Photodissociation dynamics of Cl2O: Interpretation of electronic transitions

Article abstract of DOI:10.1021/jp049334r

The photodissociation of Cl₂O has been investigated for adjacent regions in the first and the second absorption bands, respectively, at dissociation wavelengths of 5.3 and 6.0 eV. Chlorine and oxygen atoms were state-selectively detected by resonance-enhanced multiphoton ionization. Chlorine fragment kinetic energy and spatial distributions were determined from time-of-flight profile analysis. At 5.3 eV, kinetic energy distributions are broad and bimodal and differ significantly for the two spin-orbit states Cl*(²P_1/2) and Cl(²P_3/2). The decay is characterized by a positive anisotropy parameter of 0.7 ± 0.2. Excitation proceeds via the transitions 10a₁ ← 7b₂ and 10a₁ ← 9a₁. The data are in agreement with generating ClO(X) + Cl*(²P_1/2) as primary fragments. At 6.0 eV, kinetic energy distributions are narrow and structureless and are similar for the two spin-orbit states Cl*(²P_1/2) and Cl(²P_3/2). The decay is characterized by a small positive anisotropy parameter of 0.2 ± 0.2. The main dissociation channel is the three-body decay into 2C1 + O. The decay mechanism is of an asynchronous concerted type. The data suggest the primary production of ClO(A) + Cl( ²P_J). Excitation proceeds mainly via the transition 8b₂ ← 7b₂. The results show the need for refined theoretical calculations.

Full text of DOI:10.1021/jp049334r

7954
J. Phys. Chem. A 2004, 108, 7954-7964
Photodissociation Dynamics of Cl2O: Interpretation of Electronic Transitions†
Melanie Roth,‡ Christof Maul,* and Karl-Heinz Gericke
Institut fu¨r Physikalische und Theoretische Chemie, Technische UniVersita¨t Braunschweig,
Hans-Sommer-Strasse 10, 38106 Braunschweig, Germany
ReceiVed: February 13, 2004; In Final Form: March 19, 2004
The photodissociation of Cl2O has been investigated for adjacent regions in the first and the second absorption
bands, respectively, at dissociation wavelengths of 5.3 and 6.0 eV. Chlorine and oxygen atoms were state-
selectively detected by resonance-enhanced multiphoton ionization. Chlorine fragment kinetic energy and
spatial distributions were determined from time-of-flight profile analysis. At 5.3 eV, kinetic energy distribu-
tions are broad and bimodal and differ significantly for the two spin-orbit states Cl*(2P1/2) and Cl(2P3/2). The
decay is characterized by a positive anisotropy parameter of 0.7 ( 0.2. Excitation proceeds via the transi-
tions 10a1 r 7b2 and 10a1 r 9a1. The data are in agreement with generating ClO(X) + Cl*(2P1/2) as pri-
mary fragments. At 6.0 eV, kinetic energy distributions are narrow and structureless and are similar for the
two spin-orbit states Cl*(2P1/2) and Cl(2P3/2). The decay is characterized by a small positive anisotropy
parameter of 0.2 ( 0.2. The main dissociation channel is the three-body decay into 2Cl + O. The decay
mechanism is of an asynchronous concerted type. The data suggest the primary production of ClO(A) +
Cl(2PJ). Excitation proceeds mainly via the transition 8b2 r 7b2. The results show the need for refined theoretical
calculations.
+
Cl2O f Cl2(1Σg ) + O(3PJ)
(2a)
(2b)
(2c)
(3)
I. Introduction
∆H ) 13 710 cm-1 ) 1.70 eV
λthr ) 729 nm
f Cl2(1Σg ) + O(1D2)
∆H ) 29 580 cm-1 ) 3.67 eV
λthr ) 338 nm
Photochemistry and photophysics of Cl2O have recently re-
ceived considerable attention for several reasons.1-10 The
photochemistry of chlorine oxides is of general interest due to
their role in atmospheric chemistry, although Cl2O does not
directly take part in the ozone depletion process in the strato-
sphere. Discrepancies in the assignment of its excited electronic
states6,10-12 exist, as well as disagreement between experimental
and theoretical data with respect to its dissociation dynam-
ics,6,13,14 both of which have not completely been resolved until
now. A most interesting feature of Cl2O dissociation is the clean
production of very highly rotationally excited ClO radicals.9
Last, Cl2O is an ideal candidate to study three-body decay
processes,15,16 due to its absorption spectrum extending into the
easily accessible ultraviolet region and its relatively low three-
body decay threshold of 4.18 eV. In particular, that all fragments
in a three-body decay are atomic greatly simplifies the analysis
of the experimental data.
+
+
f Cl2(3Π0 ) + O(3PJ)
∆H ) 31 540 cm-1 ) 3.91 eV
λthr ) 317 nm
Cl2O f 2 Cl(2PJ) + O(3PJ)
∆H ) 33 710 cm-1 ) 4.18 eV
λthr ) 296 nm
Dissociation enthalpies, ∆H, are endoergicities for formation
of the specified products at T ) 0 K starting from Cl2O
molecules at V ) 0 forming the J (respectively the Ω)
components of the products with the lowest energy. The first
pathway is the two-body decay, where a Cl atom and a ClO
radical are generated. Excited ClO(A2ΠΩ), if produced, is
predissociative and characterized by lifetimes on the order of a
few picoseconds or below, depending on fine-structure and
vibrational states.17 The second one is another two-body decay,
which leads to a Cl2 molecule and an atomic O fragment. Since
Cl2 elimination (reaction 2) will involve passage over an
activation barrier, it is not expected to occur at the quoted λthr
values. In both cases, the dissociation enthalpies depend on the
product electronic states of the chlorine molecule and the oxygen
atom or the ClO radical and the chlorine atom, respectively.
