On the efficient generation of Arn-SH free-radical clusters via ultraviolet photolysis of closed-shell Arn-H2S complexes

Article abstract of DOI:10.1063/1.478410

Detailed investigations are described for the generation of open-shell radical complexes formed by unimolecular photolysis of closed-shell van der Waals precursors. As a specific test case, ultraviolet photolysis of slit-jet cooled Ar_n-H₂S (n≤2) complexes at both 248 and 193 nm are shown to yield Ar-SH and Ar₂-SH radical cluster species with surprisingly high efficiencies. Analysis of the laser induced fluorescence (LIF) spectra indicates that the radical complexes are produced with extensive van der Waals stretch/bend and overall rotational excitation, which is consistent with a simple ballistic model of the dissociation dynamics. The LIF spectra obtained as a function of expansion distance downstream provide clear evidence for remarkably efficient cooling of the newly formed radical cluster species by low-energy collisions with jet-cooled inert gas atoms at ≈10 K. Spectrally resolved Ar-SH and Ar₂-SH LIF signals have been investigated as a function of Ar composition, which yields information on relative branching ratios for fragmentary (e.g., Ar₂-H₂S→Ar-SH+H+Ar) and nonfragmentary (e.g., Ar₂-H₂S→Ar₂-SH+H) photolysis events.

Full text of DOI:10.1063/1.478410

On the efficient generation of Ar n – SH free-radical clusters via ultraviolet photolysis of
closed-shell Ar n – H 2 S complexes
S. R. Mackenzie, O. Votava, J. R. Fair, and D. J. Nesbitt
Citation: The Journal of Chemical Physics 110, 5149 (1999); doi: 10.1063/1.478410
View online: http://dx.doi.org/10.1063/1.478410
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/110/11?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Photodissociation of (SO2XH) Van der Waals complexes and clusters (XH = C2H2, C2H4, C2H6) excited at 32
040–32090 cm−1 with formation of HSO2 and X
J. Chem. Phys. 140, 054304 (2014); 10.1063/1.4863445
Intracluster stereochemistry in van der Waals complexes: Steric effects in ultraviolet photodissociation of state-
selected Ar–HOD/H 2 O
J. Chem. Phys. 120, 8443 (2004); 10.1063/1.1697394
Erratum: “An energy-resolved study of the partial fragmentation dynamics of Ar–HCl into H+Ar–Cl after ultraviolet
photodissociation” [J. Chem. Phys. 112, 4983 (2000)]
J. Chem. Phys. 115, 5692 (2001); 10.1063/1.1398309
Bond-breaking in quantum state selected clusters: Inelastic and nonadiabatic intracluster collision dynamics in
Ar–H 2 O→Ar+H ( 2 S)+ OH ( 2 Π 1/2,3/2 ± ;N)
J. Chem. Phys. 112, 7449 (2000); 10.1063/1.481344
An energy-resolved study of the partial fragmentation dynamics of Ar–HCl into H+Ar–Cl after ultraviolet
photodissociation
J. Chem. Phys. 112, 4983 (2000); 10.1063/1.481053
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

JOURNAL OF CHEMICAL PHYSICS
VOLUME 110, NUMBER 11
15 MARCH 1999
On the efficient generation of Arn –SH free-radical clusters via ultraviolet
photolysis of closed-shell Arn –H2S complexes
S. R. Mackenzie,a) O. Votava, J. R. Fair, and D. J. Nesbittb)
JILA, University of Colorado and National Institute of Standards and Technology, Boulder,
Colorado 80309-0440
͑Received 5 October 1998; accepted 3 December 1998͒
Detailed investigations are described for the generation of open-shell radical complexes formed by
unimolecular photolysis of closed-shell van der Waals precursors. As a specific test case, ultraviolet
photolysis of slit-jet cooled Arn –H2S (nр2) complexes at both 248 and 193 nm are shown to yield
Ar–SH and Ar2–SH radical cluster species with surprisingly high efficiencies. Analysis of the laser
induced fluorescence ͑LIF͒ spectra indicates that the radical complexes are produced with extensive
van der Waals stretch/bend and overall rotational excitation, which is consistent with a simple
ballistic model of the dissociation dynamics. The LIF spectra obtained as a function of expansion
distance downstream provide clear evidence for remarkably efficient cooling of the newly formed
radical cluster species by low-energy collisions with jet-cooled inert gas atoms at Ϸ10 K. Spectrally
resolved Ar–SH and Ar2–SH LIF signals have been investigated as a function of Ar composition,
which yields information on relative branching ratios for fragmentary ͑e.g.,
Ar2–H2S→Ar–SHϩHϩAr͒ and nonfragmentary ͑e.g., Ar2–H2S→Ar2–SHϩH͒ photolysis events.
© 1999 American Institute of Physics. ͓S0021-9606͑99͒00810-7͔
I. INTRODUCTION
͑For a comprehensive review of recent work on rare gas-
radical hydride complexes, see Ref. 21.͒ Much of this experi-
mental work has been made possible by techniques for high
density production of transient species under jet-cooled mo-
lecular beam conditions, spectroscopically sampling inter-
molecular potentials in excited states of these complexes.3–21
More recently, multilaser experiments involving stimulated
emission pumping ͑SEP͒ and fluorescence depletion have
been performed in order to extract similarly detailed infor-
mation on the ground state potential energy surface.22 These
efforts have proven quite successful and have yielded con-
siderable spectroscopic and dynamical information about this
class of open-shell clusters.21,23
These previous studies have all used qualitatively similar
methods of preparing open-shell species,3–21 specifically by
photolysis, electron impact, or electric discharge in the high
pressure region of a free-jet expansion. The radicals are
therefore formed initially quite hot, both translationally and
internally, but can be efficiently cooled by collisions in the
dense region of the expansion and clustered with rare-gas
diluent. The LIF spectroscopy is then obtained by interrogat-
ing the open-shell van der Waals species considerably further
downstream, typically many centimeters from the valve ori-
fice. Cluster rotational temperatures on the order of a few
Kelvin are routinely achieved, which collapses the resulting
spectra down to transitions from only the very lowest quan-
tum states. This obviously has important advantages for
spectroscopic analysis of these clusters but also restricts the
range of quantum states that can be accessed for dynamical
studies. For example, chemical reactions in clusters often
require sampling reaction barriers and intermolecular geom-
etries far from equilibrium, which even for the ‘‘floppiest’’
of intermolecular potentials may be poorly accessed
There has been much experimental and theoretical atten-
tion focused on the study of radical van der Waals
complexes.1–23 While some of these studies have been per-
formed on complexes of stable open-shell species such as
NO and O2 , recent improvements in experimental sensitivity
have also permitted work on clusters of highly reactive tran-
sient radicals such as OH, CN, SH, and CH. The special
interest in this class of open-shell clusters results from the
unusual nature of the bonding in such systems, which is an-
ticipated to be intermediate between ͑i͒ purely physical ͑i.e.,
electrostatic and dispersion͒ and ͑ii͒ chemical in character.1
Particularly tantalizing is the possibility of studying clusters
of chemically reactive open-shell intermediates,7 whereby
high resolution spectroscopy can provide new insights into
the entrance, exit, and transition state regions for fundamen-
tal radical reaction systems.
