Quantum yields for the photodissociation of iodine in compressed liquids and supercritical fluids

Article abstract of DOI:10.1524/zpch.2001.215.2.183

Quantum yields of photodissociation were determined for iodine in compressed liquid n-alkanes and supercritical CO₂ and xenon after laser excitation at 532 nm. The quantum yield decreases nearly linearly with increasing density. No cluster effe

Full text of DOI:10.1524/zpch.2001.215.2.183

Zeitschrift für Physikalische Chemie, 215, 2, 183–195 (2001)
by Oldenbourg Wissenschaftsverlag, München
Quantum Yields for the Photodissociation
of Iodine in Compressed Liquids and
Supercritical Fluids
∗
By D. Schwarzer , J. Schroeder and C. Schröder
Max-Planck-Institut für Biophysikalische Chemie, Am Faßberg, D-37077 Göttingen,
Germany
Dedicated to Prof. Dr. Dr. h.c. mult. Jürgen Troe
on the occasion of his 60th birthday
(Received July 28, 2000; accepted August 15, 2000)
Photolytic Cage Effect
Quantum yields of photodissociation were determined for iodine in compressed liquid
n-alkanes and supercritical CO2 and xenon after laser excitation at 532 nm. The quantum
yield decreases nearly linearly with increasing density. No cluster effects at low density
were observed at this excitation energy. Our results are compared with molecular
dynamics simulations and point at the importance of kinematic processes involved in the
recombination dynamics of iodine.
1. Introduction
The photodissociation and recombination of iodine is one of the most exten-
sively studied chemical reactions in liquids and, therefore, served as a model
system for understanding various aspects of condensed phase reaction dynam-
ics. The interest goes back to Franck and Rabinowitch proposing the possibility
that after photoexcitation separating iodine atoms could be trapped in a sol-
vent cage reducing the dissociation quantum yield with respect to the gas
phase [1]. Subsequent quantitative investigations of the cage effect by Noyes
were aimed at understanding the dependence of iodine dissociation quantum
yields on the wavelength of the excitation light and solvent viscosity [2–4].
In systematic studies over wide density ranges Troe and co-workers [5, 6] and
van den Bergh and co-workers [7–9] found quantum yields smaller than unity
at surprisingly low gas pressures, namely, one or two orders of magnitude be-
low liquid densities. Similar results were obtained with excitation at 694 nm
*
Corresponding author. E-mail: dschwar@gwdg.de

184
D. Schwarzer, J. Schroeder and C. Schröder
and 581 nm, i.e. about 2000 cm−1 and 4800 cm−1 above the dissociation limit
of ground state iodine, respectively. The observed drop in quantum yield was
explained by a cluster mechanism proposing the formation of iodine-solvent
complexes which after laser excitation fragment and thereby dissipate the ex-
cess energy [10, 11].
In liquids at high pressures the quantum yield φdiss was found [10, 12] to
steeply decrease as
φdiss ∝ η−2,
(1)
where η is the solvent viscosity. The latter observation was explained [10, 11]
with a diffusion model including an energy relaxation rate constant for the
geminately recombined iodine atoms which is inversely proportional to the
solvent self diffusion coefficient D. This model suggests that the effective col-
lision frequency Z for vibrational energy relaxation in dense media increases
as
Z ∝ D−1.
(2)
A detailed microscopic picture of the photodissociation and geminate recom-
bination of iodine bases on picosecond experiments which are available since
the mid 1980s (see review [13] and references cited therein). According to
these and later studies [14–16] the dynamics is characterised by a variety
Fig. 1. Photodissociation and recombination of iodine in solution (adapted from [13]).

