may either escape or
cross to the ionic surface, from where it may either return to starting
material or form CIP.
It is further important to emphasize that these free energy
surfaces correspond to the solvent equilibrated reaction path-
ways.23 However, for reaction processes that occur on the time
scale of the solvent reorganization, these surfaces will be greatly
modified. The question of nonequilibrium solvation of the
ground state surface has been explicitly addressed by Hynes.22
In the Hynes solvation model, the solute electronic structure
instantaneously feels the solvent electronic polarization, Pel, but
the solvent orientational polarization, Por, may be out of
equilibrium with the solute electronic structure. In the tert-
butyl chloride study, if the solvent is not allowed to structurally
reorganize about the developing ion pair upon heterolytic
dissociation, the ion pair energy is destablized by 55 kcal/mol.22
For diphenylmethyl chloride, this would correspond to having
the CIP higher in energy that the radical pair. Thus, it is only
the reorganization of the solvent structure that allows for the
ion pairs to become more stable that the radical pairs.
pair is subtracted. Thus, the initial amount of CIP formed from
the excited singlet state is 0.28 for DPMC, 0.43 for MeO, and
0.34 for DiMeO.
Combining these results for the initial yields of radical pairs
and ion pairs derived from the excited singlet state, we find
that the total quantum yield for the production of transient
species is 0.65 for DPMC, 1.05 for MeO, and 0.88 for DiMeO.
It is noteworthy that while almost all of the decay of the excited
singlet states for MeO and DiMeO give rise to either radical
pairs or ion pairs, this does not hold for DPMC.
Solvent Dependence of the Potential Energy Surfaces for
Radical Pairs and Ion Pairs. To understand in greater detail
the possible reaction pathways for photoinduced bond homolysis
and heterolysis in the condensed phase, it is informative to
examine the solvent dependence of the free energy surfaces for
these processes. Regrettably little is known, either experimen-
tally or theoretically, about solvent dependence of the electronic
structures associated with radical and ion processes of diphen-
ylmethyl chloride. Some insights into the nature of these
reactions may be gleaned from the theoretical studies of the
homolysis and heterolysis of tert-butyl chloride by Hynes and
co-workers.22-24 Their theoretical approach entails the coupling
of two gas phase diabatic surfaces, one purely covalent and one
purely ionic, to generate adiabatic surfaces through a nonlinear
Schrodinger equation that accounts for the mutual influence of
the solute electronic structure and solvent polarization under
both equilibrium and nonequilibrium conditions. The two
reaction coordinates are the carbon-chloride bond extension,
r, and a solvent coordinate, s, which reflects the change in the
solvation upon the production of ionic species. The potential
of mean force for equilibrium solvation pathway (ESP) is
obtained by invoking that with each incremental extension of
the carbon-chlorine bond the solvent equilibrates, thus satisfy-
ing the condition
Given the above picture of the solvent dependence of the
potential energy surfaces associated with diphenylmethyl chlo-
ride homolysis and heterolysis, we can speculate as to the
mechanism for the production of radical pairs and ion pairs
following photolysis. Prior to irradiation there will be an
equilibrium solvent structure about DPMC which will be
maintained upon photoexcitation of DPMC to S1. Presumably
this initial solvent structure would not support the development
of the CIP but would allow for the formation of the radical pair
whose energetics should show only a minor solvent dependence.
This assertion is based upon the observation that radical pairs
are formed within less than 20 ps in both cyclohexane and THF.
Following excitation to S1, the solvent structure may then begin
to structurally evolve resulting in a lowering of the ionic surface.
The time scale for solvent reorganization could be as fast as
the longitudinal relaxation time of the solvent, 0.2 ps for
acetonitrile.14 Concomitant with the lowering of the ionic
surface is either the development of a conical intersection
between the radical and ion surfaces or the development of an
avoiding crossing leading to a strong electronic coupling
between the two surfaces facilitating a nonadiabatic transition.25
It is during the development of the conical intersection or the
strongly electronically coupled avoided crossing that the excited
state population passes onto the ionic surface, which, upon
further evolution in the solvent structure, allows for the creation
of the CIP.