A particularly active area of research in the field of
open-shell clusters has been the electronic spectroscopy of
diatomic radicals bound to rare gas atoms. One of the first
such diatomic radical clusters successfully studied in the gas
phase was Ar–OH, nicely complementing the condensed
phase matrix laser induced fluorescence ͑LIF͒ work from a
decade earlier.3–5 Since then, the first electronic bands of
many van der Waals species of open-shell hydrides have
been investigated, including M–OH ͑OD͒,4–11 M–SH
͑SD͒,12–15 M–NH;16,17 M–CH,18–20 where MϭAr, Ne, Kr.
a͒
NATO Postdoctoral Fellow, 1996. Present address: Physical and Theoret-
ical Chemistry Laboratory, South Parks Road, Oxford, OX1 3QZ, United
Kingdom.
b͒
Staff member, Quantum Physics Division, National Institute of Standards
and
Technology,
Boulder,
CO
80309;
Electronic
mail:
djn@jila.colorado.edu
0021-9606/99/110(11)/5149/10/$15.00
5149
© 1999 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

5150
J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
by the lowest few eigenstates populated under jet-cooled
conditions.
An intriguing alternative method of synthesis of such
radical clusters is by direct photolysis of closed-shell precur-
sor clusters already formed in the supersonic expansion. One
potential advantage of such an approach is that the resulting
radical cluster can be generated with significant levels of
internal excitation. Indeed, unimolecular photolysis can re-
lease orders of magnitude more energy than necessary to
fragment the weakly bound cluster; thus, one might antici-
pate this to be an extremely inefficient process for forming
bound radical clusters. However, as this article demonstrates,
photolysis of hydride precursors proves to be surprisingly
efficient at removing excess energy from the weakly bound
species, yielding a remarkably high survival probability for
the resulting open-shell clusters.
The central idea of this method is as follows. The van
der Waals clusters of closed-shell hydrides with other closed-
shell subunits are formed and cooled in a slit supersonic
expansion, followed by ultraviolet ͑UV͒ pulsed laser pho-
tolysis of the radical precursor. In the minority of events,
direct intracluster collisions of the departing H atom with the
other subunit leads to prompt fragmentation of the weak van
der Waals bond. In most of the events, however, the H atom
escapes the cluster without directly sampling such hard
sphere, repulsive interactions. As a result of the large mass
disparity between the two photofragments, the H atoms carry
away the vast majority of the energy released. Thus, the re-
maining radicalϩclosed-shell subunit are left gently recoil-
ing with only a small fraction of the excess photolysis en-
ergy, which therefore promotes a much higher survival
probability for the nascent cluster.
In an earlier communication, we reported the prelimi-
nary results for such a synthesis method, which indicated
that open-shell van der Waals complexes can be produced
with surprising efficiency by the UV photolysis of appropri-
ate closed-shell cluster.24 In particular, we demonstrated the
production of Arn –SH radical complexes from small
Arn –H2S clusters by the following unimolecular reactions:
FIG. 1. Schematic representation of the experimental apparatus used in
these studies. H2S ͑0.5%͒ and Ar ͑2%–10%͒ seeded in first run neon ͑30%
He, 70% Ne͒ are expanded through a slit jet forming closed-shell van der
Waals complexes. These complexes are photolyzed 1.3 cm downstream and
the free-radical clusters produced are probed with a tunable laser collinear
with the photolysis laser. Laser induced fluorescence produced as the probe
laser is scanned is collected on two cylindrical mirrors and detected on a
PMT.
nonfragmentary channels, for which no Ar atom is ejected as
a third body, can be efficient pathways for the production of
highly internally excited, albeit bound radical complexes.
The present work elaborates on the previous study, specifi-
cally ͑i͒ obtaining the spectra of the radical clusters formed
by such a photolysis event ͑ii͒ characterizing the efficiency
of such a gentle recoil process, and ͑iii͒ extracting informa-
tion on the subsequent energy transfer dynamics due to low-
energy collisions with jet-cooled Ar atoms.
The remainder of this article is organized as follows. In
Sec. II, details of the experimental apparatus for gentle recoil
photolysis of open-shell radical clusters is described. The
spectroscopic results for Arn –SH clusters are presented in
Sec. III, followed by estimates of the photolysis efficiency in
Sec. IV. The collisional relaxation dynamics of these radical
clusters with jet cooled Ar atoms are investigated in Sec. V,
while the role of fragmentary ͑i.e., Ar ejecting͒ versus non-
fragmentary ͑Ar conserving͒ photolysis dynamics are briefly
evaluated in Sec. VI. Section VII presents some directions
for future work and concluding remarks.
Arn –H2Sϩh␯→Arn –SHϩH,
͑1͒
͑2͒
Arn –H2Sϩh␯→ArnϪ1 –SHϩArϩH,
where the former process leaves the radical cluster intact,
while the latter process results in multiple ͑three body or
more͒ continuum channels being accessed. Throughout this
article, we will distinguish these two classes of events as ͑1͒
nonfragmentary and ͑2͒ fragmentary pathways. This previ-
ous report concentrated on the dynamical implications of
such events and proposed a simple ‘‘gentle-recoil’’ mecha-
nism by which it might proceed. Under this scheme, the H
atom, as a result of its small mass relative to that of the SH
diatom, carries away у95% of the excess energy released in
the photolysis event itself. The effective recoil energy of the
SH moiety is reduced even further by subsequent transfor-
mation into the ArnϩSH center-of-mass frame, down to a
low enough energy to permit some fraction of the Arn –SH
clusters to remain bound. In particular, this simple ballistic
model of the dissociation dynamics predicts that even the
II. EXPERIMENT
The experimental apparatus used in these studies is
shown schematically in Fig. 1. A mixture of H2S ͑0.5%͒ and
Ar ͑2%–10%͒ in a balance of first-run neon ͑70% Ne: 30%
He͒ at backing pressures in the range of 100–1200 Torr ͑1
Torrϭ133.32 Pa͒ is expanded through a 10 Hz pulsed slit
valve ͑4 cmϫ50 ␮m͒ into an evacuated chamber. The slit
valve has been described in detail elsewhere and has been
shown to be an efficient source for spectroscopy and dynam-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
5151
ics of van der Waals clusters.25 The photolysis laser used in
these experiments is an excimer laser operating with either
KrF ͑248 nm, 10 mJ͒ or ArF ͑193 nm, 2 mJ͒. The tunable
probe laser is a frequency doubled dye laser operating with-
out intracavity etalon ͑linewidth ϳ1 cmϪ1͒ pumped by a Nd
yttrium–aluminum–garnet ͑YAG͒ laser. With a mixture of
DCM/LDS 698 dyes, more than 1 mJ light is available in the
doubled 320–332 nm region ͑the 0←0 band of the SH
A 2⌺←X 2⌸ transition͒, although in practice this is reduced
with neutral density filters by an additional 2 to 3 orders of
magnitude to avoid optical saturation effects.
The two collinear and counter-propagating lasers are ad-
mitted to the diffusion pumped vacuum chamber via baffle
arms, which are fitted with adjustable irises to reduce the
scattered light in the interaction region. The intersection re-
gion between the laser and expansion is Ϸ1.3 cm ͑Ϸ260 slit
widths͒ downstream of the valve and parallel to the 4 cm slit
axis, by which point clustering within the beam has long
been completed. By tuning the laser, LIF is generated from
both bare SH as well as Arn –SH clusters. The resulting fluo-
rescence is focused onto a photomultiplier tube ͑PMT͒ with
high efficiency cylindrical optics, with the spectra recorded
by monitoring the total fluorescence level ͑or time-gated
components thereof͒ as the probe laser is scanned. Optical
notch filters in front of the PMT are used to reduce the scat-
tered light arising from the excimer pulse.