Quantum Yields for the Photodissociation of Iodine in Compressed Liquids and ...
185
of competing processes summarised in Fig. 1. After light absorption +in the
range 450–600 nm the dissociative Bꢀꢀ 1u( 1Π) or predissociative B 0u ( 3Π)
electronic state is populated causing the iodine atoms to separate on repul-
sive potentials. A fraction of atoms collide with surrounding solvent particles
and recombine geminately in the cage. Partially they get trapped in weakly
bound A 1u( 3Π) and Aꢀ 2u( 3Π) electronic states from which they can relax to
+
the X 0g ( 1Σ+) ground state by radiative or nonradiative transitions. Partially
they immediately recombine to the ground state where the iodine molecules
are stabilised by vibrational energy relaxation. A fraction of separating atoms
succeeds to escape from the solvent cage. These atoms can still recombine
geminately within several tens of picoseconds or separate by diffusion and fi-
nally nongeminately recombine on a microsecond time scale.
The experimental investigations were accompanied by extensive theoret-
ical studies which in the past often were based on Monte Carlo or classical mo-
lecular dynamics (MD) simulations [17–19]. Recently, even the complicated
curve crossing problem involved in the recombination process was dealt with
in coupled quantum-classical MD simulations by taking into account nonadia-
batic transitions among 23 Hund’s case (c) covalent molecular states [20].
Most computer simulations and time resolved experimental investiga-
tions suggest that the quantum yield for dissociation is mainly determined
by the primary in-cage recombination. The partitioning between cage es-
cape and recombination, however, is very fast (1–2 ps), i.e., this process is
dominated by highly energetic, nonthermal collisions emphasizing that dif-
fusive models of geminate recombination are inappropriate. On the other
hand the interpretation of photodissociation quantum yield measurements of
iodine [10] and bromine [21] just bases on such models. Furthermore, an
energy relaxation rate constant was deduced from these quantum yield meas-
urements suggesting that the collision frequency Z in dense media is pro-
portional to D−1, see Eq. (2). Recent experiments on the density dependence
of the vibrational energy relaxation rate of highly excited azulene, however,
clearly show an increase of Z with the local solvent density in the immedi-
ate surrounding of the solute [22, 23] which is much weaker than expressed
by Eq. (2).
Even more puzzling is the drop of the iodine photodissociation quan-
tum yields observed at relatively low densities. The interpretation that this
effect is caused by cluster formation [10] seemed to be supported by experi-
ments with I2-Ar van der Waals complexes formed in supersonic molecular
beams [24, 25]. In these experiments, photoexciting the region 0–1400 cm−1
above the B 0+( 3Π) state dissociation limit of the bare I2 molecule, fluores-
cence from a uvibrational distribution within the B state was observed. The
implication of these experiments is that fragmentation of the van der Waals
complex can leave the I2 molecule below its dissociation limit while the ex-
cess energy is deposited in the I2-solvent relative translational motion. For
a long time it was unclear what dynamical mechanism was responsible for

186
D. Schwarzer, J. Schroeder and C. Schröder
this so-called one-atom cage effect, since over the years many of the no-
ble gas-I2, -Br2 and -Cl2 structures were found to be of triangular (or T-
shaped) geometry (see [26] and references cited therein) which make a kine-
matic interpretation of the phenomenon difficult. Attempts to interpret the
I2-Ar data using triangular geometry by means of non-adiabatic models were
made [27], but now it is known that the one-atom cage effect is due to a lin-
ear isomer [28]. Recent wave packet and quasiclassical trajectory calculations
confirm that the I2-Ar photodissociation-recombination can be explained as
a purely kinematic process on the B state electronic surface of the collinear
cluster [29, 30].
In contrast to these molecular beam experiments, the low density quantum
yield measurements of Troe and co-workers on iodine were performed in the
bulk at T ≥ 314 K [5, 6]. Although it is feasible that a sufficient number of van
der Waals complexes is present under conditions where the cluster effect was
observed [31], the probability that they appear in a linear geometry is very un-
likely due to thermal excitation of the clusters. Furthermore, the excess energy
in these studies was considerably higher than in the molecular beam experi-
ments where the one-atom cage effect was observed. This brings us back to
the problem to imagine a mechanism which accounts for the photoabsorptive
dissociation of the complex into a solvent fragment and a vibrationally excited
iodine molecule in the bound A, Aꢀ, or ground state (attempts into this direction
have been made by Dardi and Dahler [32]).
In order to elucidate the reasons for the discrepancies discussed above, we
re-determined iodine dissociation quantum yields over wide pressure ranges
after photoexcitation at 532 nm. We will show that in contrast to earlier studies
the quantum yield φdiss decreases nearly linearly with density and that at low
densities there is no evidence for a cluster effect. Experimental results are in
agreement with model calculations basing on MD-simulations.
2. Experimental technique
In our experiments I2 was photolyzed by nanosecond laser pulses at 532 nm.
The quantum yield for I2 dissociation was measured directly after the laser
pulse by monitoring the bleaching of molecular iodine absorption at 514 nm.
The experimental setup is shown schematically in Fig. 2. The excitation light
source was a frequency doubled Q-switched Nd:YAG laser (Quantel Brillant)
operating at a repetition rate of 25 Hz with pulse lengths of 5 ns. An aperture
placed in front of the sample cell restricted the diameter of the laser beam to
5.0 mm generating a nearly homogenous circular profile. The pulse energy be-
hind the aperture was typically 6 mJ and passed into the sample cell by means
of dielectric mirrors. Laser energies were measured in front of and behind the
sample cell by means of a calibrated pyroelectric detector. The sample cell
was placed in a dual beam absorption spectrometer. Analysis light at 514 nm

