Influence of solvent viscosity on the photoisomerization of a novel trans-stilbene derivative with hindered single bond torsion

Article abstract of DOI:10.1016/S0009-2614(99)00588-6

Fluorescence life-time measurements were performed at elevated pressure and temperature to examine the viscosity dependence of the rate of photoinduced isomerization, k_iso, in a stilbene derivative in which the single bond torsion is hindered by the introduction of CH₂ bridges. The results obtained with n-hexane and methylcyclohexane as solvents (0.2 mPa s<η<2 mPa s) could not be fitted satisfactorily on the basis of Kramers' expression. A good global fit was, however, achieved when applying the empirical relationship k_iso=C·(η/mPa s)^-a·exp(-E_A/RT). The value for a drops from about 0.30 in n-hexane to about 0.15 in methylcyclohexane thus pointing to solvent specific contributions to the friction. The derived activation energies E_A are about 12.8 and 12.5 kJ mol^-1 in n-hexane and methylcyclohexane, respectively.

Full text of DOI:10.1016/S0009-2614(99)00588-6

23 July 1999
Ž
.
Chemical Physics Letters 308 1999 211–217
www.elsevier.nlrlocatercplett
Influence of solvent viscosity on the photoisomerization of a
novel trans-stilbene derivative with hindered single bond torsion
a
a
a
b
S. Schneider a,), B. Brem , W. Jager , H. Rehaber , D. Lenoir b,1, R. Frank
¨
a
Institut fur Physikalische und Theoretische Chemie, Friedrich-Alexander-UniÕersitat, Egerlandstr. 3, D 91058 Erlangen, Germany
¨
¨
b
¨
Institut fur Okologische Chemie, GSF-Forschungszentrum Neuherberg, D 85758 Oberschleißheim, Germany
¨
Received 19 April 1999; in final form 20 May 1999
Abstract
Fluorescence life-time measurements were performed at elevated pressure and temperature to examine the viscosity
dependence of the rate of photoinduced isomerization, kiso, in a stilbene derivative in which the single bond torsion is
Ž
hindered by the introduction of CH2 bridges. The results obtained with n-hexane and methylcyclohexane as solvents 0.2
.
mPa s-h-2 mPa s could not be fitted satisfactorily on the basis of Kramers’ expression. A good global fit was, however,
y
a
Ž
.
Ž
.
achieved when applying the empirical relationship kisosCP hrmPa s Pexp yEArRT . The value for a drops from
about 0.30 in n-hexane to about 0.15 in methylcyclohexane thus pointing to solvent specific contributions to the friction.
The derived activation energies EA are about 12.8 and 12.5 kJ moly in n-hexane and methylcyclohexane, respectively.
1
q 1999 Elsevier Science B.V. All rights reserved.
w x
1. Introduction
though the Grote–Hynes model 3 can adequately
describe the course of the rate constant of isomeriza-
tion, kiso, it suffers from the fact that unrealistically
small frequencies are obtained for the barrier. More
w x
Since the early theoretical work by Kramers 1 ,
the influence of solvents on the rate of chemical
reactions has been a main topic of chemical kinetics.
Over the years, the photoinduced isomerization of
trans-stilbene has become the most intensively stud-
ied reaction and the major test for theoretical models
w x
recently, Anderton and Kauffmann 4 pointed out
that, because stilbene is more polar in the transition
Ž
.
state than in the planar trans configuration, correct
evaluations of barrier energies in alcohols require
data not only for isoviscous but also isodielectric
Ž
of activated barrier crossing for a review see Ref.
2 . The essential result of all studies is that the
hydrodynamic Kramers equation does not adequately
model the friction dependence. It was therefore sug-
gested to introduce a frequency dependent friction
rather than the macroscopic shear viscosity. Al-
w x.