FIG. 2. LIF temporal profiles captured on a digital storage oscilloscope. ͑a͒
Small scattered light signal arising from Rayleigh scattering of the probe
beam ͑recorded in the absence of the photolysis laser͒. ͑b͒ LIF signal arising
from open-shell Arn –SH species showing relatively long fluorescence de-
cay. The exact lifetime of the fluorescence is dependent on the upper state
2
ϩ
accessed ͑Ref. 12͒. ͑c͒ Rapid predissociation of the SH A
⌺
( ϭ0) state
v
results in a short fluorescence lifetime for the SH monomer. In order to
display the SH and Arn –SH signals on the same vertical scale, the trace in
͑c͒ is recorded on a very weak SH monomer transition.
The majority of the experiments discussed below are
performed with a probe laser delay with respect to the pho-
tolysis pulse of typically 10–100 ns. This permits the nascent
distribution of photofragments to be interrogated on a time
scale several orders of magnitude shorter than the binary
collision rate. Furthermore, additional experiments have
been performed with much longer delays ͑up to у10 ␮s͒
between photolysis and probe pulses in order to observe the
subsequent cooling of the nascent photofragment distribution
by binary collisions with jet-cooled Ar gas. For these latter
studies, the geometry of the experimental apparatus is modi-
fied to displace the photolysis laser 1–5 mm upstream of the
probe laser in order to allow the products to fly into the
collection region and be detected. This choice of adjustable
photolysis or fixed probe laser geometries allows one to
maintain maximum detection sensitivity for LIF detection
along the axis of the cylindrical mirror collection system.
A typical LIF temporal profile obtained in these studies
is shown in Fig. 2. Scattered light from both the probe and
the photolysis lasers gives rise to small peaks at tϭ0 that
arise not from fluorescence but rather from Rayleigh scatter-
ing and/or finite scattered window light rejection from the
components disappear in the absence of the excimer laser
pulse ͓as shown in Fig. 2͑a͔͒, indicating that additional pho-
tolysis by the probe laser is insignificant at these laser pow-
ers. The long lifetime fluorescence component has been ob-
served in previous studies11 and on the basis of rotationally
resolved spectra at higher resolution has been attributed to
small clusters of SH such as Ar–SH and Ar2–SH.12–15
The explanation for the vastly different fluorescence life-
times of the SH monomer and Arn –SH excited states is pro-
vided in Fig. 3. The SH monomer A 2⌺ϩ state ͑bold line͒ is
crossed in the region of the zero-point energy by a repulsive
a 4⌺Ϫ state that correlates with ground state products. Nona-
diabatic coupling to this state is believed to be responsible
for the rapid predissociation of the excited SH radical, result-
ing in the laser limited fluorescence decay profile observed
for the monomer species ͓Fig. 2͑c͔͒. Clustering with even a
single Ar atom has a profound effect on these potential en-
ergy curves, particularly that of the excited state. For ex-
ample, extensive previous work on Ar–OH indicates that the
presence of the Ar atom produces a dramatic increase in the
binding energy upon electronic excitation.3–10 In the case of
Ar–SH this has the important effect of lowering the A-state
potential ͑dashed line͒ with respect to the dissociative state,
thereby energetically shutting off the predissociation channel
for nonradiative loss of the excited state. Consequently, the
Ar–SH A 2⌺ϩ state fluorescence slows down dramatically,
to a rate approaching the intrinsic SH radiative lifetime
͑ϳ800 ns͒ determined from gas phase absorption and matrix
isolation studies. This quenching of predissociation in
Ar–SH has been the subject of a detailed jet-cooled LIF
baffle arms. The SH A2 ⌺(0)←X 2⌸( ϭ0) transition is
v
pumped in the spectral region from 320 to 332 nm, the ex-
cited state of which has been shown to have a fluorescence
lifetime that is short ͑i.e., 1–3 ns͒ with respect to the 10 ns
laser pulse duration.26–31 As a result, when the probe laser is
resonant with an SH monomer transition, an intense peak
appears promptly in the LIF temporal profile, with a lifetime
too short to be temporally resolved ͓Fig. 2͑c͔͒. In addition to
this contribution, there is a second component ͓Figs. 2͑b͒ and
2͑c͔͒ whose fluorescence lifetime is considerably longer than
that of the monomer. Both prompt and delayed fluorescence
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

5152
J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
FIG. 3. Approximate potential energy curves for the ground and first excited
2
ϩ
states of SH and Ar–SH showing the A
⌺
←X 2⌸ electronic transition. In
the case of the SH monomer ͑upper state bold line͒, the repulsive 4⌺Ϫ state
FIG. 4. Total backing pressure dependence of both fluorescence compo-
nents. ͑a͒ The monomer LIF signals increase linearly with pressure until
about 1000 Torr, above which slight deviations from linearity are evident
due to loss of H2S monomer. ͑b͒ The Arn –SH signals increase much more
rapidly than linearly in backing pressure, consistent with the expected pres-
sure dependent behavior for cluster formation. In each case the composition
was 0.5% H2S/10% Ar in a balance of first run Ne.
crosses the excited state in the vicinity of the ϭ0 level. Adiabatic coupling
between these states leads to predissociation of the A state. The effect of
clustering is to produce a remarkable increase in binding energy upon elec-
v
tronic excitation with the result that the Ar–SH A ( ϭ0) state ͑dashed line͒
v
lies sufficiently below the crossing that it survives predissociation to fluo-
resce at a rate approaching the inherent SH radiative lifetime.
study by Yang et al.,12 in which the fluorescence lifetimes of
the first five van der Waals vibrational levels have been de-
termined.
There have been previous indications of such radical
cluster photolysis channels in the literature. For example,
evidence for transient open-shell I–HI complexes formed by
unimolecular photolysis of (HI)n clusters has been observed
in H atom energy distributions from high-Rydberg time-of-
As a result of this strongly cluster dependent predisso-
ciation dynamics, the spectra of the SH monomer and
Arn –SH cluster species can be readily distinguished from
one another by time gating the fluorescence signals. By way
of example, Fig. 4 shows the dependence on total backing
pressure of the LIF signals obtained by gating regions I and
II ͑see Fig. 2͒, for a sample mixture of H2S ͑0.5%͒, Ar ͑5%͒
in a balance of first-run Ne. The integrated intensity of the
SH signal ͓I, Fig. 4͑a͔͒ increases nearly linearly with stagna-
tion gas pressure until about 1000 Torr, after which it
slightly deviates from linearity due to clustering of the H2S
monomer. Conversely, the open-shell Arn –SH signals ob-
served in gate II ͓Fig. 4͑b͔͒ grow in with a nonlinear depen-
dence in backing pressure, i.e., characteristic of signals aris-
ing from cluster species. This behavior can be easily
rationalized by noting that one mechanism for forming
weakly bound Ar–H2S clusters requires three-body colli-
sions of H2S with Ar atoms, the total rate of which ͑for a
fixed dilution͒ increases roughly with the cube of the total
gas density. Alternatively, these clusters may be formed by
flight experiments.32 In addition, Garcıa–Vela has recently
´
published a theoretical study of the dissociation dynamics of
Ar–HCl, in which the partial fragmentation channel yielding
Ar–Cl was found to represent the dominant channel at high
energy.33 Furthermore, the theoretical dynamics calculations
by Bowman and co-workers on Ar–H2O suggest that H atom
ejection to form Ar–OH may play a significant channel in
the photolysis event.34,35 In a dynamically related direction,
Janda and co-workers have intriguing indirect evidence for
intact photoejection of weakly bound Ne2 from photolysis of
Ne2Cl2 clusters.36 To the best of our knowledge, however,
these studies provide the first direct experimental detection
and spectroscopic characterization of the resulting weakly
bound radical clusters formed by photolysis of closed-shell
van der Waals precursor species.
