Photochemistry with fast sample renewal using cluster beams: Formation of rare-gas halides in charge-transfer reactions in NF3-doped rare-gas clusters

Article abstract of DOI:10.1016/S0009-2614(99)00390-5

Charge transfer reactions in free clusters are observed in a photoluminescence study on doped rare-gas clusters (Rg clusters, Rg = Ar, Kr and Xe). Following photoexcitation into the first absorption bands of Rg clusters, fluorescence from free RgF* excimers ejected from the clusters and from Rg₂F* excimers localized in the interior of the clusters is observed. The results show that the reaction dynamics in clusters differs considerably from that in the gas and solid phase.

Full text of DOI:10.1016/S0009-2614(99)00390-5

28 May 1999
Ž
.
Chemical Physics Letters 305 1999 327–333
Photochemistry with fast sample renewal using cluster beams:
formation of rare-gas halides in charge-transfer reactions in
NF3-doped rare-gas clusters
a
a
b
b
L. Moussavizadeh , K. von Haeften , L. Museur , A.V. Kanaev b,1, M.C. Castex ,
c
a,c,)
R. von Pietrowski , T. Moller
¨
a
II. Institut fur Experimentalphysik, UniÕersitat Hamburg Hamburg, Germany
¨
¨
b
Laboratoire de Physique des Lasers CNRS, 93430 Villetaneuse, France
c
Hamburger Synchrotronstrahlungslabor HASYLAB am Deutschen Elektronen-Synchrotron DESY, Notkestrasse 85, D-22603 Hamburg,
Germany
Received 21 January 1999; in final form 25 March 1999
Abstract
Charge transfer reactions in free clusters are observed in a photoluminescence study on doped rare-gas clusters Rg
Ž
.
clusters, RgsAr, Kr and Xe . Following photoexcitation into the first absorption bands of Rg clusters, fluorescence from
free RgFU excimers ejected from the clusters and from Rg2FU excimers localized in the interior of the clusters is observed.
The results show that the reaction dynamics in clusters differs considerably from that in the gas and solid phase. q 1999
Elsevier Science B.V. All rights reserved.
1. Introduction
condensed matter often rather difficult because the
composition of the sample changes during the exper-
Photochemical processes in gas-phase and solid
samples differ considerably in various aspects. One
particular observation, which is often observed in
matrix studies, is the presence of long-lived radicals
and highly reactive atoms being formed in photodis-
iment. Thanks to the fast sample renewal in cluster
y6
Ž
beams typically clusters spend only 10
s in the
.
interaction region with the photon beam photochem-
ical processes in clusters are not affected by the
presence of reaction products. This allows the inves-
tigation of photochemical processes in condensed
matter under conditions similar to gas-phase experi-
ments.
w
x
sociation processes 1,2 . Due to their low mobility
in solid samples, these species are metastable against
recombination. This makes the detailed investigation
of relaxation dynamics and reaction pathways in
In this Letter, we report on the observation of
Ž
.
charge transfer CT reactions in clusters covering a
large-size range, namely the formation of RgFU and
Rg2 FU excimers in NF3-doped Rg clusters RgN s
Ž
)
Corresponding author. E-mail: thomas.moeller@desy.de
1 Present address: LIMHP CNRS, Universite Paris-Nord, 93430
Villetaneuse, France.
.
Ar, Kr; Xe; Ns2–500 . Analogous reactions are
w
x
very well characterised in the gas 3–5 and solid
0009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 99 00390-5

(
)
328
L. MoussaÕizadeh et al.rChemical Physics Letters 305 1999 327–333
w
x
.
phase 6–11 . Based on their single, weakly-bound
slits equipped with a liquid nitrogen cooled CCD
valence electron, excited RgU atoms can be regarded
Ž
.
camera Princeton Instruments . Spectra were col-
lected over 300 s. Simultaneously, the total VUV
fluorescence was detected by a channelplate detector
as pseudo-alkalis. In the gas phase, the reactions
between excited rare-gas atoms Rg and halogen
U
Ž
.
Ž .
Ž
.
Ž
X containing molecules MX , so-called ‘harpoon-
ing’ reactions, proceed through a curve crossing of a
covalent RgU –XM entrance potential with an Rgq–
Xy ion-pair potential. Subsequent dissociation of the
complex leads to the formation of RgXU excimers
coated with CsI behind a LiF window energy range,
.