Ž
w x
solvents ‘isodielectric-Kramers–Hubbard 5 analy-
.
w
x
sis’ . According to Schroder et al. 6,7 the pressure
¨
dependence of the isothermal photoisomerization rate
reveals a significant change of the activation energy
even in n-alkanes with solvent density, caused by
the electrostatic solute–solvent interactions.
w
x
Hynes and co-workers 8,9 and Carmeli and
)
Corresponding author. Fax: q49 9131 852 8307; e-mail:
w
x
Nitzan 10 investigated a two-dimensional model in
which a reactive mode is coupled to a nonreactive
schneider@chemie.uni-erlangen.de
1 Also corresponding author.
0009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 99 00588-6

(
)
212
S. Schneider et al.rChemical Physics Letters 308 1999 211–217
mode. Both modes feel Markovian friction; however,
as the coupling between modes increases, the influ-
ence of friction on the reaction rate becomes weaker.
is even further reduced than in the five-membered
Ž .
ring derivative 5 . In this Letter, we will report on
fluorescence life-time measurements performed at
variable temperature and pressure aimed at elucidat-
ing the influence of solvent viscosity on the isomer-
Later, Schroder concluded that the observed "devia-
¨
tions from the one-dimensional Kramers model are
Ž .
manifestations of multidimensional barrier crossing
ization kinetics of the stilbene derivative 4 contain-
w x
w x
Ž
.
processes" 7 . Agmon and Kosloff 11 suggested a
model in which the motion is viscosity dependent in
only the reaction coordinate and its barrier height
dependent on the value of the perpendicular coordi-
nate, e.g. the phenyl twist angle. To provide further
evidence for such a model which was adopted by
several authors to explain their experimental results
ing two four-membered rings one CH2 group each .
According to model calculations, this derivative is
essentially planar in the ground state and thus well-
suited to elucidate the role of single bond rotation
during the double bond isomerization process.
w
x
12–14 it seems appropriate to study stilbene deriva-
tives in which the phenyl torsion is hindered. So far,
2. Materials and methods
Ž
only two derivatives of this kind are available see
.
Scheme 1 . In the first example, termed ‘stiff stil-
bene’ by Saltiel et al. 15 , the single bonds are
bridged by a five-membered ring containing two
2.1. Synthesis of new compound
w
x
Ž Ž
Ž
.
..
CH2 groups, 5 trans-1- 1-indanylidene -indane .
The title compound 4 was synthesized by reduc-
This derivative exhibits an excited state life-time
tive coupling of the corresponding ketone using low
w
x
w x
much shorter than trans-stilbene 16 . Pump–probe
experiments applying high pressure provided evi-
dence for the existence of two excited conformers
valent titanium species 19 .
The McMurry reagent was prepared in the follow-
w
x
Ž
.
ing way 20 : TiCl4 0.77 ml, 7 mmol and subse-
w
x
Ž .
quently zinc powder 1.33 g, 4 mmol were carefully
with different kinetic behaviour 17 .
Ž
.
The derivative containing two six-membered rings
added at 08C to anhydrous THF 75 ml . A solution
x Ž
Ž
w .
of benzocyclobuten-1-one 21 0.708 g, 6 mmol in
with three CH2 groups each, trans-1- 1-tetralin-
.
Ž . w
x
Ž
.
ylidene tetralane 6 18 , appeared to be non-fluo-
rescent. Since the oscillatory strength of the S0–S1
absorption band is of similar size as in trans-stil-
bene, one can conclude that the excited state life-time
THF 25 ml was added and the resulting black
slurry was refluxed for 21 h. After cooling, the
Ž
reaction mixture was quenched with K2CO3 75 ml
.
of a 10% aqueous solution . The mixture was ex-
Ž
.
tracted with diethylether 4 times with 50 ml each .
The etheral phase was washed with H2O, dried over
MgSO4, and the solvent evaporated in vacuo. The
Ž
.
crude product 0.72 g was purified by chromatogra-
Ž
.
phy over silica gel 50 g eluent: n-hexaneracetone .
Ž
.
Trans-4 and cis-4 0.42 g, 22% were obtained as a
1:1 mixture, as shown by GCrMS. A small amount
of the mixture of isomers was separated by
semipreparative HPL-chromatography: Gilson 350
Ž
apparatus with Merck LiChrospher columnn RP18,
.
particle size 5 mm , elution with acetonitrilerwater
2:1. Pure trans-4 was obtained by a modified purifi-
cation scheme as white crystals; mp 181–1838C,
q
Ž
.