III. SPECTROSCOPIC RESULTS AND ANALYSIS
preferential
displacement
reactions
such
as
Figure 5 shows the LIF spectrum of the long-time fluo-
rescence decay component ͑gate II in Fig. 2͒ as a function of
the probe laser wave number for an 800 Torr expansion of a
0.5% H2S: 10% Ar in mixture in first run neon. At the reso-
lution of the UV probe laser, the ‘‘nascent’’ collision-free
spectrum consists of a largely unresolved, strongly blue
Ar2ϩH2S→Ar–H2SϩAr, which would be expected to scale
more rapidly with total gas density than P3. In any event, this
long time fluorescence component in gate II can be unam-
biguously attributed to photolysis of van der Waals clusters
formed in the molecular expansion.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
5153
FIG. 5. LIF spectrum of the nascent distribution of open-shell complexes
produced by unimolecular photolysis of bound closed-shell van der Waals
complexes. The assignments shown above are taken from Yang et al. ͑Refs.
12,13͒. From the much colder spectra obtained by collisional cooling with
Ar atoms in the slit jet, the peaks attributed to larger clusters are almost
certainly due to Ar2–SH ͑see the text͒.
FIG. 6. Effect of delaying the probe laser pulse on the spectrum of the
open-shell cluster species. No significant changes are observed in the spec-
trum for delays up to 800 ns, indicating that any effects due to two-body
collisional cooling or three-body secondary clustering are negligible on this
time scale.
bimolecular collision rate for SH radicals with Ar atoms
would only be 1 per 300 ns, which is already substantially
longer than the ⌬␶ϭ10–100 ns delay between photolysis and
probe lasers. More relevant to the possibility of subsequent
cluster formation downstream of the photolysis, the time
scale for three-body collisions is several orders of magnitude
longer ͑i.e., roughly 1 every 10 ms͒. Given that at least one
three-body collision is required for cluster stabilization, one
can safely conclude that ArϩHSϩM→Ar–SHϩM pro-
cesses, which would skew the results after the photolysis
event, represent a completely negligible channel on the ⌬␶
ϭ10–100 ns time scale.
By way of experimental verification, the spectrum of the
open-shell cluster species has been recorded as a function of
excimer-probe delay times, for a range of delays from ⌬␶
ϭ50–800 ns ͑see Fig. 6͒. If the radical cluster signals were
the result of photolysis of H2S followed by collisionally me-
diated clustering of HS radicals, then the overall signal in-
tensity would increase more or less linearly with photolysis-
probe delay. As demonstrated in Fig. 6, there is no evidence
for such an increase. In fact, no significant changes are evi-
dent in the spectrum for delays up to у800 ns, indicating
that secondary clustering of photofragments makes a negli-
gible contribution to the overall signals at early times.
Since unimolecular photolysis is responsible for produc-
tion of the open-shell species observed, it is of interest to
characterize the size of the precursor clusters from which
they are formed. In order to do this, the growth in the open-
shell cluster signal as a function of Ar and dilution of H2S
concentration and total backing pressure has been investi-
gated. One surprising result of these studies is the dynamic
range of backing pressure and dilution of H2S and Ar con-
centration over which the cluster LIF signals are observed.
For example, the characteristic fluorescence spectrum with
associated asymmetric band profile is clearly visible in mix-
tures as lean as 0.5% H2S/2% Ar, even down to total stag-
nation pressures of only a few hundred Torr. Based on ex-
tensive high resolution studies of Arn –HF and Arn –HCl
shaded band. However, several sharp features are clearly vis-
ible, which provide important evidence of the identity of the
species produced. Specifically, the prominent progression of
bands at 30 609, 30 673, 30 730 cmϪ1 etc. can be readily
identified by comparison with previous jet-cooled studies as
the A 2⌺ϩ( Ј )ϪX 2⌸( Љ ϭ0) excited state van der
vvdw
vvdw
Waals progression in Ar–SH.12–15 However, even allowing
for differences in laser linewidth, these same features appear
at least an order of magnitude broader in the present study
than in those of Miller and co-workers.12–15 Furthermore, as
will be shown later in Sec. V, these linewidths narrow by an
order of magnitude upon subsequent collisions with Ar at-
oms in the jet expansion. This provides a strong indication
that the complexes are initially produced with high levels of
end-over-end rotational excitation, which is of course consis-
tent with the gentle-recoil model.24
A second series of peaks is visible in Fig. 5 at 30 580,
30 645, 30 695 cmϪ1 etc. These peaks were also observed in
previous jet-cooled studies,12–14 where it was noted that this
second series exhibited more congested rotational structure
at high resolution and grew in with Ar concentration at a
more rapid rate than the rotationally well resolved Ar–SH
bands. As a result, this second progression was tentatively
interpreted by Yang et al. as the corresponding A–X elec-
tronic transition in Ar2–SH trimer. Such an assignment is
entirely consistent with the present results, and also implies a
finite population of Arn –H2S species with nϭ2 or more
present in the supersonic expansion. The point of specific
relevance to the current study is that the trimer ͑and larger͒
clusters can also have additional fragmentary channels ͑i.e.,
the ‘‘boiling off’’ of Ar atoms͒ by which open-shell clusters
can be formed.
One central result of this study is that the open-shell
species arise entirely from unimolecular processes. Specifi-
cally, the photolysis event takes place Ϸ260 slit widths
͑Ϸ1.3 cm͒ downstream of the nozzle orifice; even for a con-
servatively large hard sphere cross-section of 100 Å2, the
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

5154
J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
luster ArϩSH collision, the departing H atom carries away
Ϸ97% of this excess energy. Furthermore, there is an addi-
tional Ϸ twofold loss in transforming to the Ar–SH center-
of-mass ͑cm͒ frame, which leaves only an extremely small
fraction ͑Ϸ1.6%͒ of the initial energy to be accommodated
in the cluster.
To address this issue more quantitatively, we have per-
formed the following experiments to estimate the efficiency
of this gentle recoil process. First of all, based on previous
high resolution studies on Ar–H2O clusters in which the ab-
solute clustering efficiency can be directly determined by IR
laser absorption measurements,43 we estimate that roughly
2% of the initial H2S forms closed-shell complexes ͑pre-
dominantly Ar–H2S or Ar2–H2S͒ under typical expansion
conditions for these studies. This is also corroborated ͑in Fig.
4͒ by the onset of slight deviations away from linearity ex-
perimentally observed for SH LIF signals with total backing
pressure. Second, by recording spectra of the laser induced
fluorescence from SH monomer and cluster Arn –SH species,
one can determine the relative integrated intensities of each
component. More specifically, based on equivalent oscillator
strengths, phototube detection efficiencies, etc. the sensitivity
to SH monomer signals ͑with fluorescence decay lifetime
FIG. 7. Dependence on %Ar composition of the total integrated cluster LIF
signal. Each point is obtained at a total backing pressure of 1000 Torr. For
all other expansion conditions held fixed, the density of Arn –H2S species in
the perturbative clustering limit should increase with Ar n. Based on this
proportionality, the slope of the log–log graph reflects the effective number
of Ar atoms contained in the Arn –H2S complex being photolyzed.