Quantum Yields for the Photodissociation of Iodine in Compressed Liquids and ...
187
Fig. 2. Experimental setup.
was obtained from a cw argon ion laser (Lexel). 20 mW of the laser light was
divided into two paths. The signal beam passed through the sample cell to
a photodiode (Hamamatsu, S5972) while the reference beam traveled directly
to a second photodiode. A neutral density filter wheel in front of the reference
photodiode was used to match the laser intensities. Both photodiodes were pro-
tected against excitation light by means of narrow band interference filters and
were part of an electronic circuit subtracting their photocurrents ∆I(t) = I0 − I
as shown in Fig. 2. In this way most of common-mode laser noise was rejected
from the signal. The difference signal was amplified and recorded by a digi-
tal oscilloscope (Tektronics, TDS 640, 500 MHz). After averaging about 2000
laser excitation pulses changes of the iodine optical density were derived ac-
cording to ∆OD(t) = log(I0/(I0 −∆I(t))) were I0 was determined from the
low frequency current monitor of the reference photodiode. The over all time
resolution of the absorption spectrometer was 10 ns which is slow enough to
fade out geminate recombination and vibrational relaxation dynamics but suf-
ficient to give reliable information about photodissociation quantum yields of
iodine.
Experiments in supercritical fluids between 1 and 30 bar were performed in
a heatable stainless steel cell with an optical path length of 20 cm. At higher
pressures and for liquids a heatable high pressure cell with an optical path
length of 22 mm was used. Both cells were equipped with sapphire windows
of thickness 7 mm and aperture 5 mm. The windows were held by conical
fittings. Temperature was measured with an accuracy of 0.2 K, the precision
determining pressures was 1%. I2 concentrations inside the cell were adjusted
to ≤ 10−3 mol/l and solvents (p.a. grade, J. T. Baker) and gases (Messer–
Griesheim) were used without further purification.

188
D. Schwarzer, J. Schroeder and C. Schröder
3. Results and discussion
3.1 Photolysis quantum yields
In Fig. 3 typical absorption signals of molecular iodine in supercritical xenon
at 41 and 1022 bar after excitation with a 5 ns laser pulse at 532 nm are shown.
The temperature was 313 K. Assuming that geminate cage recombination, vi-
brational energy relaxation, and other fast phenomena are complete after 10 ns,
iodine photodissociation quantum yields φdiss were derived from absorption
jumps ∆OD at t = 0 and the number of excitation photons absorbed in the
sample. The latter was derived from the laser pulse energies in front of and be-
hind the high pressure cell. In the analysis, reflection losses at air-sapphire and
sapphire-sample interfaces were taken into account using refractive indices of
sapphire of n = 1.78 and of the samples as derived from the Lorentz–Lorenz
formula. The slow increase of the absorption signals in Fig. 3 at t > 0 is due
to nongeminate recombination of iodine which at high densities is a diffusion
controlled process and well understood [10, 21].
Fig. 3. Recombination of iodine in xenon at 41 and 1022 bar after photodissociation with
laser pulses at 532 nm (observation wavelength: 514 nm; T = 313 K).