Ž
.
Ž .
MS: mres204 74, M , 203 100 , 202 97 .
1
Ž
.
Ž
HyNMR 400 MHz, CDCl3 ds3.80 s, 4 H,
.
Ž
.
2yH , 7.11y7.33 ppm ABCD-system, 2 m, 8 H .
Scheme 1. Structures of compounds referred to in the text and
abbreviations used.
13
Ž
.
C-NMR 100.6 MHz, CDCl3 ds145.03, 144.86

(
)
S. Schneider et al.rChemical Physics Letters 308 1999 211–217
213
Ž
.
Ž
.
2s, C-2a, C-6a . 128.07, 127.52 2s, C-3, C-4 ,
Ž
.
Ž
.
Ž
.
127.45 s, C-1 , 122.74 s, C-5 , 118.77 s, C-6 ,
37.62 ppm s, C-2 . Anal. Calcd. for C16 H12: C,
Ž
.
94.07; H, 5.92. Found: C, 93.89; H, 6.30.
2.2. Fluorescence measurements
Fluorescence decay curves were recorded by
means of a home-built time-correlated single photon
counting apparatus using the frequency-doubled out-
put of a cavity-dumped, synchronously-pumped rho-
Ž
damine 6G dye laser tp ;10 ps, repetition rate 800
Fig. 1. Comparison of UVrvis absorption spectra of trans-stil-
.w
x
kHz 22 . The fluorescence was dispersed by a Zeiss
monochromator and detected by a channel plate mul-
tiplier Fa. Kemnitz, Berlin, model ELDY EM1 . In
combination with coventional electronics Tennelecr
Ortec a fwhm of the instrument response function of
Ž
.
Ž .
and the bridged derivative 4 – – – in
bene 1
Ž
.
n-hexane for structures see Scheme 1 .
Ž
.
Ž
.
less than 100 ps was achieved. Data analysis was
performed by the usual least-squares fitting proce-
dure assuming a multi-exponential decay law. The
high pressure optical cell was built according to a
trans-stilbene at low temperature in a highly viscous
environment. Accordingly, the 0–0 transition ap-
pears more pronounced in 4 and the vibrational
bands are better resolved. In the room temperature
emission spectra, this distinction is conserved in the
sense that the 0–0 transition is relatively stronger
than in 1. The fluorescence quantum yields of 4 are
Ž
design provided to us by Dr. D. Schwarzer Max-
Planck Institut fur Biophysikalische Chemie,
¨
.
Gottingen . It contains three windows at right angles
¨
w
x
and can be used for pressures up to 3.5 kbar. The
sample region and pressure producer are separated
by a transducer with the total sample volume being
approximately 8 ml.
higher than those of the parent trans-stilbene 23 ,
namely 0.06 versus 0.036 in hexane and 0.05 versus
0.016 in polar solvents.
w
x
Recently, Hohlneicher et al. 24 reported the
fluorescence spectra of 4 matrix-isolated in argon at
15 K. They analyzed the low frequency modes in
comparison with the Raman spectra and found that
"the most striking difference with respect to the
parent stilbene is the lack of pronounced progres-
sions of low frequency ag modes in the fused ring
derivatives".
Stationary fluorescence spectra were recorded at
Ž
.
the same excitation wavelength 308 nm on a Perkin
Elmer model LS50 spectrofluorimeter. Solvents were
Ž
Ž
.
used as provided by the suppliers n-hexane Ferak ,
Ž
..
methylcyclohexane Aldrich . Solutions were freed
of oxygen by bubbling argon through for an ex-
tended period of time.
3.2. Time-resolÕed fluorescence measurements
3. Results and discussion
As one may anticipate on the basis of the increase
in fluorescence quantum yield, the fluorescence de-
cay times of 4 at ambient temperature and pressure
are found to be longer than those of trans-stilbene,
namely 140 ps in n-hexane and 120 ps in methylcy-
clohexane. As in case of trans-stilbene, we find a
second component in the fit of the decay curves
3.1. Stationary absorption and emission spectra
In Figs. 1 and 2, the UVrvis absorption and
fluorescence spectra of 4 in n-hexane are compared
to those of the parent trans-stilbene 1. It has been
w x
pointed out before 2 that those substitutions which
Ž
constrain the twisting of the phenyl rings change the
spectral features in the absorption spectra such that
they become more similar to those observed for
exhibiting a life-time of about 1.6 ns fairly indepen-
.
dent of pressure and temperature . Since its relative
amplitude increases with the length of the measure-

(
)
214
S. Schneider et al.rChemical Physics Letters 308 1999 211–217
Ž
.
and
.