͓
͔
␶
͒ and Arn –SH cluster signals ͑with fluorescence decay
mon
lifetime ␶ ͒ in a detection gate ⌬␶ from 0 to td can be
approximated as
clusters ͑nϭ1, 2, 3, and 4͒ in slit jet expansions,37–40 one
would expect these dilute H2S/Ar conditions and low back-
ing pressures to yield predominantly binary clusters,
Ar–H2S, with perhaps some additional Ar2–H2S trimers
forming at higher pressures. By way of supporting evidence,
Fig. 7 shows the dependence of the integrated radical cluster
LIF signal on %Ar concentration, at a total backing pressure
of 1000 Torr for 248 nm photolysis. For all other total pres-
sures and concentration conditions held fixed, previous
studies41 have demonstrated that M –X is roughly propor-
clust
t
t
0
͐
͐
d exp Ϫ
dt
dt
ͩ
ͪ
ͪ
Nc0lust
Nm0 on
Iclust
␶
␶
mon
ϭ
,
͑3͒
Imon
t
t
0
d exp Ϫ
ͩ
clust
where N0mon and N0clust are the number of SH monomer and
Arn –SH open-shell cluster species produced by photolysis,
respectively, and Imon and Iclust are integrated band intensities
over the monomer and cluster signals.
͓
͔
n
n
tional to M , in the limit of small fractional clustering.
͓
͔
In order to use Eq. ͑3͒, some estimate of the fluorescence
decay lifetimes for each component is required. In the last
decade there has been considerable discussion in the litera-
Hence, the number of Ar atoms contained in the precursor
cluster can be roughly estimated from a log–log graph of the
fluorescence signal intensity versus ͓Ar͔, as shown in Fig. 7.
Even at the highest pressures sampled in these studies
͑Ϸ1000 Torr͒, Fig. 7 demonstrates LIF signals varying be-
tween linear and quadratic in Ar concentration in the range
of ͓Ar͔ϭ1%–10%. This would indicate that the predominant
closed-shell species leading to LIF signals under these ex-
pansion conditions are Ar–H2S and Ar2–H2S, although the
possibility of trace amounts of larger cluster species cannot
be excluded.
2
ϩ
ture concerning the lifetime of the SH
⌺
( ϭ0)
v
level.26–31 It is now widely accepted that the fluorescence
lifetime is completely limited by predissociation for the rea-
sons outlined above ͑see Fig. 3͒. Specifically, Ubachs and ter
Meulen31 have recently produced a definitive study in which
the fluorescence lifetimes of the first 10 rotational levels of
SH A 2⌺ϩ ( ϭ0) have been reliably determined to range
v
from 3.2Ϯ0.3 ns ͑for Nϭ0͒ down to ϳ0.9 ns ͑for Nϭ9͒.
The effective fluorescence lifetime for Arn –SH cluster
species is somewhat more involved to estimate. Yang et al.12
measured the fluorescence lifetimes of the Ar–SH A 2⌺ϩ
IV. EFFICIENCY OF GENTLE RECOIL SYNTHESIS
͑
vdwϭ0, 1, 2, 3, and 4͒ vibrational states and found a strong
v
It seems initially counterintuitive that a weakly bound
van der Waals cluster could survive the UV photodissocia-
tion of one of the subunits, which in the case of H2S at 248
nm releases roughly 2 orders of magnitude more energy than
required to break the van der Waals bond.42 However, as
argued previously, a simple gentle-recoil model of such a
photolysis event provides a very different perspective.24
Even for a worst case scenario where the event can be con-
sidered as ͑i͒ the photolysis of essentially free H2S in a weak
Ar–H2S van der Waals complex followed by ͑ii͒ an intrac-
reduction in fluorescence lifetime with increasing van der
Waals excitation, from ␶fϭ600Ϯ50 ns for vdwϭ0 to ␶
ϭ49Ϯ20 ns for vdwϭ4. This trend is easily rationalized,
v
f
v
since the high van der Waals stretch excited states should
asymptotically behave more like an isolated HS monomer
with Ϸ1 to 3 ns predissociation limited lifetimes. Based on
this reasoning, the fluorescence lifetime for Ar2–SH should
be even longer than that of Ar–SH since it models a more
extensive ‘‘cage’’ of Ar atoms. In support of this model, the
fluorescence lifetime for SH A 2⌺ϩ ( ϭ0) in an Ar matrix
v
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
5155
TABLE I. Estimated efficiency ͑in %͒ for the generation of open-shell com-
plexes by unimolecular photolysis of closed-shell van der Waals clusters.
These calculations assume ͑i͒ an overall 2% precursor clustering efficiency,
based on previous values experimentally determined for Ar–H2O clusters
and ͑ii͒ a range of estimates for A-state fluorescence lifetimes in both the SH
monomer and the open-shell cluster ͑see text for details͒.
␭
193 nm
248 nm
photolysis
␶ ͑Ar–SH͒
␶ ͑SH͒ϭ1 ns
␶ ͑SH͒ϭ3 ns
800 ns 400 ns 50 ns 800 ns 400 ns
50 ns
Ϸ100%a
Ϸ100%a
1.5%
4%
2%
6%
11%
32%
15%
46%
20%
60%
aThe LIF intensities would imply more than 100% efficiency for forming
open-shell radical clusters and are therefore not consistent with this choice
of fluorescence lifetimes.
has been measured to be 425Ϯ20 ns at 17–32 K.29 When
corrected for Stokes shift and matrix dielectric constant, this
translates into a pure radiative lifetime of 700 ns, i.e., in
good agreement with the 820Ϯ240 ns value estimated by
Friedl et al.26 Thus, for the distribution of Ar–SH and
Ar2–SH internal states populated by photolysis, one might
reasonably expect lifetimes ranging from this upper limit of
Ϸ800 ns down to the fastest observed values of Ϸ50 ns.
The LIF spectral data for photolysis of H2S/Arn –H2S at
248 and 193 nm has been analyzed according to Eq. ͑3͒,
based on the estimated range of Arn –SH ͑i.e., 800–50 ns͒
and SH A 2⌺ϩ ͑i.e., 1–3 ns͒ fluorescence lifetimes noted
above. The resulting values are listed in Table I and under-
score two important points. First of all, the probability of
forming open-shell clusters is surprisingly high, with even
the lowest estimate at 248 nm in excess of 15%. Second, the
predicted efficiencies are roughly tenfold lower for photoly-
sis at 193 nm than at 248 nm. This qualitative trend is ex-
pected, since in the gentle recoil model, there is Ϸ twofold
more excess energy to accommodate at the shorter wave-
length. It is worth stressing that the magnitude of this shift
with wavelength is independent of all the above assumptions
of clustering efficiency and lifetime estimates and thus
should represent a rigorous prediction.
FIG. 8. The effect of collisional cooling on the LIF spectrum. Extended
delays between the photolysis pulse and the probe allow the nascent distri-
bution of photolysis product quantum states to relax with remarkably high
efficiency via low energy collisions with the cold inert gas diluent atoms.
300 ns͒ this offers the opportunity to systematically probe
the spectral evolution of these internally hot clusters as a
function of number of collisions with low temperature rare
gas diluent atoms ͑Ϸ10 K͒ in the slit-jet expansion.
Figure 8 shows the effects of increased delay between
the photolysis pulse and the probe on the LIF spectrum.
Whereas essentially no spectral changes are observed for
short delays ͑50–800 ns, see Fig. 6͒ the spectra taken at
much longer delays ͑1–10 ␮s͒ demonstrate a dramatic nar-
rowing of rotational band contours and decrease of spectral
congestion. Two quantitative trends are clear from Fig. 8.