The third pathway, the three-body decay,15 produces two
chlorine atoms and one oxygen atom.
For photon energies below 6.2 eV, several dissociation paths
are energetically accessible:
Cl2O f Cl(2PJ) + ClO(X2ΠΩ)
∆H ) 11 530 cm-1 ) 1.43 eV
λthr ) 867 nm
f Cl(2PJ) + ClO(A2ΠΩ)
∆H ) 43 180 cm-1 ) 5.35 eV
λthr ) 231 nm
(1a)
(1b)
† Part of the special issue “Richard Bersohn Memorial Issue”.
‡ Present address: Hahn-Meitner-Institut Berlin, Glienicker Strasse 100,
14109 Berlin, Germany.
Early photochemical work has been reviewed by Renard and
* Corresponding author. E-mail: c.maul@tu-braunschweig.de.
Bolker.18 The Cl2O(X1A1) ground-state electron configuration
10.1021/jp049334r CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/23/2004

Photodissociation Dynamics of Cl2O
J. Phys. Chem. A, Vol. 108, No. 39, 2004 7955
Figure 2. Experimentally observed Cl fragment anisotropy parameters
and its limiting values as derived from theoretical calculations without
considering simultaneous excitation of several states. Theoretical work
is represented by open squares and the dotted line, respectively;
experimental data are marked by circles. Filled circles are data obtained
in this work. Experiments yield positive anisotropy parameters through-
out the energy range between 3.5 and 6.5 eV. Theory predicts zero or
negative limiting values between 4.5 and 6.5 eV.
Figure 1. Logarithmic representation of the Cl2O absorption spectrum28
and the three lowest electronic transitions suggested by Tomasello et
al.5 Transitions are mainly from nonbonding Cl(3p) orbitals to the lowest
antibonding σ*(a1) orbital. Squares represent the calculated oscillator
strengths of the transitions at 2.92, 4.42, and 4.81 eV. Vertical axes
have been scaled independently of each other.
4.18 eV restricted to reaction 1a. The data displayed in Figure
2 clearly demonstrate the existing discrepancies concerning the
ultraviolet dissociation of Cl2O and the necessity to gather
additional experimental information. Whereas the photodisso-
ciation and the corresponding states at low excitation energy
below 4 eV (above 300 nm) are well characterized today, the
situation is less clear for shorter wavelengths. Therefore,
excitation energies of 5.3 and 6.0 eV (235 and 207 nm), which
lie on different sides of a minimum in the absorption spectrum,
were chosen so that dissociation should start on different
potential energy surfaces of the molecule. The goal of the present
work is to analyze the nature of the relevant excited states of
Cl2O on the basis of the state-specific spatial fragment distribu-
tions and to investigate the competition between two- and three-
body decay on the basis of the fragment kinetic energy
distributions by monitoring Cl fragments from Cl2O photodis-
sociation by the state-selective resonance-enhanced multiphoton
ionization (REMPI) technique and analyzing the widths and the
shapes of the corresponding time-of-flight (TOF) profiles.
At a dissociation energy of 2.9 eV (423 nm), Davis and Lee
employed the photofragment translational spectroscopy tech-
nique and observed a negative anisotropy parameter, â ) -1,
for atomic chlorine fragments resulting from reaction 1a, thus
confirming excitation of Cl2O from the X1A1 ground state into
a state of B1 symmetry.29 Around 3.6 eV (340 nm), ground-
state ClO(X2ΠΩ) fragments also resulting from reaction 1a were
state-selectively detected by Aures et al. with extremely high
rotational angular momentum.4 The ClO anisotropy parameter
was found to be slightly positive, indicating a mild preference
for fragment ejection parallel to the electric field vector of the
dissociation laser light. The positive value of the anisotropy
parameter is evidence of accessing a state of B2 symmetry, as
predicted by the majority of ab initio studies.5,10,12 However,
the experimental results of Aures et al. disagree with the
association of the absorption around 350 nm (3.5 eV) with
excitation into the 1B1 state as suggested by Nickolaisen et al.11
Okumura and co-workers examined the dissociation at 308 (4.0
eV), 248 (5.0 eV), and 193 nm (6.4 eV) by photofragment
translational spectroscopy (PTS).13,30 With the exception of the
three-body decay channel, which is not allowed energetically
at 308 nm, all channels (1) through (3) are energetically
accessible at all wavelengths. For all excitation wavelengths,
two-body decay yielding ClO radicals was observed. Addition-
is [core] (6a1)2 (5b2)2 (7a1)2 (6b2)2 (8a1)2 (2b1)2 (2a2)2 (9a1)2
(7b2)2 (3b1)2 (10a1)0 (8b2)0. The six highest occupied molecular
orbitals are essentially contributed to by Cl(3p) and O(2p)
electrons, which are in principle nonbonding, whereas the two
lowest unoccupied orbitals can be identified as Cl-O antibond-
ing σ* orbitals. The energetic order of the occupied orbitals
has unequivocally been determined by photoelectron spec-
troscopy.1,2,19-21 Although photofragmentation is possible for
wavelengths below 867 nm, Cl2O absorption only starts with a
very weak band centered at 510 nm, which is probably
associated with a singlet-triplet excitation. Four more absorp-
tion features can be distinguished at 410, 285, 256, and 171
nm.22-28 The absorption spectrum of Cl2O below 500 nm is
shown in Figure 1, together with the three lowest singlet-singlet
transitions suggested by Tomasello et al.5 Accordingly, the first
strong absorption peak at 410 nm can be identified as 10a1 r
3b1 excitation into the 11B1 state, whereas the broad feature
with apparent maxima at 285 and 256 nm involves contributions
from two transitions, the 10a1 r 7b2 excitation into the 11B2
state being the dominant one, accompanied by the much weaker
10a1 r 9a1 transition into the 21A1 state. The 10a1 r 2a2
transition, which is calculated to have a similar excitation
energy, is dipole-forbidden and not assumed to contribute to
the absorption spectrum. Most probably, the next main feature
at higher energies in the absorption spectrum is predominantly
caused by the three dipole-allowed transitions from the same
7b2, 9a1, and 2a2 orbitals into the second 8b2 Cl-O antibonding
orbital.5,11 The corresponding 8b2 r 3b1 transition is unlikely
to play a significant role because it is dipole-forbidden.