7–11.5 eV . Excitation spectra of UV–IR fluores-
cence 2–6 eV covering the RgFU and Rg2 FU
fluorescence were recorded with a photomultiplier
Ž
.
Ž
.
Ž
.
Hamamatsu R943-02 with a GaAs Cs photo cath-
w x
4 . Previous attempts to study CT reactions in small
ode and a quartz window. For a given stagnation
pressure and temperature of the gas before expan-
sion, the mean cluster size N was calculated using
w
x
clusters, e.g. in XeN Cl2 12–14 or ArN O2 , have
shown that the characteristic excimer fluorescence
which is emitted after the CT reaction takes place is
only observed for small complexes with Ns1 or 2.
This already gives strong evidence that photochem-
istry in clusters and solids differs considerably.
w
x
the experimental calibration curve 15 .
3. Results and discussion
Fluorescence spectra of NF3-doped XeN , KrN and
ArN clusters recorded after excitation in the first
2. Experiment
Ž
.
absorption band of the pure clusters at 8.38 eV Xe ,
Ž
.
Ž
.
The experiments were performed at the CLULU
10.33 eV Kr and 12.11 eV Ar are presented in
Fig. 1. Under our experimental conditions, the clus-
ters are doped with only one NF3 molecule. This has
been checked by measuring the excitation spectra for
different NF3 densities in another set of measure-
Ž
Ž
.
cluster luminescence arrangement at HASYLAB
. w
x
Ž
.
DESY 15 . In short, RgN clusters Ns2–500
are prepared in a continuous free-jet expansion of
pure Ar, Kr or Xe gas at a stagnation pressure up to
5 bar and a temperature between 120 and 300 K
w
x
ments 17 . At sufficiently low densities, spectral
features due to NF3 dimers are absent. Several broad
and narrow emission bands can be seen in Fig. 1.
Ž
.
through a conical nozzle ds100 mm, 2ws308 . A
beam of NF3 molecules, prepared in an expansion
through a 300 mm nozzle, intersects the cluster beam
15 mm downstream. Mixed clusters were formed by
two different techniques. NF3-doped Ar and Kr clus-
ters are prepared by a co-expansion of 0.1% NF3 in
Ar or Kr gas or by a pick-up technique. Doped Xe
clusters are only formed by a pick-up technique. On
Ž
.
Ž
.
The narrow bands at 3.54 eV Xe and 5.0 eV Kr
are attributed to the B™X transition of XeFU and
KrFU. A similar band at 6.4 eV is observed for
Ž
NF3-doped Ar clusters the small band at 3.2 eV in
Fig. 1 is due to the second order of the monochro-
mator of this band 17,18 . These bands are the
prominent emissions in the gas-phase NF3rRg mix-
tures providing intense UV light in excimer lasers
. w
x
w
x
the basis of our results for Xe-doped Ar clusters 16 ,
we assume that the clusters are solid if they are
prepared by co-expansion. In view of the results
presented in the next paragraph, we assume that the
clusters are also solid if they are prepared by the
pick-up technique. The background pressure was
kept below 10y4 mbar during the experiment by a
continuous pumping of the interaction-volume. Tun-
able synchrotron radiation in the vacuum ultraviolet
w x
4 . The emissions are due to bound free fluorescence
from the lowest electronically excited CT RgqF state
labelled B to the repulsive ground state X. Weak
Ž
.
Ž
.
broad emissions at 2.56 eV Xe and 5.64 eV Ar
are assigned to the corresponding gas-phase C™A
w x
transitions 4 . Within the error bars, the measured
transition energies in the cluster agree with the gas-
Ž
.
Ž
.
Ž .
VUV range Dlexc s0.25 nm was focused on the
phase values see Table 1 . The broad and intense
Ž .
emissions at 1.53 eV Xe , 2.58 eV as well as 2.79
Ž
.
cluster beam. The UV–visible UVV fluorescence
was analysed by a Czerny–Turner-type monochro-
mator fs275 mm, 150 lrmm grating, 300 mm
Ž
.
Ž
.
eV Kr and 3.87 eV Ar are energetically close
2
2
Ž
Ž
.