First of all, as the delay increases, there is a systematic de-
crease in intensity at the blue end of the spectrum. According
to our modeling of the Franck–Condon factors, this is where
the majority of hot band transition intensity resides, which
would therefore be consistent with rapid collisional cooling
of these excited internal state populations. What is especially
worth noting is the efficacy of such a cooling process; for a
10 ␮s delay, a typical radical cluster would only have expe-
rienced approximately 30 collisions, yet this is apparently
sufficient to collapse the spectrum from essentially the ‘‘high
temperature’’ limit to a spectrum more consistent with Ϸ10
K. It is clear that the low energy collisions in the slit jet
expansion with slow inert gas diluent ͑He, Ne, Ar͒ atoms
must be remarkably efficient at removing residual energy left
in the intermolecular modes of the weakly bound radical
clusters. Interestingly, this is reminiscent of the high energy
transfer efficiencies noted for vibrational relaxation of highly
excited I2 with low energy collisions with rare gas diluent far
downstream in supersonic jets.44–46
V. SUBSEQUENT COLLISIONAL COOLING OF
NASCENT Arn –SH RADICAL COMPLEXES
The ballistic picture for this gentle recoil synthesis
mechanism clearly predicts the open-shell clusters to be pro-
duced with high internal and end-over-end rotational
excitation.24 In our earlier report, the gross features of the
LIF spectrum could be simulated by fitting the van der Waals
coordinate of both the ground and excited states of Ar–SH to
one-dimensional Morse potentials and calculating Franck–
Condon factors for the transitions between the two surfaces.
The simplest ‘‘high temperature’’ model of states equally
populated up to the dissociation limit yielded a reasonable fit
to the frequencies and nicely reproduced the characteristic
band contour asymmetries in the experimental spectrum. To
further elucidate the collisional dynamics of such highly ex-
cited radical cluster species, several additional experiments
have been performed at extended delays between the pho-
tolysis and probe pulses. Since the bimolecular collision
rates in the slit jet can be estimated reliably ͑roughly 1 per
Second, at these longer delays, the peaks in the nascent
spectra are strongly enhanced with respect to the background
spectral congestion due to rotation. If one considers a typical
peak in isolation, e.g., the A 2⌺(
ϭ3)ϪX 2⌸( vdwϭ0)
vvdw
v
at 30 673 cmϪ1, it is evident that the rotational contour nar-
rows considerably in the first 5 ␮s of delay, down to widths
comparable to those seen in the jet cooled studies12,13 of
Yang et al. at a few Kelvin. The rotational contours of indi-
vidual vibronic transitions are now clearly visible and serve
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

5156
J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
to further confirm the spectral assignment offered in Sec. III.
Most importantly, all progressions in the long time jet spec-
tra can be readily assigned to either Ar–SH or Ar2–SH radi-
cal clusters. This proves useful for the analysis in the follow-
ing section, which attempts to elucidate the ratio of
fragmentary versus nonfragmentary photolysis pathways
forming these two species.
It is worth briefly considering whether this cooling oc-
curs by ‘‘selective predation’’ of hot Arn –SH species near
the dissociation limit, or by collisional removal of internal
energy from the clusters. With a frequency doubled probe
laser line width of Ϸ1 cmϪ1, rotational structure in the LIF
spectra is not resolved, even at the longest of time delays and
lowest of cluster temperatures. However, at these longest
time delays, the LIF rotational contours exhibit a full-width
at half-maximum ͑FWHM͒ of ϳ2.3 cmϪ1. At the present
resolution, this would be consistent with the ϳ1.8 cmϪ1
FWHM band head structures observed at a few Kelvin by
Yang et al. as well as collisional equilibration with diluent
inert gas in the 10 K slit jet expansion. In any event, the
mean energies ͑kT Ϸ5–10 cmϪ1͒ are an order of magnitude
below that necessary to bind the cluster, which would not be
consistent with a purely predative mechanism. Thus, though
selective destruction of the hottest clusters may in fact be
responsible for some of the initial loss, there is clearly a
significant component that occurs via inelastic energy re-
moval dynamics with the cold jet gas.
Furthermore, one might question whether this time de-
pendent cooling has any appreciable component due to slow
unimolecular decomposition of Ar–SH or Ar2–SH clusters
above the dissociation limit. However, from both Rice–
Ramsperger–Kussel–Marcus ͑RRKM͒ models47 and classi-
cal trajectory studies, such unimolecular decomposition of
Ar–SH and Ar2–SH species would take place very much
faster than the ␮s time scales observed. Since these unimo-
lecular decomposition rates slow dramatically with cluster
size, there could be possible contributions on the ␮s time
scale from Arn –SH clusters with nϾ2, though based on the
fractional Ar concentration dependence of these signals, such
effects would be appear to be very small.
FIG. 9. The effect of Ar concentration on the LIF spectrum. The peaks
tentatively assigned as transitions in higher order Ar2–SH clusters grow in
more rapidly with %Ar content than do the Ar–SH transitions, in good
agreement with the analysis of Yang et al. In order to observe these effects,
the spectra are recorded with a 10 ␮s photolysis probe delay, which allows
efficient collisional cooling to take place with cold Ar atoms in the slit jet
expansion.
cluster species that yield bare SH radical, but these can be
readily discriminated against in the LIF spectra by time gated
detection in region II, i.e., at times greater than the monomer
predissociation window. Furthermore, as demonstrated in the
previous section, it is easy to identify peaks in the cluster
LIF spectrum at long delay times arising from Ar–SH and
Ar2–SH, respectively. The relative branching ratios between
these two species can then be investigated as a function of
fractional Ar concentration in the stagnation region, holding
the LIF probe detection geometry and all other expansion
conditions constant.
Figure 9 shows the free-radical cluster LIF spectrum re-
corded at a 10 ␮s probe delay time for a series of different
Ar compositions photolyzed at 248 nm, with a similar data
set obtained at 193 nm. To make visual comparisons easier,
the spectra have all been normalized to the integrated inten-
sity of assigned Ar–SH peaks. Such a scaling makes it im-
mediately apparent that the second series of peaks due to
Ar2–SH grows in considerably more rapidly with Ar concen-
tration than does the Ar–SH progression. This is consistent
with these transitions arising from Arn –SH species with n
Ͼ1, in good agreement with the original assertion of Yang
et al.12,13
VI. FRAGMENTARY VERSUS NONFRAGMENTARY
PHOTOLYSIS DYNAMICS
One final issue posed by this study is the role of frag-
mentary versus nonfragmentary processes in the photolysis
event. For example, photolysis of Ar2–H2S trimer and
Ar–H2S dimer can occur by different pathways and be de-
tected via LIF as Ar–SH or Ar2–SH radical clusters:
If we assume equivalent UV photolysis cross-sections
for Ar–H2S and Ar2–H2S, one can readily show that the
ratio of ͓Ar–SH͔ and Ar –SH free-radical species ob-
͓
͔
2
served is
Ar–H2Sϩh␯→HϩAr–SH,
Ar2–H2Sϩh␯→HϩAr2–SH,
Ar2–H2Sϩh␯→HϩArϩAr–SH,
͑4a͒
͑4b͒
͑4c͒
Ar–SH͔ Pa Ar–H S ϩP Ar –H S
͔
2
͓
͓
͔
͓
2
c
2
ϭ
,
͑5͒
͓Ar2–SH͔
Pb Ar –H S
͓ ͔
2 2
where Pa , Pb , and Pc represent the probabilities for the
three processes in Eq. 4͑a͒–4͑c͒, respectively. To make use
of Eq. ͑5͒, one must be able to relate the concentrations of
the two Ar–H2S and Ar2–H2S photolysis precursors to the
experimental variable, ͓Ar͔. Based on previous cluster work
with high resolution IR laser absorption measurements, the
where the first two pathways represent nonfragmentary pho-
tolysis channels that eject only a H atom and the third rep-
resents a fragmentary photolysis channel where an additional
Ar atom is also ‘‘boiled off’’ to stabilize the resulting cluster.