Fig. 2. Fluorescence spectra of trans-stilbene 1
Ž
.
Ž
the bridged derivative 4 – – – in n-hexane lex s308 nm .
Fig. 3. Rate of excited state isomerization, kiso, as derived from
Ž
.
pressure dependent fluorescence decay times solvent: n-hexane .
Solid lines represent a global fit using the parameters given in
Table 1.
Ž
.
ment time of irradiation we assign this component
to an unidentified photoproduct. In case of trans-
stilbene, it is well-known that the cis-isomer, either
present from the very beginning or produced by
photoinduced isomerization from the trans-isomer,
can react via 9,10-dihydrophenanthrene to phenan-
threne. The corresponding final products have also
been postulated in case of photodegradation of the
bene. In an analogous procedure, we have deter-
mined k0 s3.6=108 sy1 for compound 4 and we
Ž
will use this value independently of solvent n-hexane
.
or methylcyclohexane and temperature or pressure.
Consequently, the rates for kiso displayed in Figs.
3 and 4 are based on the equation:
w
x
derivatives with aliphatic bridges 25 . Due to its low
amplitude and the large difference in decay time, the
presence of this additional component did not influ-
ence the result for the fluorescence decay time of the
investigated compound. It should also be mentioned
that no additional shorter lived component at-
tributable to fluorescence from the cis-isomer could
be detected.
8
y1
w
x
kiso s1rtF y3.6=10 s
.
2
Ž .
The necessary values of the solvent viscosity were
determined from the published dependencies of h on
w
x
pressure and temperature 27 .
The rate for isomerization, kiso, is usually derived
from the experimentally determined fluorescence de-
cay times tF via the relation:
kiso s1rtF ykr ykISC ykIC s1rtF yk0 .
1
Ž .
For the numerical evaluation of kiso in trans-stilbene
and derivatives it is assumed that the rate for internal
conversion, kIC , can be neglected. The intersystem
crossing rate, kISC , was determined as approximately
3=108 sy1 26 and is therefore of the same order
of magnitude as the rate for radiative decay, kr. The
latter is determined with the help of the fluorescence
w
x
Ž
.
yield FF kr sFFrtF and is assumed to be con-
Fig. 4. Rate of excited state isomerization, kiso, as derived from
stant and independent of temperature and solvent
Ž
pressure dependent fluorescence decay times solvent: methylcy-
Ž
.
viscosity pressure . Several authors, e.g. Schroder
¨
.
clohexane . Solid lines represent a global fit using the parameters
0
8
y1
w
x
17 have chosen k as 6.2=10 s for trans-stil-
given in Table 1.

(
)
S. Schneider et al.rChemical Physics Letters 308 1999 211–217
215
Ž
.
Inspection of Figs. 3 and 4 immediately reveals
that the data points follow in the logrlog plot of kiso
versus h an approximately linear dependence with
ane, he found 12.1 kJrmol hs1.00 mPa s and
Ž
.
10.1 kJrmol hs10.0 mPa s thus pointing to the
importance of the molecular structure and specific
properties of the solvent.
If one wants to compare the results obtained in
this work for the bridged compound 4 with the
many, in part apparently contradictory results for
trans-stilbene 1, one should recall that the evaluation
of the viscosity and temperature dependence of the
Ž
.
Ž
slopes around 0.3 n-hexane and 0.15 methyl-
.
cyclohexane .
Ž
.
This qualitative finding suggests a global fit of
the data by means of the generally accepted relation-
ship:
kiso sF h Pexp yE rRT
Ž .
,
3
Ž .
Ž
.