Quantum Yields for the Photodissociation of Iodine in Compressed Liquids and ...
189
For the calculation of photodissociation quantum yields, absorption coef-
ficients at excitation and observation wavelengths of ε532 = 888 l mol−1 cm−1
and ε514 = 898 l mol−1 cm−1, respectively, were used. Since both wavelengths
are very close to the centre of the broad I2 absorption band and this band was
shown to hardly shift with pressure, we assumed ε532 and ε514 to be constant
over the whole density range. Then φdiss was calculated according to Lambert–
Beer’s law:
πr2 NA
φdiss = −∆OD(t = 0)
,
(3)
ε514 NP
where r, NP, and NA are the radius of the excitation laser beam in the sample,
the number of absorbed photons, and Avogadro’s number, respectively.
Our results for photodissociation quantum yields of iodine in liquid n-
pentane, n-hexane, and n-heptane at room temperature are shown in Fig. 4.
Within experimental error φdiss linearly decreases with density for all liquids as
indicated by dashed lines which intersect at φdiss(ꢀ = 0) = 1. For comparison,
data from [10, 12] are shown which were obtained in n-heptane after laser exci-
tation at 590 nm. These quantum yields are considerably below our results and
strongly deviate from a linear density dependence.
In Fig. 5 photodissociation quantum yields for iodine in supercritical CO2
at 323 K are shown (full squares). At each density several experiments were
performed. The error bars indicate the standard deviation of the measured
quantum yields. As for the alkanes, φdiss decreases nearly linearly with density.
A comparison with older data shows strong deviations. At low densities strong
cluster effects in CO2 were found using excitation wavelengths of 694 nm [5]
Fig. 4. Photodissociation quantum yields of iodine in compressed liquid n-alkanes at
298 K (full symbols: excitation at 532 nm, this work; open circles: excitation at 590 nm,
from [10, 12]; dashed lines intercept at φdiss(ꢀ = 0) = 1).

190
D. Schwarzer, J. Schroeder and C. Schröder
Fig. 5. Photodissociation quantum yields of iodine in CO2 ( : λexc = 532 nm at 323 K, this
work; : λexc = 590 nm at 298 K [10]; ꢁ : λexc = 581 nm at 330 K [6]; : λexc = 694 nm at
◦
314 K [5]).
(open squares) and 581 nm [6] (open triangles), respectively. Also at liquid
densities, at an excitation wavelength of 590 nm the reported quantum yields
were much lower [10] (open circles). In supercritical xenon at 313 K we ob-
served a similar density dependence as for CO2. As shown in Fig. 6, φdiss is
monotonously decreasing with density. Note that at low densities no cluster ef-
fect can be observed, although xenon is a highly polarizable solvent. Even at
the highest densities the photodissociation quantum yield is still around 25%.
Slightly higher quantum yields were found in xenon at 323 K by van den Bergh
and co-workers [8, 9] after excitation to the A or Aꢀ state at a wavelength of
694 nm (open triangles).
A comparison of iodine photodissociation quantum yields in different sol-
vent is most illuminating. In Fig. 7 all our data of Figs. 4–6 are plotted versus
the reduced solvent density ꢀr = ꢀ/ꢀc, where ꢀc is the critical density of the
corresponding solvent. This density scale is a measure of the packing fraction
of the respective fluid. As is evident from Fig. 7, all the data nearly fall on a sin-
gle line, indicating that φdiss is mainly proportional to the free space which is
available to the iodine atoms when they separate during the dissociation pro-
cess. The mass of the solvent is of minor importance as was already concluded
from early Monte Carlo simulations [17].
3.2 Molecular dynamics simulations
In the following we will compare our results in supercritical xenon with quan-
tum yields from classical molecular dynamics (MD) simulations carried out
by standard methods. In order to simplify the calculations we have neglected
any curve crossings during the dissociation process of iodine. Although elab-