There will of course also be fragmentary channels for both
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
5157
TABLE II. Fragmentary vs nonfragmentary photolysis dynamics from a
weighted least-squares fit of Eq. ͑3͒ to the data in Fig. 10. The ͑i͒ slopes and
͑ii͒ intercepts at the two different wavelengths are proportional, respectively,
to ͑i͒ the ratio of fragmentary to nonfragmentary photolysis channels in
Ar2–H2S and ͑ii͒ the ratio of nonfragmentary photolysis channels between
Ar–H2S and Ar2–H2S ͑see text for details.͒
Intercepta
Slopeb
248 nm
0.8
0.14
193 nm
248 nm/193 nm ratio
2.1
0.4͑2͒:1
0.064
2.1͑3͒:1
aProportional to the ratio of fragmentary to nonfragmentary photolysis of the
Ar2–H2S species.
bProportional to the ratio of nonfragmentary photolysis channels for Ar–H2S
and Ar2–H2S.
wavelength, which in turn requires a more efficient removal
of the excess energy in the photolysis event to stabilize the
cluster below the dissociation limit.
Similarly, the slopes derived from Fig. 10 ͑see Table II͒
are proportional to the probability ratio for nonfragmentary
photolysis of Ar–H2S (ϰPa) and Ar2–H2S (ϰPb) com-
plexes. First of all, these slopes are nonzero at both wave-
lengths, which indicates a finite probability for Ar–H2S clus-
ters surviving UV photolysis to yield Ar–SH free-radical
complexes. Second, the slopes decrease by more than two-
fold between 248 and 193 nm, which indicates that the im-
portance of nonfragmentary pathways in Ar2–H2S increases
with respect to that of Ar–H2S with increasing photolysis
energy. Taken in isolation, there are several scenarios con-
sistent with this result. However, in conjunction with the
previously determined tenfold decrease in open-shell cluster
generation efficiency between 248 and 193 nm photolysis,
the only possibility is that ͑i͒ each of the probabilities are
decreasing functions of photolysis energy, but that ͑ii͒ non-
fragmentary photolysis of Ar2–H2S decreases more slowly
than nonfragmentary photolysis of Ar–H2S.
FIG. 10. Relative Ar–SH͔/͓Ar –SH signals as a function of reciprocal Ar
͓
͔
2
concentration following ͑a͒ 248 and ͑b͒ 193 nm photolysis. Data from the
weighted least-squares fit is collated in Table II and provides information on
the relative branching ratios of fragmentary and nonfragmentary photolysis
as a function of photolysis energy ͑see the text for details͒.
ratio of adjacent oligomer populations ͕i.e., Ar
͓
nϩ1
ϪX͔/͓Ar ϪX ͖ in the weak clustering limit is proportional
͔
n
to the Ar density. In the present system, this is equivalent to
assuming that
Ar –H S͔/͓Ar–H S͔Ϸ␤͓Ar
͑6͒
͓
͔
2
2
2
Two simple physical pictures nicely reinforce this con-
clusion. First of all, by virtue of stabilization due to the ad-
ditional Ar–Ar interaction, the binding energy of Ar2–SH
͑with respect to Ar–SHϩAr͒ is Ϸ twofold larger than that of
Ar–SH. This would allow it preferentially to absorb more
recoil energy than Ar–H2S without undergoing secondary
fragmentation, in agreement with experiment. Second, based
on the gentle recoil model, an H-atom pushing off from the
more massive Ar2–HS photofragment carries away an even
greater fraction of excess energy than is dynamically the case
for Ar–H2S, which leaves less to be accommodated in the
final open-shell cluster. In all likelihood, both of these fac-
tors play a role in the photolysis event, and it remains an
interesting task for classical trajectory calculations to help
isolate and interpret the key dynamical pathways. Such cal-
culations are currently being performed in our group and will
be reported elsewhere.48
for a fixed total backing pressure and H2S concentration.
When implemented into Eq. ͑5͒, this yields
Ar–SH͔
Pa
Pc
Pb
͓
ϰ
ϩ
,
͑7͒
ͫ
ͬ
͓Ar2–SH͔ P ␤ Ar
͓
͔
b
where ␤ is the proportionality constant in Eq. ͑6͒. Thus, fits
of Eq. ͑7͒ to the experimental data can be used to predict the
relative propensity for fragmentary versus nonfragmentary
events at different UV photolysis wavelengths.
Figure 10 shows the ratio of the integrated intensities of
Ar–SH lines to Ar2–SH lines in the LIF spectrum as a func-
tion of reciprocal Ar percentage composition for both 248
and 193 nm photolysis. As predicted, the data are linear
functions of 1/͓Ar͔; the relevant quantitative information
from weighted least squares fits is presented in Table II. The
two intercepts yield information on the relative probability
for fragmentary (ϰPc) versus nonfragmentary photolysis
(ϰPb) of Ar2–H2S. Specifically, the data indicate a 2.5-fold
greater propensity for fragmentation of the trimer radical
cluster at 193 nm photolysis than at 248 nm. This is again
consistent with the much greater exoergicity at the shorter
VII. SUMMARY AND CONCLUSION
Ultraviolet photolysis of small closed-shell Arn –H2S
van der Waals clusters has been demonstrated to be a sur-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47

5158
J. Chem. Phys., Vol. 110, No. 11, 15 March 1999
Mackenzie et al.
prisingly efficient method for the unimolecular generation of
open-shell Arn –SH complexes. The probability for generat-
ing stable open-shell clusters from the photolysis event ap-
pears to be at least 15% for 248 nm, with much higher values
predicted from a range of reasonable fluorescence lifetime
estimates. At 193 nm, there is much more excess energy to
be accommodated in the photolysis event, which leads to a
roughly tenfold lower probability for open-shell cluster sur-
vival. The distribution of closed-shell precursors present in
the molecular beam has been characterized by composition-
and pressure-dependent studies which in turn suggest that
nonfragmentary photolysis ͑i.e., for which no Ar atoms are
lost, represent an important channel in the formation of the
open-shell complexes observed. Based on a simple kinetic-
and composition-dependent model of the LIF signals, the
relative probabilities of fragmentary versus nonfragmentary
photolysis of Ar–H2S and Ar2–H2S have also been investi-
gated at 248 and 193 nm. Analysis of the LIF spectrum for
the photofragments indicates that radical complexes exhibit
extremely high end-over-end rotational and intermolecular
vibrational excitation, consistent with a gentle recoil model
of the formation dynamics. As a result, this novel cluster
‘‘synthesis’’ approach can provide alternate access to highly
excited states of open-shell clusters for further dynamical
and/or spectroscopic studies. Furthermore, this internal and
rotational excitation can subsequently be quenched with sur-
prisingly high efficiency, as evidenced by the cold LIF spec-
tra of Ar–SH and Ar2–SH obtained after р30 low energy
collisions with rare gas diluent atoms in the slit jet expan-
sion. By simple choice of time delay, therefore, this method
can be used to generate high densities of open-shell clusters
under conditions smoothly varying from the high tempera-
ture limit immediately after the photolysis event down to
conventional slit jet temperatures on the order of Ϸ10 K.