A
Ž .
isomerization rate as expressed in Eq. 3 is done by
Ž .
with the viscosity dependent term F h being ap-
using different assumptions. We have chosen to eval-
uate our data with the minimum set of free fit
parameters, i.e. temperature independent values for a
w
x
proximated by a power law 28,29 :
ya
F h sCP hrmPas
Ž .
with 0(a(1 .
4
Ž .
Ž
.
Ž
.
and EA cf. Table 1 . It is obvious that by allowing
slightly temperature dependent slopes in the single
curve fitting we would produce viscosity dependent
activation energies from isoviscous Arrhenius plots.
Thus, our fits demonstrate that in case of the bridged
compound 4 the temperature effects on a andror EA
must be so small that they do not show up at the
level of accuracy of our experiment.
The solid lines in Figs. 3 and 4 represent the result of
a global fit using the same values for a and EA for
all temperatures and viscosities. That such a global
fit is possible in good quality is astonishing in view
of the different results obtained by various groups
for the parent trans-stilbene. Previous investigations
at ambient pressure yielded slopes of a ;0.32 in
w
x
higher alkanes 16,29–33 with viscosities ranging
between 0.1 and 4 mPa s. From transient absorption
measurements at variable pressure in n-hexane,
With these assumptions, we find that the activa-
tion energy of 4 is lower by about 0.3 kJ moly1 in
methylcyclohexane than in n-hexane. This trend par-
w
x
Schroder 17 derived a temperature dependent slope
allels the finding by Schroder for trans-stilbene and
¨
¨
Ž
.
Ž
rising from as0.29 Ts298 K to as0.46 Ts
points to the importance of specific solute–solvent
interactions for the magnitude of the activation bar-
rier.
.
453 K . It must, however, be mentioned that the
scatter in Schroder’s data collected at the highest
temperatures 398 and 453 K is fairly large and
consequently the values of the slopes at higher tem-
peratures are subject to some uncertainty.
Sundstrom and Gillbro 31 probed the trans-
stilbene isomerization kinetics by picosecond tran-
sient absorption spectroscopy in longer chain n-al-
¨
Ž
.
The small reduction of the activation barrier in 4
when compared to that given by Schroder for 1
¨
Ž
.
especially in n-hexane is interesting in view of
w
x
earlier discussions which proposed that the torsion
around the single bond could facilitate isomerization
around the double bond. On the other hand, in the
derivative comprising the five-membered rings, 5,
the thermal activation barrier is lowered even more,
¨
Ž
.
kanes ns12,14,16 . Their isoviscosity plot for hs
1.64 mPa s yielded an intrinisic barrier for isomeriza-
namely to only about 6.3 kJ moly1 15–17 . Since in
contrast to the latter compound, the rate for isomer-
w
x
w
x
tion of only 10.1 kJrmol. Courtney et al. 32 found
an activation barrier of 15 kJrmol in hexane. In
w
x
contrast, Park and Waldeck 33 reported a value of
about 18 kJ moly1 as average activation energy for
Ž
.
hydrocarbon solvents 350 K-T-278 K . They
found an increase in activation energy with the length
of the alkane chain, i.e. with increasing viscosity.
The origin of the latter trend was commented as
being opposite to that expected for a solvation effect.
Table 1
Parameters derived from a global fit of the pressure dependent
rates of isomerization Figs. 3 and 4 by Eqs. 3 and 4
Ž
.
Ž . Ž .
n-hexane
0.93=1012
0.26
Methylcyclohexane
1.08=1012
0.14
y1.
Ž
C s
w
x
According to Schroder 7,17 , the activation energy
in n-hexane drops from 14.3 hs0.33 mPa s to
11.8 kJrmol hs1.15 mPa s . In methylcyclohex-
¨
a
Ž
.
.
y1
Ž
.
EA kJ mol
12.8
12.5
Ž

(
)
216
S. Schneider et al.rChemical Physics Letters 308 1999 211–217
ization is lower in 4 than in 1, the pre-exponential
bridged compound 4 than in the parent trans-stil-
bene. Whereas for the latter EA was found to de-
crease with solvent viscosity, it appeared to be vis-
cosity independent in the derivative 4. Fits applying
Kramers’ expression yielded unreasonable values for
the barrier frequency v0 and too high activation
energies EA.