However, it has also been suggested that this second group of
transitions should significantly contribute to the first absorption
feature below 5.5 eV.10 To resolve this issue by investigating
the dissociation dynamics in the energy range between 5 and 6
eV has been one of the main objectives of the present work.
Experimentally, the photochemistry of Cl2O has previously
been investigated by several groups.4,13,14,29,30 The dominating
dissociation channel depends on the excitation energy and the
excited state associated with the respective transition. Results
concerning the fragment spatial distribution, based on the â
anisotropy parameter, are summarized in Figure 2. Data gener-
ally refer to reactions 1 or 3, with excitation energies below

7956 J. Phys. Chem. A, Vol. 108, No. 39, 2004
Roth et al.
ally, at 248 and 193 nm three-body decay was observed, whereas
two-body decay yielding molecular Cl2 was only found to occur
at 193 nm. The spatial fragment distribution was monitored for
the ClO fragment for the longer wavelengths and for the O atom
at 193 nm. In all cases, the anisotropy parameter, â, was found
to be positive, indicating a preference for fragment ejection
parallel to the electric field vector of the dissociation laser light.
The positive value of the anisotropy parameter for the ClO
products formed at 308 and 248 nm is again evidence of
accessing a state of B2 symmetry from the A1 ground state, in
agreement with the large oscillator strength found for the 11B2
r X1A1 transitions in all available quantum mechanical
calculations.5,10-12 Last, Tanaka et al.14 showed state-specific
energy distributions for the Cl(2P3/2) ground state and the
Cl*(2P1/2) excited spin-orbital state, markedly different from
each other, by monitoring chlorine fragments in a photofragment
ion-imaging experiment at an excitation energy of 5.3 eV (235
nm). From the Cl fragment kinetic energy distributions, they
propose a competition between two- and three-body decay upon
excitation of a single excited state. From the observed aniso-
tropic behavior characterized by a large positive â parameter,
this state was again identified to be of B2 symmetry. Whereas
the 5.3 eV experimental results are in reasonable agreement with
the theoretical work of Nickolaisen et al.11 and of Tomasello et
al.,5 unfortunately these findings contradict dynamical calcula-
tions by Toniolo et al.10 in which quasi-classical trajectory
calculations on the initially excited potential energy surface
yielded a negative anisotropy parameter.
fragments were analyzed by the resonance-enhanced multipho-
ton ionization (REMPI) technique. Cl atoms were detected by
(2+1) REMPI in the 2P3/2 ground state and in the 2P1/2 excited
state. For the 2P3/2 ground state, the (4p2 2 3/2 r 2P3/2) transition
D
2
at 235.336 nm was employed; for the P1/2 excited state, the
2
(4p2 2S1/2 r P1/2) transition at 235.205 nm was used.35,36
O
3
atoms were detected by (2+1) REMPI in the P2 state via the
(3P2 r 3P2) transition at 225.57 nm.37 The probe laser light was
delivered by an excimer laser pumped dye laser (Lambda Physik
LPX 100, LPD 3000, Coumarin 47) and was frequency doubled
by a â barium borate (BBO) crystal. The probe pulse energy
was kept in the range between 100 and 200 µJ and was focused
into the ionization region of the TOF spectrometer by a lens of
80 mm focal length. For the 235 nm dissociation, the probe
laser was also used as the dissociation laser (one-color experi-
ment). For the 207 nm dissociation, light from a second,
Nd:YAG laser pumped dye laser was used (Coherent Infinity,
Lambda Physik Scanmate, Furan 2), which counter-propagated
the probe laser beam (two-color experiment). The dissociation
laser light was frequency doubled by a second BBO crystal.
Typically, 200 µJ/pulse was focused into the TOF spectrometer
by a lens of 90 mm focal length. A delay generator (Stanford
Research DG 535) adjusted a delay of typically 10 ns between
dissociation and probe pulses to ensure collision-free detection
of the fragments. A λ/2 wave plate allowed for the rotation of
the polarization of the dissociation laser in order to realize
parallel and perpendicular detection geometries. Here, the angle,
γ, of the electric field vector, EB, of the dissociating laser with
respect to the TOF spectrometer axis, z, defines the terms
parallel (γ ) 0°) and perpendicular (γ ) 90°).
II. Experimental Section
Cl2O gas was generated by the method of Cady,31 where a
gas mixture of N2(900 mbar) and Cl2(100 mbar) flows through
a column filled with Hg(II)O and the reaction n HgO + 2Cl2
f Cl2O + HgCl2‚(n - 1)HgO takes place. The reaction product
is purified by vacuum distillation at a temperature of ca. 200
K, maintained by a mixture of methanol and liquid nitrogen.