-100 meV to the well known G™1,2 G transi-

(
)
L. MoussaÕizadeh et al.rChemical Physics Letters 305 1999 327–333
329
Ž
.
what higher 0.3–0.5 eV energies are observed in
w x
the gas phase at high pressure 4 . The width of these
emission bands from clusters is in good agreement
with the values for the gas and the condensed phase
Ž
.
see Table 1 . Therefore, we attribute the intense
broad emission bands in clusters to triatomic ex-
cimers. The large spectral shift of the triatomic
excimer emission with respect to the diatomic one is
due to the extra binding energy of Rgq2 Fy relative to
RgqFy which is approximately the binding energy
of Rgq2 . Interestingly, diatomic RgFU excimer emis-
sions are only observed in liquid and solid samples if
w x
ternary mixtures, e.g. Xe, Kr and F2 are prepared 6 .
In the case of NF3-doped ArN clusters, a few weak
yet unassigned bands between 2.1 and 3.0 eV are
Ž
observed the intense band at 1.95 eV is presumably
due to the second order from the Ar2 FU fluorescence
.
at 3.9 eV .
The simultaneous appearance of diatomic and tri-
atomic emissions from clusters can be explained in
the following way. Diatomic emissions are not shifted
relative to the gas-phase values and are accordingly
attributed to diatomic RgFU excimers desorbing from
the cluster. The absence of RgFU fluorescence in an
adsorbed configuration is not surprising since in the
adsorbed configuration RgFU can easily be con-
verted into Rg2 FU excimers which are considerably
Ž
Fig. 1. Fluorescence spectra of NF3-doped Rg clusters cluster size
Ž
.
Ž
.
Ž
..
Nf100 Xe ; Nf30 Kr ; Nf150 Ar . Assignments of dif-
ferent emission bands are labelled. Some very sharp bands at 1.79
eV Xe , 2.50 eV Kr , 2.0 and 3.22 eV Ar labelled by arrows
are due to the second order of the monochromator.
Ž
.
Ž
.
Ž
.
w x
lower in energy 4 . The large spectral red-shift of
approximately 0.5 eV of the Rg2 FU fluorescence
relative to the gas phase indicates that the triatomic
excimers do not desorb from the cluster. Thus, Rg2 FU
tion of the triatomic Rg2 FU excimers in liquid and
solid samples. Similar broad bands emitted at some-
Table 1
2
Comparison between the emission energy E and the width D E FWHM of diatomic B–X transitions and triatomic 4 G–1,2 2 G transitions
Ž
.
of Rg halides under various conditions
ArF
Ar2 F
KrF
Kr2 F
XeF
Xe2 F
E
D E
E
D E
E
D E
E
D E
E
D E
E
D E
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Ž
eV
.
Gas
6.43a
0.1a
4.3a
4.0b
0.68a
0.52b
5.0a
0.1a
2.95a
2.79b
2.74d
2.79c
2.59c
0.5a
3.55a
0.15a
2.0a
1.8b
1.6e
1.54c
0.3a
Liquid
Solid
Cluster
0.38b
0.37d
0.27c
0.19c
0.16b
0.29e
0.35c
6.43c
0.1c
3.9c
0.61c
5.0c
0.1c
3.55c
0.15c
a
w
x
From Ref. 27 .
b
w x
From Ref. 6 .
c This work.
d
w x
From Ref. 2 .
e
2
2
U
w
x
From Ref. 10 . In the case of Kr2 F , the 4 G–1,2 G transitions is split into two components.

(
)
330
L. MoussaÕizadeh et al.rChemical Physics Letters 305 1999 327–333
emissions are due to triatomic excimers at the sur-
face or in the interior of the cluster.
Cluster size-dependent fluorescence spectra of
NF3-doped XeN clusters are displayed in Fig. 2. The
excitation energy was tuned to the first excitonic
absorption band of XeN clusters at approximately
Furthermore, the intensity ratio J1,2 between XeFU
and Xe2 FU depends on the excitation energy. We
have measured the ratio for bulk and surface excita-
tions. It turned out that the ratio for surface and bulk
Ž
.
excitations here labelled Js,b can be described by
17 Js,b s1.05=Ny1r3 which is characteristic for
w
x
U
w
x
w x
8.4 eV 19 . The increase in fluorescence intensity
reflects the increase of the particle density of clusters
in the interaction region. This has been checked by
comparing the luminescence intensity of intrinsic
the surface to bulk ratio 20 . Since XeF is due to
desorbing excimers while Xe2 FU is due to excimers
in adsorbed states, we conclude that the desorption
rate of XeFU decreases with increasing cluster size.