Quantum Yields for the Photodissociation of Iodine in Compressed Liquids and ...
191
Fig. 6. Photodissociation quantum yields of iodine in xenon ( : λexc = 532 nm at 313 K,
this work; ꢁ : λexc = 694 nm at 323 K [8, 9]; : in krypton with λ = 694 nm at 298 K
◦
exc
(density scaled) [5]). The experimental data are compared with MD simulations of iodine
in xenon ( —— : excitation to the Πu state (R = 0.55 nm); ······ : excitation to the Πu
1
1
state (upper curve: R = 0.45 nm, lower curve: R = 0.65 nm); ------ : excitation to the Aꢀ
1
state; : excitation to the Π state [19]; : excitation in the ground state [18]).
•
u
Fig. 7. Photodissociation quantum yields of iodine versus the reduced solvent density (ꢀc
is the critical density of the solvent).
orate methods are available to handle this problem [20, 33] they are still too
expensive to calculate trajectories of several tens of picoseconds length neces-
sary for modelling the geminate recombination process. In order to elucidate

192
D. Schwarzer, J. Schroeder and C. Schröder
the influence of the laser wavelength on φdiss we conducted two types of sim-
1
ulations, namely excitation to the repulsive Πu state by a 532 nm photon and
excitation to the weakly bound Aꢀ state by a 694 nm photon. The model system
used in this study consisted of one I2 molecule immersed in 254 and 498 xenon
atoms at low and high (> 12 mol/l) densities, respectively. A cubic simula-
tion box with periodic boundary conditions was used. The simulations were
performed in a microcanonical ensemble employing the leap-frog Verlet algo-
rithm. All particles interacted via pairwise additive potentials with a cut-off
radius of half the box length. Xenon-xenon and xenon-iodine interactions were
treated by Lennard-Jones potentials. For the iodine ground and excited Aꢀ state
1
Morse potentials were used. The repulsive iodine Πu state was of the form
c(r/α)−9.5 [34]. Potential parameters used in this study are listed in Table 1.
The simulations were started on the iodine ground state potential at a spec-
ified density from a face centred cubic structure of 256 or 500 particles, two
of which were iodine atoms separated by 0.260 nm, and the remaining were
xenon atoms. The initial velocities of the particles corresponded to 300 K.
After 10000 integration steps (1 fs step size) with velocity scaling the desired
temperature of about 313 K was usually reached. The system was further equi-
librated for 100 ps.
Initial conditions for trajectories on the excited surfaces were prepared in
accord with the Franck-Condon principle for vertical transitions, i.e. the ini-
tial I-I distance rI-I was derived from hν = Ve(rI-I)− Vg(rI-I), where hν is the
Table 1. Potential parameters.
Xe-Xe1)
4ε[(σ/r)12 −(σ/r)6]
σ = 0.4055 nm
ε/kB = 229 K
I-Xe1)
4ε[(σ/r)12 −(σ/r)6]
σ = 0.394 nm
ε/kB = 328.6 K
I-I (X-state)1)
D{exp[−2β(r −re)]−2 exp[−β(r −re)]}
D/kB = 18044.61 K
β = 18.67 nm−1
re = 0.2668 nm
I-I (Aꢀ-state)2)
D{exp[−2β(r −re)]−2 exp[−β(r −re)]}
D/kB = 3609 K
β = 21.04 nm−1
re = 0.3073 nm
I-I ( 1Πu-state)3)
c(r/α)−9.5
c/kB = 1.238·108 K
α = 0.1 nm
1
2
3
[36], [16, 37], [34].

Quantum Yields for the Photodissociation of Iodine in Compressed Liquids and ...
193
photon energy and Ve and Vg are potential energies of the excited and ground
state, respectively. To reach these conditions, trajectories on the ground state
were simulated for random times between 5 and 10 ps and continued until the
desired value of rI-I was approached. In this way for each density a pool of
500 starting configurations and velocities were generated which subsequently
were propagated for 100 ps on the excited states. During the simulation the I-I
distance rI-I(t) was recorded.
1
For the repulsive Πu state the resulting trajectories were analysed in the
following way. A distance R in the range 0.45–0.65 nm where the potential
curves of Fig. 1 converge was defined at which the separating iodine atoms
are assumed to return to the ground state surface. Once rI-I(t) exceeded R and
subsequently fell below this distance the iodine atoms were counted as recom-
bined. This seems to be a crude simplification of the recombination dynamics,
because the iodine atoms in principle can approach each other on a repulsive
potential such that after some time they will separate again. Furthermore, even
if the iodine atoms remain on the attractive surface, vibrational energy relax-
ation has to occur to stabilise the I2 product. At low densities the probability
for stabilisation during an encounter is not necessarily one. Both deficiencies
of the model would increase the photodissociation quantum yield. Therefore, if
they play a role, from our calculations we get lower limits of φdiss. In Fig. 8 time
dependent photodissociation quantum yields derived from this analysis with
R = 0.55 nm are shown for various densities. The characteristic time scale for
recombination is picoseconds to tens of picoseconds in agreement with earlier
calculations [17–19, 35] and time resolved experiments [13, 15]. The quantum
1
yields at the end of the calculations (at t = 100 ps) on the repulsive Πu state
are shown in Fig. 6 and compared with our experimental values and data from
Fig. 8. Time dependent photodissociation quantum yields of iodine in xenon at various
1
densities (MD simulations, excitation to the Πu state with R = 0.55 nm).