This procedure leaves Cl2, HOCl, and OClO as known impuri-
ties. After purification, the absorption spectrum of the sample
was measured with a Lambda 9 Perkin-Elmer spectrometer.
Fitting a superposition of the known absorption spectra of Cl2O
and the suspected impurities to the measured spectrum yielded
that the sample consists of 51% Cl2O, 33% Cl2, 15% HOCl,
and 1% OClO. In the wavelength region of interest (200-240
nm), only Cl2O and HOCl have a nonvanishing absorption cross
section, with the Cl2O cross section being larger by a factor of
6. Therefore, the contribution to the signal caused by dissociation
of HOCl cannot exceed 2.5% and can, therefore, be neglected
in the data analysis procedure. It is evident, however, that for
experiments carried out at longer dissociation wavelengths,
additional effort has to be spent to remove interference from
impurities.
The experimental setup is of the typical pump and probe type
and has been described in detail elsewhere.32,33 Cl2O gas, as
produced in the procedure described above, was mixed with
argon in a ratio of approximately 1:10 to a total pressure of 105
Pa and was expanded into the time-of-flight (TOF) spectrometer
region of a vacuum chamber in a supersonic jet produced by a
pulsed nozzle (General Valve Series 9). The base pressure in
the vacuum chamber without the jet was 0.1 Pa. At typical
repetition rates of 10 Hz, the base pressure increased by 1 or 2
orders of magnitude, depending on the nozzle operating condi-
tions. Typically, under these conditions, molecules are cooled
to temperatures below 15 K.34
The energy-sensitive home-built single-field TOF spectrom-
eter has a total length, s, of 57 cm and consists of acceleration
and drift regions, separated from each other by a grid, with a
length ratio of 1:2. This arrangement ensures space-focusing
conditions. The spectrometer axis, z, is oriented perpendicular
to the plane defined by the molecular and laser beams. The
acceleration field could be varied between 0 and 20 kV/m,
resulting in minimum times of flight, t0, of 5 µs for chlorine
ions. Ions were monitored by a 4 cm active-diameter micro-
channel plate, and the transient, energy-broadened mass peak
was recorded as the corresponding TOF profile Pt(t) by a 350
MHz digital oscilloscope (LeCroy 9450), interfaced to a personal
computer. The width and the shape of the transient TOF profiles
Pt(t) are measures of the initial kinetic energy and spatial
distributions of the fragments obtained in the dissociation
process. With respect to profile width, for our spectrometer
geometry, the deviation, ∆t, of the observed time of flight, t,
from the center, t0, of a TOF profile is proportional to the initial
velocity component, Vz, of the observed fragment along the
spectrometer axis, z:
2
∆t ) 3t0 Vz/(8s)
(4)
Thus, the distribution, Pz(Vz), of the velocity component, Vz,
onto the fragments is directly reflected in the experimental TOF
profile: Pz(Vz)
Pt(∆t). Since Pz(Vz) depends on the chosen
detection geometry, the initial spatial fragment distribution can
be investigated by rotating the dissociation laser polarization.
The energy resolution of the spectrometer is typically 5%,
depending on the acceleration voltage. All TOF profiles were
monitored for different acceleration fields in order to exclude
experimentally induced artifacts.
To increase the kinetic energy resolution of the experiment,
TOF profiles obtained at moderate acceleration field strengths
of 5 kV/m were analyzed. While increasing the energy resolu-
The supersonic molecular beam in the TOF spectrometer was
crossed by the dissociation and the probe lasers, and the

Photodissociation Dynamics of Cl2O
J. Phys. Chem. A, Vol. 108, No. 39, 2004 7957
Pθ(V,θ,æ) distribution is expressed by its dependence on the
angle θ:
PV(V)
4πV2
Pθ(V,θ,æ) )
[1 + â(V)P2(cos θ)]
(5a)
[Equation 5a ensures that both Pθ and PV are normalized:
∫∫∫ V2 sin θ dV dθ dφ Pθ(V,θ) ) ∫ dV PV(V) ) 1.] Here, θ is
the angle of the initial velocity vector of the fragment with the
electric field vector, EB, of the dissociation laser, Pv(V) is the
speed distribution of the fragments, and P2 is the second
Legendre polynomial P2(x) ) 1/2(3x2-1). If the coordinate
system is rotated by the angle, γ, so that it is fixed to the
spectrometer axis, z, rather than to the electric field vector, then
the spatial distribution can be written in terms of the velocity
component Vz:
PV(V)
2V
Vz
V
Pγ(V,Vz) )
1 + â(V)P
P (cos γ)
(5b)
2( )
[
2
]
The distribution, Pz(Vz), of the z component of the velocity,
which is directly proportional to the TOF profile, Pt(∆t), is
obtained from Pγ by integrating eq 5b over the speed V:33
Figure 3. Cl atom TOF profiles for Cl2O dissociation at 5.3 eV (235
nm). All profiles have been normalized with respect to their integrated
areas. Profiles are shown for both Cl(2P3/2) and Cl*(2P1/2) spin-orbit
states with perpendicular (γ ) 90°) and parallel (γ ) 0°) detection
geometry at an acceleration voltage Uacc ) 3 kV. Magic angle profiles
(γ ) γm) were computed from the weighted sum of the raw profiles.