The Ny1r3 of Js,b indicates that excimers formed at
the surface desorb. Furthermore, it gives evidence
that NF3 molecules are present in the surface as well
as being solvated in the interior sites of the cluster.
Here we assume that the clusters are solid for the
following reason: there is no significant difference
between the measured fluorescence spectra and fluo-
rescence excitation spectra of NF3-doped Ar and Kr
clusters prepared either by co-expansion or by the
Ž
fluorescence of the clusters in the VUV Xe2 centres
inside the cluster 19 and from the XeFU and
Xe2 FU excimers in the UV. With increasing cluster
size, the contribution of XeFU decreases. The inten-
sity ratio J1,2 of XeFU and Xe2 FU excimer fluores-
w
x.
cence can be described by J1,2 s0.25=Ny1r3 17 .
w
x
Ž
.
pick-up technique see below . If the clusters were
melted upon collisions with NF3, we would expect
significant differences between the spectra. The same
argument should hold for Xe clusters which are more
tightly bound. Which site in Xe clusters is preferen-
tially populated depends on the strength and the
bond length of the Xe–NF3 interaction in compari-
w
x
son with the Xe–Xe interaction 21 . Unfortunately,
there are no data available for the interaction poten-
tials between Xe and NF3. In view of the small size,
Ž
large polarizability and the dipole moment as3.62
=10y24 cmy3, ps0.235 debye, 28 of NF3 we
expect that it can stay at the surface as well as be
solvated in the interior of XeN clusters. In the case
of KrN clusters, we have performed measurements
on NF3-doped clusters prepared by either a pick-up
technique or conventional co-expansion. Fluores-
cence spectra recorded after excitation of clusters
w
x.
w
x
prepared by either method 17 show no significant
difference, indicating that NF3 molecules can, to a
certain extent, migrate inside the clusters.
Fluorescence excitation spectra of the intrinsic
KrN cluster fluorescence in the VUV and of Kr-halide
excimer fluorescence in the energy range 2–5 eV are
compared in Fig. 3. The strongest bands are due to
w
x
the well-known excitonic absorption bands 22 of
Ž
.
KrN clusters Nf500 labelled 1s, 1t and 1l and the
corresponding spin–orbit states 1sX, 1tX and 1lX. The
suffix ‘s’ denotes surface states, ‘l’ and ‘t’ longitudi-
Fig. 2. Fluorescence spectra of NF3-doped XeN clusters as the
function of size.

(
)
L. MoussaÕizadeh et al.rChemical Physics Letters 305 1999 327–333
331
tion on the Kr–NF3 complex and subsequent dissoci-
ation into CT states of the type KrqyFy or Krq2 Fy.
The absence of the sharp lines due to Kr atoms and
dimers in the fluorescence excitation spectrum of
KrFU and Kr2 FU fluorescence can be interpreted in
Ž .
two different ways. i Under our experimental con-
ditions, namely with a low concentration of NF3,
molecular complexes containing one NF3 molecule
and one or two Kr atoms are not formed or their
absorption is substantially shifted relative to the ab-
Ž .
sorption of bare Kr atoms and dimers. ii Larger
complexes favour the formation of the Rg halides.
On the other hand, it is well known that Rg halide
formation from gas-phase Kr and NF3 enclosed in a
w x
gas cell is very efficient 4 . Thus, we conclude that
under collision-free conditions, namely with a low
concentration of NF3, a minimum cluster size of
w
x
three to five Kr atoms 17 is required for the forma-
tion of van der Waals KrN–NF3 complexes and
subsequent formation of Rg halides. Interestingly,
the opposite has been observed for the XeN–Cl2
w
x
system 14 . In this case, only monomers and dimers
led to the formation of XeClU.