It has been speculated that a good agreement
between Kramers’ model and experimental data can
be found in two types of cases: first, if the barrier is
low enough that other degrees of freedom do not
open up comparably efficient pathways. Second, if
the isomerization is more clearly restricted to one
dimension. Although the latter requirement should be
factor must also be lower. For trans-stilbene, Court-
12
ney et al. 32 determined Cs3.5=10 sy1 with
w
x
the corresponding as0.32. Our analysis yields Cs
0.93=1012 n-hexane and 1.08=1012 methyl-
Ž
. Ž
.
Ž
.
cyclohexane cf. Table 1 .
If the value of the exponent a is taken as a
measure for the friction effect, then isomerization of
4 is subject to similar interactions as trans-stilbene
in n-hexane, but subject to stronger interaction in
w
x
methylcyclohexane. Saltiel et al. 34 investigated the
isomerization kinetics of trans-stilbene, trans-4,4X-
dimethylstilbene and trans-4,4X-dimethoxystilbene in
a series of n-alkane solvents. Applying a viscosity
Ž .
Ž
.
dependent term as given by Eq. 4 , the exponents a
better fulfilled in the bridged derivative 4, the
in that series of compounds decreased from 0.33 via
0.26 to 0.22 in contrast to the expectation of free
volume theory. A possible explanation offered by
these authors assumes a solvent and temperature
dependence of the activation entropy. A solvent me-
diated entropy change could also possibly explain
the differences in the pre-exponential factor of 4 in
n-hexane and methylcyclohexane.
Before closing the discussion, it should be men-
tioned that the observed viscosity dependence of kiso
could not be fitted by the modified Kramers’ expres-
sion
simple Kramers’ model does not yield a satisfactory
description. In our high pressure studies of the har-
pooning mechanism of semiflexible bridged electron
donor–acceptor systems, we found that a proper
description of the friction effect on the large ampli-
0.7
Ž
.
w x
35 . In
tude motion is possible by the term Arh
contrast to the present results, the same exponent
could be used for the fit of rate constants measured
in n-hexane, n-octane and methylcyclohexane. This
shows once more that general statements about the
proper description of friction effects are impossible.
F h sAP
Ž .
hrB 2 q1 1r2 y hrB
,
5
Ž
.
Ž
.
Ž .
½
5
Acknowledgements
except if completely unreasonable parameters were
used.
The authors wish to thank Dr J. Schroder and Dr.
¨
Ž
.
D. Schwarzer Gottingen for many stimulating
¨
discussions and the members of the machine shop
for excellent work in constructing the high pres-
sure apparatus. Financial support by Deutsche
Forschungsgemeinschaft and Fonds der Chemie is
also gratefully acknowledged.
4. Summary
In order to learn more about the influence of
torsion around the single bonds onto the excited state
kinetics of trans-stilbene, a new derivative was syn-
thesized in which the torsion around the single bond
is hindered by a CH2 bridge. The fluorescence decay
times could be determined in n-hexane and methyl-
References
w x
Ž .
H.A. Kramers, Physica 7 1940 284.
1
Ž
w x Ž .
cyclohexane at varying temperatures 283 K(T(
2 D.H. Waldeck, Chem. Rev. 91 1991 415.
w x
Ž .
H.F. Grote, J.T. Hynes, J. Chem. Phys. 73 1980 2715.
3
.
Ž
.
343 K and pressures 1 bar(p(3600 bar . The
w x
Ž
.
4
R.M. Anderton, J.F. Kauffman, Chem. Phys. Lett. 237 1995
145.
analysis of the rates of isomerization suggests that
only the power law F h sCPh
all data points satisfactorily. From the global fits,
ya
Ž .
is suited to fit
w x
Ž .
5
w x
6
P. Hubbard, Phys. Rev. 131 1963 1155.
J. Schroder, J. Troe, P. Vohringer, Z. Phys. Chem. 188
¨ ¨
Ž
.
one can conclude that EA is slightly smaller in the
1995 287.

(
)
S. Schneider et al.rChemical Physics Letters 308 1999 211–217
217
w x
Ž .