Profiles reflect a positive â-anisotropy parameter and a larger average
kinetic energy release onto Cl*(2P1/2) than onto Cl(2P3/2).
PV(V)
2V
Vz
V
Pt(∆t) P (V ) ) ∞dV
1 + â(V)P
P (cos γ)
2
∫
{
2( )
}
z
z
[
]
Vz
(5c)
From eqs 5b and 5c, it is obvious that both the spatial
fragment distribution, Pγ, and the experimental TOF profile, Pt,
become independent of the anisotropy parameter, â, if P2(cos
γ) ) 0, i.e., if γ is the magic angle of 54.7°. The same result
tion, the correspondingly larger TOF causes isotopic overlap
and ion fly-out. Both effects require a modification of the
standard data analysis procedure. Details of the employed least-
squares fit method employed in the correct analysis of the high-
resolution data can be found in refs 33 and 38.
1
is obtained for the sum Pzm(Vz) ) /3[Pz(Vz,γ ) 0°) + 2Pz(Vz,γ
) 90°)]. The lowest trace in each panel in Figure 3 shows the
magic angle TOF profile for Cl and Cl* fragments, accordingly
composed from the experimentally observed profiles. Any
influence of the spatial fragment distribution on the profile shape
has been eliminated; thus the shape is only determined by the
speed distribution of the fragments according to:
III. Results
Time-of-flight (TOF) profiles, Pt(t), were monitored for both
Cl(2P3/2) ground state and Cl*(2P1/2) excited spin-orbit state
fragments for parallel (γ ) 0°) and for perpendicular (γ ) 90°)
excitation geometries. Complete sets of four TOF profiles were
obtained for each of the two dissociation wavelengths of 235
and 207 nm.
PV(V)
dPzm(Vz)
dVz
dPtm(∆t)
d∆t
) -2
-2
(6)
V
Figure 3 shows a set of TOF profiles for the dissociation
wavelength of 235 nm at an acceleration voltage of 3 kV. The
upper two traces in each panel represent the experimental raw
data for parallel and perpendicular geometries for both 35Cl and
37Cl isotopes. For both spin-orbit states, the wings of the
profiles of each isotope are more pronounced for the parallel
excitation geometry than for the perpendicular excitation
geometry. For any given fragment speed, V, the wing of a TOF
profile, which is characterized by large ∆t values, is contributed
to by those fragments with large Vz components initially flying
along the spectrometer axis, z, whereas the center of a TOF
profile, which is characterized by small ∆t values, is contributed
to by those fragments with small Vz components initially moving
perpendicular to the spectrometer axis, z. Hence, from Figure
3, it is evident that, for dissociation at 235 nm, fragments in
both spin-orbit states are preferentially ejected along the electric
field vector EB of the dissociation laser: Profile wings are
enhanced for parallel detection geometries; centers are enhanced
for perpendicular detection geometries.
Basically, the speed distribution is given by the derivative
of the magic angle TOF profile. Equation 6 only holds for an
ideal experiment with unbiased detection of all fragments and
if the apparatus response function is negligibly small compared
to the width of the TOF profiles. More often than not, the
simultaneous fulfillment of these conditions is not possible, and
the data analysis becomes significantly more complex than the
simplicity of eq 6 might suggest.33 Nevertheless, for the
conditions under which the profiles presented in Figure 3 were
obtained, one can draw some qualitative conclusions based on
the relationship established by eq 6. It is evident that the speed
distributions for Cl ground state and for Cl* excited spin-orbit
state fragments differ significantly from each other. While the
ground-state Cl profiles are rather smooth and narrow, the
excited-state Cl* profiles are broader and exhibit some modula-
tion in the wings. Therefore, the corresponding Cl* speed
distribution must be bimodal and contain contributions from
fragments which are significantly faster than their Cl ground-
state counterparts.
This situation is commonly described by a positive value
of the anisotropy parameter, â, where the spatial fragment
Figure 4 shows a set of TOF profiles for the dissociation
wavelength of 207 nm at an acceleration voltage of 3 kV,

7958 J. Phys. Chem. A, Vol. 108, No. 39, 2004
Roth et al.
Figure 5. Experimental oxygen atom TOF profiles for Cl2O dissocia-
tion wavelengths of 207 nm (filled circles) and 225.57 nm (open circles)
for the O(3P2) spin-orbit state at an acceleration voltage of 3 kV. The
207 nm data also contain some contribution from dissociation at the
225.57 nm detection wavelength. Profiles at both wavelengths are
narrow and structureless and do not depend on the detection geometry,
within experimental error. Signal intensities of the other spin-orbit
states were too low to be analyzed.
Figure 4. Experimental chlorine atom TOF profiles for Cl2O dissocia-
tion at 6.0 eV (207 nm) for both spin-orbit states and detection
geometries. Dashed lines are raw profiles and contain contributions
from the 207 nm dissociation, as well as from the dissociation by the
analysis laser at 235 nm. Dotted profiles are corresponding one-color
profiles obtained for 235 nm dissociation alone. Solid profiles are
obtained by subtracting the one-color signal from the two-color signal.
equivalent to the ones presented in Figure 3 for the dissociation
wavelength of 235 nm. In Figure 4, the experimental profile
contains contributions from both the 207 nm dissociation laser
and the 235 nm probe laser. Therefore, the experiment was
performed with (dashed line) and without the 207 nm dissocia-
tion laser present (dotted line), and in order to exclusively extract
the profiles for 207 nm dissociation, the signals were subtracted
from each other. Raw data and the corresponding difference
profiles are both shown in Figure 4. Surprisingly, the difference
profiles reflecting the dissociation dynamics at 207 nm (solid
line) are narrower than the profiles for 235 nm dissociation.