From spectrally resolved measurements of bare
KrN clusters, it is well known that the trimer and
larger clusters emit fluorescence from vibrationally
Ž .
a Fluorescence excitation spectra of intrinsic lumines-
Fig. 3.
relaxed KrU2 24 . Moreover, the ratio of the sum of
Kr2 FU and KrFU emission versus KrU2 fluorescence
Ž
.
Ž
.
w
x
cence energy range, 7–11.5 eV of KrN clusters full line and of
KrFU and of Kr2 FU excimer fluorescence dashed line, energy
Ž
.
range 2–6 eV formed in a photochemical reaction inside the
increased almost linearly with N1r3for N)5 see
Ž
clusters. Surface and bulk excitons of the clusters labelled s,sX
. w
x
X
X
Fig. 4 17 . In other words, the formation of Rg
halide excimers increases linearly with the radius of
the cluster. This gives evidence that their formation
is considerably larger in the interior of the cluster
than at the surface. In view of the fast relaxation of
the primary excited states into KrU2 , it seems to be
likely that Kr2 FU is formed in a two step process.
After excitation, KrU2 is formed inside the clusters. In
a second step, Kr2 FU is formed in a reaction with
NF3.
Ž
w
.
Ž
.
surface and t,t and l,l transverse and longitudinal bulk states
X
x
Ž
.
22,23 , the atomic resonance lines 5s,5s and dimer absorption
1
1
bands B Ýu and C Ýqu ,1u 26 are indicated. Experimental
. w
q
Ž
x
Ž .
b The
conditions: coexpansion of 0.1% NF3 in Kr; Nf500.
intensity ratio of the sum of KrFU, Kr2 FU fluorescence to the
intrinsic fluorescence of KrN clusters for N(500. It is clearly
visible that bulk states contribute significantly more to the forma-
tion of krypton halide excimers than surface states. The dashed
lines are only to guide the eye.
nal and transverse bulk exciton states, respectively
The intensity of the KrFU and Kr2 FU fluores-
cence, normalised to the intrinsic KrN fluorescence,
is presented in the lower part of Fig. 3. The increase
with increasing excitation energy indicates that there
is an energy threshold for the formation of the
halides. A threshold energy of 9.9"0.05 eV was
derived from the data in Fig. 3 by a linear fit of the
data between 10 and 10.4 eV. Furthermore, one can
see that surface and bulk excitons of the clusters
w
x
23 . The rather sharp lines in the excitation spectrum
Ž
.
of VUV radiation 7–11.5 eV are due to Kr atoms
and dimers which are always present in the beam. In
the fluorescence excitation spectrum of KrFU and
Kr2 FU these lines are absent, while the excitonic
absorption bands of KrN clusters are prominent. This
demonstrates that KrFU and Kr2 FU are formed after
excitation of KrN clusters, localization of the excita-

(
)
332
L. MoussaÕizadeh et al.rChemical Physics Letters 305 1999 327–333
are due to one-step processes, while in the solid a
two-step excitation is required. At present, we can
only speculate as to why in the solid the two-step
process is much more efficient than the one-step
process. One plausible explanation could be that fast,
radiationless transitions depopulate very efficiently
electronically excited states. In the cluster, these
processes may be not as efficient as in the solid,
allowing the harpooning reaction to take place.
In the cluster, it may happen as already discussed
that the excitation of the clusters becomes localized
on RgU2 self-trapped molecular centres 19,23 before
the energy is transferred to the NF3 molecule. In
addition, the branching ratio between formation from
RgU or RgU2 can depend on the excitation energy.
This could, e.g. explain the observed energy depen-
dence in Fig. 3b. If the formation of Rg halides
proceeds via RgU2 the available energy is consider-
ably reduced by the energy needed for the relaxation
into RgU2 and might not be sufficient in order to
allow the formation of the halide excimers. At pre-
sent, we cannot make a definite statement if the
formation of Rg halides proceeds via RgU or RgU2 .
Time resolution may help to solve this problem,
w
x
Fig. 4. The intensity ratio between the sum of KrFU and Kr2 FU
fluorescence in the UV and visible and the intrinsic luminescence
of KrN clusters in the VUV spectral range as a function of N1r3
.
The increase with N1r3 gives evidence that the formation of
krypton halide excimers is more favourable in the interior of the
cluster than at the surface. The dashed lines are only to guide the
eye.