194
D. Schwarzer, J. Schroeder and C. Schröder
earlier MD-simulations in xenon at 280 K [19]. General agreement is very
good. Our simulation results are quite insensitive to the actual choice of R: the
full line is obtained with R = 0.55 nm, upper and lower dotted lines indicate
quantum yields with R = 0.45 nm and R = 0.65 nm, respectively.
After exciting iodine to the Aꢀ state, the separating atoms do not need to
reach the ground state surface to recombine. Instead, due to collisions with
xenon atoms, the I2 molecule can be trapped directly in the attractive well of
the Aꢀ potential of 3609 K depth. In our MD simulations the photodissociation
quantum yield after excitation to the Aꢀ state was derived from the number of
trajectories for which after 100 ps the distance rI-I was above 0.6 nm relative to
the total number of trajectories. The corresponding density dependence of φdiss
is shown in Fig. 6 by a dashed line. The results are in agreement with classical
trajectory simulations on the I2 ground state surface in the presence of 22 rare
gas atoms using 694 nm excitation as well [18] (full circle), whereas the experi-
mental data of van den Bergh et al. seem to high. In Fig. 6 also experimental
results for φdiss obtained in krypton at 298 K using 694 nm excitation [5] are
presented (open circles). To account for the smaller Lennard-Jones diameter
of krypton (σKr = 0.366 nm) with respect to xenon (σXe = 0.4055 nm), kryp-
ton densities were multiplied by a factor (σKr/σXe)3. The agreement with our
calculations is excellent supporting the implication of Fig. 7 that φdiss is mainly
determined by the free space available to the separating iodine atoms rather
than the mass of the solvent particles.
Our simulations show that excitation to the Aꢀ state leads to smaller quan-
1
tum yields than excitation to the repulsive Πu state. The reason for this is that
1
on the Πu state the I2 population first has to reach distances where surface
hopping to the ground state can occur. At these distances the attractive forces
supporting recombination are weak (in our calculations they were essentially
zero). In contrast, on the Aꢀ state recombination appears to happen already at
the repulsive limb of the potential. This effect is most pronounced at high den-
sities were most trajectories on the Aꢀ potential are characterised by strongly
damped oscillations from the Franck-Condon region to the potential well.
4. Conclusions
In the present study we investigated the geminate recombination dynamics of
iodine by determining photodissociation quantum yields φdiss after laser excita-
tion at 532 nm in compressed liquid n-alkanes and supercritical fluids CO2 and
xenon. For all solvents φdiss was found to decay nearly linearly with increas-
ing density. A plot of φdiss versus the reduced density shows that the quantum
yield is independent of the nature of the solvent but mainly determined by the
free space available to the separating iodine atoms during the dissociation pro-
cess. Our results are supported by molecular dynamics simulations and clearly
show that the geminate recombination of iodine can be understood in terms of
a purely kinematic effect.