J. Schroder, Ber. Bunsenges. Phys. Chem. 95 1991 233.
¨
w
x
7
w x
8
24 G. Hohlneicher, I. Kautz, B. Tillmanns, D. Lenoir, R. Frank,
Ž
.
Ž .
J. Mol. Struct. 348 1995 191.
R.F. Grote, J.T. Hynes, J. Chem. Phys. 74 1981 4465.
w x
Ž
Ž
.
.
w
x
Ž
.
9
G. van der Zwan, J.T. Hynes, Chem. Phys. 90 1984 21.
25 C.F. Mallory, F.B. Mallory, Org. React. 30 1984 1.
w
w
w
x
w x
26 L.A. Brey, G.B. Schuster, H.G. Drickamer, J. Am. Chem.
10 B. Carmeli, A. Nitzan, Chem. Phys. Lett. 106 1984 329.
x
Ž
.
Ž .
Soc. 101 1979 129.
11 N. Agmon, R. Kosloff, J. Phys. Chem. 91 1987 1988.
x
w
w
w
w
x
Ž
.
12 J. Schroder, D. Scharzer, J. Troe, F. Voß, J. Chem. Phys. 93
27 D.W. Brazier, G.R. Freeman, Can. J. Photochem. 47 1969
¨
Ž
.
1990 2393.
893.
w
w
x
Ž
.
x
28 D. Gegiou, K.A. Muszkat, E. Fischer, J. Am. Chem. Soc. 90
13 N.S. Park, D.H. Waldeck, Chem. Phys. Lett. 168 1990 379.
x
Ž
.
14 A.Z. Szarka, N. Pugliano, D.K. Palit, R.M. Hochstrasser,
1968 , 12, 3907.
Ž
.
x
Chem. Phys. Lett. 240 1995 25.
29 G.R. Fleming, S.H. Courtney, M.W. Balk, J. Stat. Phys. 42
w
w
x
Ž
.
Ž
.
15 J. Saltiel, J.T. D’Agostino, J. Am. Chem. Soc. 94 1972
1986 83.
x
6445.
30 M. Lee, A.J. Bain, P.J. McCarthy, C.H. Ilan, J.N. Haseltine,
x
Ž
.
16 G. Rothenberger, D.K. Negus, R.M. Hochstrasser, J. Chem.
A.B. Smith III, R.M. Hochstrasser, J. Chem. Phys. 85 1986
Ž
.
Phys. 79 1983 5360.
4341.
w
w
w
w
w
x
w
x
31 V. Sundstrom, T. Gillbro, Ber. Bunsenges. Phys. Chem. 89
¨
17 J. Schroder, Habilitationsschrift, Universitat, Gottingen, 1991.
¨
¨
¨
x
Ž
.
Ž
.
18 P. Lemmen, D. Lenoir, Chem. Ber. 117 1984 2300.
1985 222.
x
Ž
.
w
x
19 D. Lenoir, Synthesis 1989 883.
32 S.H. Courtney, M.W. Balk, L.A. Philips, S.P. Webb, D.
x
Ž
.
Ž
.
20 J. McMurry, Chem. Rev. 89 1989 1513.
Yang, D.H. Levy, G.R. Fleming, J. Chem. Phys. 89 1988
x
21 P. Schiess, M. Heitzmann, Angew. Chem., Int. Ed. Eng. 16
. Ž
6697.
Ž
.
w
x
Ž
.
469 1977 .
33 N.S. Park, D.H. Waldeck, J. Chem. Phys. 91 1989 943.
w
x
w x
34 J. Saltiel, A.S. Waller, D.F. Sears Jr., E.A. Hoburg, D.M.
22 W. Jager, Ph.D. Dissertation, Friedrich-Alexander-
¨
Ž .
Zeglinski, D.H. Waldeck, J. Phys. Chem. 98 1994 10689.
Universitat, Erlangen, 1997.
¨
w
x
w x
35 S. Schneider, W. Jager, J.W. Verhoeven, Chem. Phys. Lett.
¨
23 S.L. Murov, J. Carmichael, G.L. Hug, Handbook of Photo-
chemistry, 2nd edn., Marcel Dekker, New York, 1993.
Ž
.
270 1997 50.