9 S. K. Bramble and P. A. Hamilton, Chem. Phys. Lett. 170, 107 ͑1990͒.
10 B.-C. Chang, L. Yu, D. Cullin, B. Rehfuss, J. Williamson, T. A. Miller,
W. M. Fawzy, X. Zheng, S. Fei, and M. Heaven, J. Chem. Phys. 95, 7086
͑1991͒.
11 B.-C. Chang, D. W. Cullin, J. M. Williamson, J. R. Dunlop, B. D. Reh-
fuss, and T. A. Miller, J. Chem. Phys. 96, 3476 ͑1992͒.
12 M.-C. Yang, A. P. Salzburg, B.-C. Chang, C. C. Carter, and T. A. Miller,
J. Chem. Phys. 98, 4301 ͑1993͒.
13 M.-C. Yang, C. C. Carter, and T. A. Miller, J. Chem. Phys. 107, 3437
͑1997͒.
14 C. C. Carter and T. A. Miller, J. Chem. Phys. 107, 3447 ͑1997͒.
15 P. P. Korambath, X. T. Wu, E. F. Hayes, C. C. Carter, and T. A. Miller, J.
Chem. Phys. 107, 3460 ͑1997͒.
16 R. W. Randall, C. C. Chuang, and M. I. Lester, Chem. Phys. Lett. 200,
113 ͑1992͒.
17 M. Yang, M. H. Alexander, C. C. Chuang, R. W. Randall, and M. I.
Lester, J. Chem. Phys. 103, 905 ͑1995͒.
18 G. W. Lemire, M. J. McQuaid, A. J. Kotlar, and R. C. Sausa, J. Chem.
Phys. 99, 91 ͑1993͒.
19 W. H. Basinger, U. Schnupf, and M. C. Heaven, Faraday Discuss. 97, 351
͑1994͒.
20 W. H. Basinger, W. G. Lawrence, and M. C. Heaven, J. Chem. Phys. 103,
7218 ͑1995͒.
21 M. C. Heaven, J. Chem. Phys. 97, 8567 ͑1993͒.
22 M. T. Berry, M. R. Brustein, M. I. Lester, C. Chakravarty, and D. C.
CLary Chem. Phys. Lett. 178, 301 ͑1991͒.
23 Dynamics of Polyatomic van der Waals Complexes, edited by N. Halber-
stadt and K. C. Janda ͑Plenum, New York, 1989͒.
24 S. R. Mackenzie, O. Votava, J. R. Fair, and D. J. Nesbitt, J. Chem. Phys.
105, 11 360 ͑1996͒.
25 D. J. Nesbitt, Annu. Rev. Phys. Chem. 45, 367 ͑1994͒ and references
therein.
26 R. R. Friedl, Wm. H. Brune, and J. G. Anderson, J. Chem. Phys. 79, 4227
͑1983͒.
27 G. W. Loge and J. J. Tiee, J. Chem. Phys. 89, 7167 ͑1988͒.
28 J. J. Tiee, M. J. Ferris, and F. B. Wampler, J. Chem. Phys. 79, 130 ͑1983͒.
29 J. Zoval, D. Imre, and V. A. Apkarian, J. Chem. Phys. 98, 1 ͑1993͒.
30 W. Ubachs, J. J. ter Meulen, and A. Dynamus, Chem. Phys. Lett. 101, 1
͑1983͒.
31 W. Ubachs and J. J. ter Meulen, J. Chem. Phys. 92, 2121 ͑1990͒.
32
¨
C. Jaques, L. Valachovic, S. Ionov, E. Bohmer, Y. Wen, J. Segall, and C.
Wittig, J. Chem. Soc., Faraday Trans. 89, 1419 ͑1993͒.
33
´
A. Garcıa-Vela, J. Chem. Phys. 108, 5755 ͑1998͒.
34 K. M. Christoffel and J. M. Bowman, J. Chem. Phys. 104, 8348 ͑1996͒.
35 D. F. Plusquellic, O. Votava, and D. J. Nesbitt, J. Chem. Phys. 101, 6356
͑1994͒.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the National Sci-
ence Foundation for the financial support of this project.
S.R.M. is further grateful to NATO and The Royal Society
for a Postdoctoral Fellowship when this work was per-
formed, and J.R.F. thanks the NSF for a Predoctoral Fellow-
ship.
36 S. R. Hair, J. I. Cline, C. R. Bieler, and K. C. Janda, J. Chem. Phys. 90,
2935 ͑1989͒.
37 A. McIlroy, R. Lascola, C. M. Lovejoy, and D. J. Nesbitt, J. Phys. Chem.
95, 2636 ͑1991͒.
38 J. T. Farrell, Jr. and D. J. Nesbitt, J. Chem. Phys. 105, 9421 ͑1996͒.
39 J. T. Farrell, Jr., S. Davis, and D. J. Nesbitt, J. Chem. Phys. 103, 2395
͑1995͒.
40 D. T. Anderson, S. Davis, and D. J. Nesbitt, J. Chem. Phys. 107, 1115
͑1997͒.
1 R. S. Mulliken, J. Chim. Phys. 61, 20 ͑1964͒.
41 M. Kappes and S. Leutwyler, in Atomic and Molecular Beam Methods,
edited by G. Scoles ͑Oxford University Press, Oxford, 1988͒.
42 B. R. Weiner, H. B. Levene, J. J. Valentini, and A. P. Baronavski, J.
Chem. Phys. 90, 1403 ͑1989͒.
2 P. D. A. Mills, C. M. Western, and B. J. Howard, J. Chem. Phys. 90, 3331
͑1986͒; W. M. Fawzy and J. T. Hougen, J. Mol. Spectrosc. 137, 154
͑1989͒; M.-L. Dubernet, D. Flower, and J. M. Hutson, J. Chem. Phys. 94,
7602 ͑1991͒.
43 M. J. Weida and D. J. Nesbitt, J. Chem. Phys. 106, 3078 ͑1997͒.
44 J. Tusa, M. Sulkes, and S. A. Rice, J. Chem. Phys. 70, 3136 ͑1979͒.
45 M. Sulkes, J. Tusa, and S. A. Rice, J. Chem. Phys. 72, 5733 ͑1980͒.
46 G. M. McClelland, K. L. Saenger, J. J. Valentini, and D. R. Herschbach, J.
Phys. Chem. 83, 947 ͑1979͒.
3 J. Goodman and L. E. Brus, J. Chem. Phys. 67, 4858 ͑1977͒.
4 W. M. Fawzy and M. C. Heaven, J. Chem. Phys. 89, 7030 ͑1988͒.
5 M. T. Berry, M. R. Brustein, J. R. Adams, and M. I. Lester, J. Chem.
Phys. 92, 5551 ͑1988͒.
6 M. T. Berry, M. R. Brustein, and M. I. Lester, Chem. Phys. Lett. 153, 17
͑1988͒.
47 P. J. Robbinson and K. A. Holbrook, Unimolecular Reactions ͑Wiley–
Interscience, New York, 1972͒.
7 R. A. Loomis, B. L. Schwartz, and M. I. Lester, J. Chem. Phys. 104, 6984
͑1996͒.
48 J. R. Fair, S. R. Mackenzie, O. Votava, and D. J. Nesbitt ͑work in
progress͒.
8 W. M. Fawzy and M. C. Heaven, J. Chem. Phys. 92, 909 ͑1990͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
132.236.27.111 On: Thu, 18 Dec 2014 03:42:47