δG(r,s)/δs ) 0
(4)
The projection of this reaction pathway onto the r coordinate
for the ground state (S0) for adiabatic bond heterolysis should
lead to a ground state potential surface for heterolytic dissocia-
tion.23 On the basis of the tert-butyl chloride study, the
diphenylmethyl chloride potential energy surface for heterolysis
should appear as shown in Figure 11. The energy of the CIP
is estimated to be 11 kcal/mol.7 The barrier to collapse of the
CIP to form the carbon-chlorine bond is 3.0 kcal/mol, a value
obtained in our prior picosecond study of DPMC in acetonitrile.5
The free energy profile for the first excited singlet state, S1,
was not derived by Hynes in these first studies. However, a
qualitative description for S1 may be deduced from the free
energy for bond homolysis, 64 kcal/mol, and the optical
transition of 110 kcal/mol for S0 f S1. The magnitude of the
electronic coupling, â, between the two diabatic surfaces at the
One puzzling aspect of this model, as it is presented, is the
entire S1 population either dissociates to radical pairs or ion
pairs. Such appears to be the case for MeO and DiMeO.
However, for DPMC, 35% of the initial population of S1 does
not appear as either geminate radical pairs or CIP. One possible
explanation is that the fluctuation in the solvent to a solvent
orientation that allows for the surface crossing is not the structure
that supports of development of the CIP; it is only upon further
evolution in the solvent structure that the CIP is formed.
However, it is feasible that once the system is on the ground
state surface, the solvent structure may reorient in such a manner

3586 J. Phys. Chem., Vol. 100, No. 9, 1996
Lipson et al.
as to allow for the development of the carbon chlorine bond,
returning the molecule to its original state. That this latter
pathway does not appear to operate to any great extent in MeO
and DiMeO, given the quantum yields for radical and ion pair
formation, is the result of the difference in the electronic
structures which undoubtedly will effect the position along the
reaction coordinates, in either the r or s coordinates, of either
the conical intersection or the avoiding crossing.
determined by changes in both r and s coordinates. Apparently
the electron transfer within the DPMC geminate radical pair
places the system in such a region on the ground state surface
that the predominate decay channel is toward formation of the
covalent bond, while electron transfer within the MeO and
DiMeO geminate radical pairs places the system in a region on
the ground state surface that the predominate decay channel is
toward CIP. That transitions occur to different regions on the
ground state surface must reflect the different positions in the
excited state local minimum due to differing electronic structures
of the three geminate radical pairs.
An interesting question relates to the possible existence of a
local minimum on the potential energy surface associated with
the radical pair. As depicted in Figure 9, there is an energy
minimum which results from the electronic coupling between
the two diabatic states. However, this picture was developed
from the Hynes model within the context of the solvent fully
equilibrated to the ionic surface. If the reacting species are on
the homolytic surface, the solvent structure equilibrated with
the radical species will be different than that associated with
the ionic species, and thus, it is not clear whether such a
minimum should exist. A local minimum could result if there
is an ionic component to the radical interaction. A proposal
for ionic interactions in radical systems can be found in
Walling’s studies of the thermal decomposition of a large
number of diacyl peroxides where the process yields both radical
derived and ion derived products in a ratio sensitive to solvent
polarity.26 Walling proposed a single transition state for the
decomposition that cannot be described as either radical or ionic
in nature but as some resonance hybrid of the two electronic
structures which then evolves into ion pairs or into radical pairs
upon further separation, a process strongly dependent upon the
nature of the solvent. If such interactions occur within the
present system on the excited state surface, the magnitude of
the interaction will be dependent both upon the electronic
structure of the interacting species and upon the polarity of the
solvent.
Conclusion
We have established that photolysis of DPMC in acetonitrile
leads directly to the formation of both radical pairs and contact
ion pairs. Radical pairs do not directly recombine adiabatically,
but cross onto the ground state ionic surface leading to the
formation of either the contact ion pair or to the formation of
a covalent bond. Future work will concentrate on radical pair
and ion pair formation and decay pathways employing femto-
second laser absorptions spectroscopy.
Acknowledgment. This work is supported by a grant from
the National Science Foundation, CHE 9408354. We thank
Lisha Barre for the synthesis of the compounds employed in
this study.
References and Notes
(1) Cristol, S. J.; Bindel, T. H. In PhotosolVolysis and Attendant
Photoreactions InVolVing Carbocations; Cristol, S. J., Bindel, T. H., Eds.;
Dekker: New York, 1983; Vol. 6, pp 327-415.
(2) Das, P. K. Chem. ReV. 1993, 93, 119-144.
(3) Hilborn, J. W.; MacKnight, E.; Pincock, J. A.; Wedge, P. J. J. Am.
Chem. Soc. 1994, 116, 3337-3346.
(4) Peters, K. S.; Li, B. J. Phys. Chem. 1994, 98, 401-403.