Quantum Yields for the Photodissociation of Iodine in Compressed Liquids and ...
195
Acknowledgement
The authors wish to thank Professor Jürgen Troe for continued support.
References
1. J. Frank and E. Rabinowitch, Trans. Faraday Soc. 30 (1934) 120.
2. F. W. Lampe and R. M. Noyes, J. Am. Chem. Soc. 76 (1954) 2140.
3. D. Booth and R. M. Hoyes, J. Am. Chem. Soc. 82 (1960) 1868.
4. L. F. Meadows and R. M. Noyes, J. Am. Chem. Soc. 82 (1960) 1872.
5. K. Luther, J. Troe, Chem. Phys. Lett. 24 (1974) 85.
6. H. Hippler, K. Luther, M. Maier, J. Schroeder and J. Troe, Laser Induced Processes
in Molecules, Springer, Heidelberg (1979) 286.
7. C. Dupuy and H. van den Bergh, Chem. Phys. Lett. 57 (1978) 348.
8. J. M. Zellweger and H. van den Bergh, J. Chem. Phys. 72 (1980) 5405.
9. J. C. Dutoit, J. M. Zellweger and H. van den Bergh, J. Chem. Phys. 78 (1983) 1825.
10. B. Otto, J. Schroeder and J. Troe, J. Chem. Phys. 81 (1984) 202.
11. J. Schroeder and J. Troe, Ann. Rev. Phys. Chem. 38 (1987) 163.
12. K. Luther, J. Schroeder, J. Troe and U. Unterberg, J. Phys. Chem. 84 (1980) 3072.
13. A. L. Harris, J. K. Brown and C. B. Harris, Ann. Rev. Phys. Chem. 39 (1988) 341.
14. N. F. Scherer, D. M. Jonas and G. R. Fleming, J. Chem. Phys. 99 (1993) 153.
15. C. Lienau and A. H. Zewail: J. Phys. Chem. 100 (1996) 18629.
16. R. Zadoyan, Z. Li, C. C. Martens and V. A. Apkarian, Chem. Phys. Lett. 218 (1994)
504.
17. D. L. Bunker and B. Slotten Jacobson, J. Am. Chem. Soc. 94 (1972) 1843.
18. J. N. Murrell, A. J. Stace and R. Dammel, J. Chem. Soc. Faraday Trans. 2 74 (1978)
1532.
19. J. K. Brown, C. B. Harris and J. C. Tully, J. Chem. Phys. 89 (1988) 6687.
20. V. S. Batista and D. F. Coker, J. Chem. Phys. 105 (1996) 4033.
21. H. Hippler, V. Schubert and J. Troe, J. Chem. Phys. 81 (1984) 3931.
22. D. Schwarzer, J. Troe and M. Zerezke, J. Chem. Phys. 107 (1997) 8380.
23. D. Schwarzer, J. Troe and M. Zerezke: J. Phys. Chem. A 102 (1998) 4207.
24. K. L. Saenger, G. M. McClelland and D. R. Herschbach, J. Phys. Chem. 85 (1981)
3333.
25. J. L. Valentini and J. Cross, J. Chem. Phys. 77 (1982) 572.
26. A. Rohrbacher, J. Williams and K. C. Janda, Phys. Chem. Chem. Phys. 1 (1999)
5263.
27. J. A. Beswick, R. Monot, J. M. Philippoz and H. van den Bergh, J. Chem. Phys. 86
(1987) 3965.
28. A. E. Stevens Miller, C.-C. Chuang, H. C. Fu, K. J. Higgins and W. Klemperer, J.
Chem. Phys. 111 (1999) 7844.
29. J.-Y. Fang and C. C. Martens, J. Chem. Phys. 105 (1996) 9072.
30. A. J. Conley, J.-Y. Fang and C. C. Martens, Chem. Phys. Lett. 272 (1997) 103.
31. P. S. Dardi and J. S. Dahler, J. Chem. Phys. 93 (1990) 242.
32. P. S. Dardi and J. S. Dahler, J. Chem. Phys. 98 (1993) 363.
33. A. A. Buchachenko and N. F. Stepanov, J. Chem. Phys. 104 (1996) 9913.
34. J. Tellinghuisen, J. Chem. Phys. 82 (1985) 4012.
35. M. W. Balk, C. L. Brooks and S. A. Adelman, J. Chem. Phys. 79 (1983) 804.
36. W. Wang, K. A. Nelson, L. Xiao and D. F. Coker, J. Chem. Phys. 101 (1994) 9663.
37. C. Teichteil and M. Pelissier, Chem. Phys. 180 (1994) 1.