This means that the fragments generated by larger excitation
energy are released with smaller translational energy than the
fragments from 235 nm dissociation, although there is more
energy deposited in the parent molecule. This is evidence of a
new dissociation channel which is accessed at shorter dissocia-
tion wavelengths.
Only a weak dependence of profile shapes and widths on
the Cl fragment electronic state or the detection geometry is
observed. The 207 nm dissociation is, therefore, less specific
with respect to product anisotropy and quantum state.
In addition to the Cl and Cl* profiles, oxygen atoms were
also monitored. Figure 5 shows TOF profiles for dissociation
wavelengths of 225 and 207 nm. Profiles are narrow and
structureless. No influence of the detection geometry on the
profile shapes could be detected. Signal levels were sufficiently
large to be analyzed only for ground-state O(3P2) atoms. For
207 nm dissociation, a very weak, broad shoulder was observed,
which was not present for 225 nm dissociation. Otherwise, the
225 and 207 nm profiles are almost identical.
Figure 6. Chlorine atom kinetic energy distributions following Cl2O
dissociation at 5.3 eV (235 nm). Distributions are shown for both
Cl(2P3/2) and Cl*(2P1/2) spin-orbit states. Threshold energies corre-
sponding to the dissociation limits (1a) and (3) are marked by dashed
lines. For both electronic states, a bimodal behavior is observed such
that the distributions can be composed of a slow and a fast component
corresponding to channels (1a) and (3), respectively.
component which correspond to channels (1a) and (3), respec-
tively. For Cl(2P3/2) ground-state atoms, the slow component
dominates the overall distribution, whereas the opposite is the
case for excited Cl*(2P1/2) excited spin-orbit state fragments.
IV. Discussion
A. 5.3 eV Energetics. In Figure 6, state-specific kinetic
energy distributions (KEDs) of Cl(2P3/2) and Cl*(2P1/2) fragments
in both spin-orbit states are shown as obtained in the analysis
procedure. Threshold energies corresponding to the dissociation
limits (1a) and (3) are marked by dashed lines. Channel (1b) is
not accessible for a dissociation wavelength of 235 nm. For
both chlorine electronic states, a bimodal behavior is observed
such that the KEDs can be composed of a slow and a fast
This bimodal behavior has previously been observed by
Tanaka et al. employing photofragment imaging.14 For Cl(2P3/2
)
ground-state atoms, the agreement with the previous data is
excellent, whereas for Cl*(2P1/2) excited spin-orbit state atoms,
the agreement is only qualitative. Tanaka et al. observed
significantly more fast Cl* than we did, and their average kinetic
energy of the fast component is significantly larger than ours.
However, their maximum total kinetic energy in the imaging

Photodissociation Dynamics of Cl2O
J. Phys. Chem. A, Vol. 108, No. 39, 2004 7959
reliable than ours in view of the focusing conditions in our
experiment. The focus diameter df in our experiment is estimated
from the divergence D ≈ 10-3 of the laser beam and the focal
length f ) 90 mm of the focusing lens to be dfD ≈ 100 µm.
The pulse energy of 200 µJ at 235 nm corresponds to ca. 2.5 ×
1014 photons per laser pulse, resulting in a photon flux, Φ, of
3 × 1018 photons/cm2. Taking into account the absorption cross
section at 235 nm of σ235 ≈ 6 × 10-19 cm2, saturation of the
dissociation step cannot be ruled out, the effect of which would
be to reduce the value of the experimentally determined
anisotropy parameter. Unfortunately, the focusing conditions of
Tanaka et al. were not published.
Regardless of the quantitative disagreement between our
values of the anisotropy parameters and the ones observed by
Tanaka et al., that â is undoubtedly positive is sufficient to
determine the nature of the Cl2O excited state mainly involved
in the dissociation at 235 nm. The value of the â parameter for
an immediate decay on a single excited potential energy surface
depends only on the geometry of the molecule and the direction
of the transition dipole moment, which itself is determined by
the symmetries of the initial and the final states in the excitation
step. The â parameter is related to the angle, R, of the transition
dipole moment, µ, with the fragment velocity vector v by39
Figure 7. 5.3 eV (235 nm) dissociation: Comparison of experimental
Doppler profiles with Doppler profiles calculated from the kinetic
energy distributions of Figure 6. The excellent agreement confirms the
correctness of the employed analysis procedure.
â ) 2P2(cos R)
(7)
Therefore, if only one excited state is involved in the
dissociation, the â parameter of 1.2 corresponds to R ) 31°
and the â parameter of 0.7 corresponds to R ) 41°. The
sensitivity of the â parameter to the decay angle is large for the
relevant decay angles, and, thus, a large deviation in â translates
to a small deviation in R only. Therefore, experiment suggests
experiment exceeds the available energy for channel (1a).14 To
evaluate our data, we re-converted our KEDs, which have been
derived from TOF measurements, into Doppler profiles as shown
in Figure 7. This procedure yields Doppler profiles which agree
very well with our experimentally observed Doppler profiles
for both spin-orbit states if the effective laser line width for
two-photon absorption (0.3 cm-1 fwhm) is taken into account
by appropriate convolution. This agreement is not trivial since
the KED derived from the TOF data contains information not
only about the excitation into the respective resonant intermedi-
ate states of the Cl fragment but also about the ionization step,
hence, about the complete history of the fragment from frag-
mentation in the laser focus until impinging on the particle
detector. The Doppler profiles, however, contain information
about the neutral fragments only until they pass through the
resonant intermediate electronic state. Perfect agreement be-
tween these two different sets of information necessarily means
that no artifacts have been induced on the fragments following
ionization out of the intermediate resonant state. Thus, any
influence from Coulomb interaction between ionized fragments
resulting from a too large charge density inside the laser focus
can be ruled out in our experiment.
that in the 235 nm dissociation of Cl2O the decay angle, R235
,
lies within the range of 30-40°. For a simultaneous excitation
of more than one excited state, one needs to consider that the
averaged â parameter reflects the correspondingly averaged
decay angles for different symmetries.