Evidence for such a minimum should be manifested in
substituent effects upon the rate of separation of the radical pairs,
kesc. Indeed the rate of radical separation decreases as methoxy
(5) Deniz, A. A.; Li, B.; Peters, K. S. J. Phys. Chem. 1995, 99, 12209-
12213.
(6) Post, A. J.; Nash, J. J.; Love, D. E.; Jordon, K. D.; Morrison, H. J.
Am. Chem. Soc. 1995, 117, 4930-4935.
groups are added to the phenyl rings; the rate of escape, kesc
,
decreases from 1.3 × 1010 s-1 for DPMC to 4.0 × 109 s- for
DiMeO. Also this separation process should be sensitive to
both the viscosity and the dielectric of the solvent. Based upon
viscosity considerations alone, the parameter kesc should decrease
with an increase in viscosity. However, one finds that kesc
actually increases on changing the solvent from acetonitrile to
the more viscous solvent propionitrile,18 and thus, kesc appears
to be sensitive to the dielectric of the solvent. Therefore, the
increase in kesc must be the result of a decrease in the electronic
interaction within the radical pair with the decrease in polarity,
leading to a lower barrier for radical separation. Unfortunately,
it is not possible to separate the effect that the solvent has upon
the electronic barrier for radical separation from the effect that
the solvent viscosity has upon barrier for diffusional separation.
There is evidence, however, that the two radicals are electroni-
cally coupled, producing a stabilization of the radical pair, and
that this stabilization is sensitive to the electronic structure of
the radicals and to the polarity of the solvent.
(7) Bartl, J.; Steenken, S.; Mayr, H.; McClelland, R. A. J. Am. Chem.
Soc. 1990, 112, 6918-6928.
(8) Wayner, D. D. M.; McPhee, D. J.; Griller, D. J. Am. Chem. Soc.
1988, 110, 132-137.
(9) Arnold, B.; Donald, L.; Jurgens, A.; Pincock, J. A. Can. J. Chem.
1985, 63, 3140-3146.
(10) Pincock, J. A.; Wedge, P. J. J. Org. Chem. 1994, 59, 5587-5595.
(11) Walling, C. Free Radicals in Solution; Wiley: New York, 1957.
(12) Step, E. N.; Buchachenko, A. L.; Turro, N. J. J. Org. Chem. 1992,
57, 7018-7024.
(13) Peters, K. S.; Lee, J. J. Phys. Chem. 1992, 96, 8941-8945.
(14) Simon, J. D. Acc. Chem. Res. 1988, 21, 128-134.
(15) Kahlow, M. A.; Jarzeba, W.; Kang, T. J.; Barbara, P. F. J. Chem.
Phys. 1989, 90, 151-158.
(16) Castner, E. W. J.; Maroncelli, M.; Fleming, G. R. J. Chem. Phys.
1987, 86, 1090-1097.
(17) Noyes, R. M. J. Am. Chem. Soc. 1956, 78, 5486-5490.
(18) Li, B.; Peters, K. S. J. Phys. Chem. 1993, 97, 7648-7651.
(19) Winstein, S.; Clippinger, E.; Fainberg, A. H.; Robinson, G. C. J.
Am. Chem. Soc. 1954, 76, 2597.
(20) Peters, K. S.; Li, B. J. Phys. Chem. 1992, 98, 401-403.
(21) Arnold, B. R.; Noukakis, D.; Farid, S.; Goodman, J. L.; Gould, I.
Finally, there is the question as to why, upon electron transfer,
the geminate radical pairs for MeO and DiMeO predominately
form CIP while the geminate radical pairs for DPMC appear to
predominately return to the ground state. If the geminate radical
pairs exist in a local minimum on the excited state surface, the
distance of this interparticle separation would place the radical
pair above the transition state for heterolytic dissociation on
the ground state surface. Thus, the transition to the ground state
surface via electron transfer would place the system in the region
of the transition state. The product distribution then would be
R. J. Am. Chem. Soc. 1995, 117, 4399-4400.
(22) Kim, H. J.; Hynes, J. T. J. Am. Chem. Soc. 1992, 114, 10528-
10537.
(23) Kim, H. J.; Hynes, J. T. J. Am. Chem. Soc. 1992, 114, 10508-
10528.
(24) Mathis, J. R.; Kim, H. J.; Hynes, J. T. J. Am. Chem. Soc. 1993,
115, 8248-8262.
(25) Klessinger, M. Angew. Chem., Int. Ed. Engl. 1995, 34, 549-551.
(26) Walling, C.; Waits, H. P.; Milovanovic, J.; Pappiaonnou, C. G. J.
Am. Chem. Soc. 1970, 92, 4927-4932.
JP952672V