X
X
labelled s,sX surface and t,t and l,l transverse and
Ž
.
Ž
.
longitudinal bulk states , respectively, both con-
tribute to the formation of the halides. In agreement
with the results discussed before, the yield of surface
states is 35% lower.
since the lifetime of the long-lived RgU2 is depend-
Ž
.
ing on the rare gas 10–100 times longer than that of
RgU 23 .
The Rg fluorine systems are characterized by a
w
x
Finally we compare the reaction dynamics in
clusters with corresponding processes in the gas and
solid phase. Studies in the gas phase show that
Ž
large excess energy between excitation f10 eV for
U
.
Ž
Kr and emission 2.5 eV for Kr2 F and 5 eV for
photoreactions starting with RgU or MXU lead de-
KrFU . If the total relaxation energy is less than 5
Ž
.
.
pending on the pressure with high efficiency, to the
eV, the formation of KrFU and Kr2 FU is energeti-
cally possible. From this viewpoint, Rg fluorides are
good candidates for the formation of halide excimers
in clusters. These findings may explain why previous
attempts to observe excimer emission from large
formation of RgFU or Rg2 FU 4 . Triatomic ex-
w x
cimers are preferentially formed at high pressure. On
the other hand, it is now well established that in the
w
x
solid a two-step process is usually required 1,2 .
First the MX molecule is dissociated by a photon of
sufficient energy and, in a next step induced by a
second photon, the photochemical reaction between a
halogen atom and the matrix leads to the formation
of Rg2 FU excimers.
w
x
Cl2- or O2-doped Rg clusters failed 12–14 . On the
other hand, in a study on F2-doped liquid krypton
with femtosecond multiphoton excitation, evidence is
given that Kr2 FU excimers can also be formed in a
direct reaction between F2 KrU centres formed in the
w
x
This raises the question: how do the reaction
dynamics in clusters differs from those in the solid?
Firstly, we like to point out that, at least under our
experimental conditions, two-step processes involv-
ing two photons are extremely unlikely because of
the low photon flux and the fast sample renewal.
Therefore, we can assume that the photo reactions
relaxation process of krypton excitons 25 . From the
time evolution, it could be concluded that the disso-
ciation of the F2 into two F atoms is not required in
the first step of the reaction. This indicates that under
certain conditions a reaction pathway similar to the
one reported here in clusters can be observed in
liquid rare gases.

(
)
L. MoussaÕizadeh et al.rChemical Physics Letters 305 1999 327–333
333
w x
Ž .
J. Ku, D.W. Setser, Appl. Phys. Lett. 48 1986 689.
H. Jara, H. Pummer, H. Egger, C.K. Rhodes, Phys. Rev. B
5
w x
6
4. Conclusions
Ž
.
30 1984 1.
The first CT reactions in clusters containing up to
500 atoms are reported. The fluorescence of RgFU
excimers desorbing from the cluster and of Rg2 FU
excimers emitting inside the cluster are observed
following photoexcitation in the first strong absorp-
w x
7
Ž
.
.
M.E. Farjardo, V.A. Apkarian, Chem. Phys. Lett. 134 1987
51.
N. Schwentner, V.A. Apkarian, Chem. Phys. Lett. 154 1989
413.
w x
8
Ž
w x
9
G. Zerza, G. Sliwinski, N. Schwentner, Appl. Phys. B. 55
Ž
.
1992 331.
tion bands of NF3-doped Rg clusters. The strongest
w
w
x
Ž
.
10 M.E. Fajardo, V.A. Apkarian, J. Chem. Phys. 89 1988
U
Ž
.
emissions can be attributed to the B–X RgF and
4102.
2
2
U
Ž
.
4 G–1,2 G Rg2 F transitions. In the case of KrN
clusters, the energy threshold of the reaction was
determined to be 9.9 eV. The reaction dynamics in
the clusters differs considerably from that in the
gas-phase and solid samples. In contrast to gas-phase
experiments performed in a gas cell, Rg2 FU emis-
sions dominate the fluorescence spectra, at least for
¨
x
11 H. Kunz, J.G. McCaffrey, M. Chergui, R. Schriever, O.
¨
Unal, V. Stepanenko, N. Schwentner, J. Chem. Phys. 95
Ž
.