For a nonimmediate decay involving passage through a
conical intersection, the anisotropy parameter is a function of
the rotational levels of the ClO counter fragment and can be
calculated theoretically.40 It must be negative for low J values
and become positive for a large internal excitation of ClO. We
could not experimentally observe any such J dependence due
to the small rotational constant of the ClO counter fragment.
Moreover, the fastest Cl fragments, which correspond to a small
internal excitation of the ClO counter fragment, should exhibit
a negative anisotropy parameter, which is in qualitative dis-
agreement with our own observations as well as with the results
obtained by Tanaka et al.14
C. 6.0 eV Energetics. The KEDs for 6.0 eV dissociation,
after correction for 5.3 eV contribution, are shown in Figure 8.
Again, the borderlines between the two-body radical decay (1a)
and the three-body decay (3) are marked by dashed lines. The
narrow TOF profiles translate into almost exclusive fragmenta-
tion via the three-body channel (3), accompanied by a small
production of Cl in conjunction with highly excited ClO*
radicals. Consequently, the bimodality present for both Cl and
Cl* spin-orbit states is not ascribed to a competition between
decay channels (1a) and (3), but rather to the three-body
fragmentation mechanism (3) itself.
B. 5.3 eV Anisotropy. After the KEDs and the speed
distributions, PV(V), have been determined, the anisotropy
parameters, â, can be extracted from a single TOF profile from
Figure 3 for any one detection geometry. The mean anisotropy
parameters, â ) ∫ dV {PV(V)â(V)}, exhibit the same value for
both spin-orbit states and were found to be âCl ) âCl*
)
0.7 ( 0.2. Unfortunately, the resolution of our measurements
did not permit us to resolve the speed dependence, â(V), of the
anisotropy parameters. If a speed dependence exists, the
variations of â with V should be smaller than (0.2, which is
less than the variation from 0.7 to 1.5 observed by Okumura
and co-workers in the 248 nm dissociation.30 Again, there is
qualitative agreement of our data with the values of Tanaka et
al., who observed mean anisotropy parameters of 1.2 ( 0.2 for
both spin-orbit states.14 In fact, their value might be more
The oxygen atom TOF profiles shown in Figure 5 have not
been converted to KEDs due to the small signal intensity, which
did not allow an increase in the energy resolution by lowering
the acceleration field strength. However, evaluating the maxi-
mum kinetic energy is possible from the main peak full width

7960 J. Phys. Chem. A, Vol. 108, No. 39, 2004
Roth et al.
discussed mechanisms and the geometric parameters associated
with the respective mechanisms.
For a synchronously concerted three-body decay, in addition
to the energies mentioned above, one more quantity is needed
for a complete characterization of the fragmentation process.
In this case, the natural choice for this quantity is the bond (or
decay) angle of the dissociating molecule. Any one decay angle
correlates to a precise fragment kinetic energy. Therefore, from
an experimentally observed kinetic energy distribution, one can
extract a corresponding decay angle distribution if the decay
mechanism is known to be synchronous. Cl fragment kinetic
energies Ekin(Cl) acquired in a synchronous three-body decay
of Cl2O are restricted to a certain energy range, which is
determined by the fragment masses, mCl and mO, and by the
total kinetic energy, ꢀ, of all fragments. The total kinetic energy
is the available energy Eav diminished by internal fragmentation
excitation.15 Here, Eav ) hν - ∆H is defined as the difference
of the photon energy hν and the dissociation enthalpy given in
eq 3. Generally, the Cl fragment kinetic energy Ekin(Cl) is
1
1
restricted to the range /2ꢀ/(1 + 2mCl/mO) e Ekin(Cl) e /2ꢀ,
which for dissociation into Cl(2P3/2) + Cl*(2P1/2) + O(3P2) at
6.0 eV translates to 1300 cm-1 e Ekin(Cl) e 6850 cm-1 Clearly,
the quality of fitting a Gaussian distribution of decay angles to
the observed Cl KEDs is poor, as shown in the upper panel of
Figure 9. Whereas there is no a priori reason to restrict the decay
angle distribution to a Gaussian distribution, the significant
contribution of slow Cl fragments (Ekin(Cl) < 1300 cm-1) cannot
be explained by a synchronous three-body decay for any decay
angle distribution.
Figure 8. Chlorine atom kinetic energy distributions following Cl2O
dissociation at 6.0 eV (207 nm). Distributions are shown for both
Cl(2P3/2) and Cl*(2P1/2) spin-orbit states. Threshold energies corre-
sponding to the dissociation limits (1a) and (3) are marked by dashed
lines.
at base of 70 ( 10 ns. This width corresponds to a maximum
oxygen atom speed of 4000 ( 600 m/s and to a maximum
oxygen atom kinetic energy of 10000 ( 3000 cm-1, respec-
tively. Based on the conservation of linear momentum, a hypo-
thetical Cl2 partner fragment would have to carry a maximum
kinetic energy of 2300 ( 700 cm-1, leaving a minimum internal
energy of 22300 ( 3700 cm-1 in the Cl2 fragment. Since this
value supersedes the Cl2 dissociation energy of 20000 cm-1, a
significant contribution of channel (2a) with production of
ground-state Cl2(1Σg+) can be ruled out. However, the very small
shoulders extending to 3.60 and 3.75 µs are proof of a minor
contribution of channel (2a).