1991 1466.
w
w
x
12 C. Dedonder-Lardeux, C. Jouvet, S. Martrenchard-Barra, F.
Ž
.
Talbot, D. Solgadi, Chem. Phys. Lett. 250 1996 293.
x
13 C. Dedonder-Lardeux, M. Berdah, C. Jouvet, S. Martren-
chard-Barra, J.M. Mestdagh, D. Solgadi, J.P. Visticot, J.
Ž
.
Chem. Phys. 104 1996 2470.
w
w
w
w
w
x
Ž
.
14 A. Tabayashi, A. Hiraya, K. Shobotake, UVSOR Annual
large clusters N)100 . In a similar way, measure-
ments on solid samples show dominant Rg2 FU fluo-
rescence. However, in the clusters a one-step excita-
tion process involving one photon is established
while in the solid a two-step excitation with two
photons is needed for the formation of Rg2 FU ex-
cimers. Thanks to fast sample renewal in cluster
beams, photoreactions can be studied without com-
plications due to the presence of reaction products.
Ž
.
Report 1990, 23 1991 .
x
15 R. Karnbach, M. Joppien, J. Stapelfeldt, J. Wormer, T.
¨
Ž
.
Moller, Rev. Sci. Instr. 64 1993 2838.
¨
x
16 M. Lengen, M. Joppien, R. Muller, J. Wormer, T. Moller,
¨
¨
¨
Ž
.
Phys. Rev. Lett. 68 1992 2362.
x
17 L. Moussavizadeh, Diplom-thesis, Universitat Hamburg,
¨
1996.
x
18 L. Moussavizadeh, R. von Pietrowski, M. Joppien, K. von
Haeften, A.R.B. de Castro, Y.V. Kanaev, M.C. Castex, T.
Ž
.
Moller, HASYLAB Jahresbericht 1994, 183 1994 .
¨
w
x
19 R. Karnbach, M. Joppien, T. Moller, J. Chim. Phys. 92
¨
Ž
.
1995 499.
w
w
w
x
Ž
.
20 J. Jortner, Z. Phys. D 24 1992 247.
x
Acknowledgements
Ž
.
21 L. Perera, F.G. Amar, J. Chem. Phys. 93 1990 4884.
x
22 J. Stapelfeldt, J. Wormer, T. Moller, Phys. Rev. Lett. 62
¨
¨
Financial support from the Deutsche Forschungs-
gemeinschaft under Grant Mo719r1-1 is kindly
acknowledged.
Ž
.
1989 98.
w
w
x
23 N. Schwentner, E.E. Koch, J. Jortner, Electronic Excitations
in Condensed Rare Gases, Springer, Berlin, 1985.
x
24 R. Karnbach, M.C. Castex, J.W. Keto, M. Joppien, J.
Wormer, G. Zimmerer, T. Moller, Chem. Phys. Lett. 203
¨
¨
Ž
.
References
1993 248.
w
w
x
25 M.W. Hill, R. Zasoyn, V.A. Apkarian, Chem. Phys. Lett.
w x
Ž
.
1
w x
2
J. Le Calve, P. Gurtler, J. Chem. Phys. 86 1989 1847.
¨
Ž
.
200 1992 251.
H. Kunttu, J. Feld, R. Alimi, A. Becker, V.A. Apkarian, J.
x
Ž
.
26 T. Moller, M.C. Castex, in: L. Frommhold, J.W. Keto Eds. ,
¨
Ž
.
Chem. Phys. 92 1990 4856.
Spectral Line Shapes Vol. 6, Vol. 216, AIP, New York,
1990, pp. 384–397.
w x
3
Ž
.
D.W. Setser, J. Ku, in: F. Lahmani Ed. , Photophysics and
Photochemistry above 6 eV, Elsevier, Amsterdam, 1985, p.
w
w
x
27 F.K. Tittel, G. Marowsky, W.L. Wilson, M.C. Smayling,
621.
Ž
.
IEEE J. Quantum Electr. QE17 1981 2268.
w x
4
C.K. Rhodes, Excimer lasers, 2nd edn., Springer, Berlin,
1984.
x
Ž
.
28 R.C. West Ed. , Handbook of Chemistry and Physics, 68th
edn., CRC press, Boca Raton, FL, 1987.