For a sequential three-body decay, in addition to the energetic
quantities, two more quantities are required for completely
characterizing the fragmentation process. Two physically mean-
ingful quantities are the internal energy deposition, EClO, in the
short-lived ClO intermediate and the angle of rotation, æ, of
the ClO intermediate before it breaks apart. Any one set of these
quantities again correlates to a precise Cl fragment kinetic
energy. A characteristic feature of a sequential (planar) three-
body decay is a symmetric kinetic energy distribution for the
Cl fragment generated in the second step, which reflects
forward-backward symmetry of the decay of the freely rotating
ClO intermediate. For a sequential decay, the restriction for the
Cl kinetic energy is 0 e Ekin(Cl) e ꢀmClO/M where M is the
mass of the Cl2O parent molecule.15 For dissociation into
Cl(2P3/2) + Cl*(2P1/2) + O(3P2) at 6.0 eV, this energy range is
A two-body fragmentation is completely characterized by the
knowledge of the energies of the reactants, the products, and
the dissociating photon. Here, energy means not only enthalpy
of formation, but also internal and kinetic energy, which can
simultaneously be determined by the REMPI-TOF method
employed in our experiment. However, for a three-body decay,
the knowledge of these energetic quantities is not sufficient to
characterize the fragmentation process. Moreover, concerted and
sequential decay mechanisms will generally induce different
kinetic energies onto the fragments.15,16 Here, the terms
“concerted” and “sequential” are used to distinguish between
decay processes occurring within a rotational period of the parent
molecule and decay processes which are delayed by more than
one mean rotational period of the parent. Concerted decays may
be subdivided into synchronous decays where both bonds strictly
break in unison, such that the two terminating fragments behave
identically, and asynchronous decays where this is not the case.
Each decay mechanism can be characterized by a set of
physically meaningful parameters, such as the bond angle when
the molecule decays, the internal energy stored in the short-
lived ClO intermediate, and the angle of rotation of the ClO
intermediate before it breaks apart. The required analysis
procedure has been described elsewhere in detail.15,16 As a
consequence, the observed kinetic energies can be analyzed in
order to recognize the contribution of one or more of the
E
kin(Cl) e 8150 cm-1. The resulting fit is shown in the middle
panel of Figure 9. Although its quality is higher than for the
synchronously concerted case, the symmetry properties of the
second step KED allow either the low-kinetic energy side or
the high-kinetic energy side to be described well, but not both
for the same decay parameter distribution. Since the description
of the experimental KED by a sequential decay yields a more
reasonable result than for the synchronous decay, an asynchro-
nously concerted decay corresponding to a moderate restriction
of the free rotation of the ClO intermediate to the most probable
angle of rotation prior to the decay of the ClO intermediate
should result in a good description of the experimental KED.
Modeling such an asynchronous concerted decay with Gaussian
distributions of, æ, and EClO peaking at 105° and 0.75ꢀ,
respectively, the fit shown in the lower panel of Figure 9 is
obtained, which is the best description to the experimental data.
The widths (fwhm) of the Gaussian distributions are 30° and
0.14ꢀ, respectively.
The energy deposition, EClO, refers to the energy of the
separated atoms Cl + O in their respective quantum states. If

Photodissociation Dynamics of Cl2O
J. Phys. Chem. A, Vol. 108, No. 39, 2004 7961
Figure 10. Geometry, symmetry properties, and orientation of transi-
tion dipole moments for the Cl2O(X1A1) electronic ground state.
the primary fragments. The effect of the delayed fragmentation
of the intermediate ClO particle can be evaluated using the
relation41-43
1 + ω2τ2
âτ ) â
(8)
(
)
1 + 4ω2τ2
where ω is the frequency of rotation of the intermediate, and τ
is its lifetime. âτ is the accordingly reduced anisotropy
parameter, and â is the anisotropy parameter for immediate
decay, as defined in eq 7. From the previous paragraph, we
know that ωτ ≈ 0.3. Thus, the effect of delayed fragmentation
of the intermediate is estimated to be a reduction of the
immediate â parameter by 20%: âτ ) 0.8â. If we assume for
simplicity that the first fragmentation event takes place im-
mediately with the excitation, then the experimentally observed
1
anisotropy parameter, /2(âτ + â), will be 10% lower than the
anisotropy parameter, â, characterizing the excitation step. Since
the experimentally determined averaged anisotropy parameter
was determined to be 0.2 ( 0.2, this small effect might be
neglected in view of the much larger experimental uncertainty,
and we conclude that the anisotropy parameter, â, characterizing
the excitation step of Cl2O at 207 nm is 0.2 ( 0.2.
Focusing conditions that might be responsible for reducing
the anisotropy parameter observed for the 235 nm dissociation
are not likely to play an important role at 207 nm. First, the
pulse energy at 207 nm is lower than that at 235 nm, and second,
the Cl2O absorption cross section at 207 nm is smaller than
that at 235 nm. Even if the observed small deviation at 235 nm
were entirely due to saturation of the dissociation step, a
possible, but necessarily even smaller, effect at 207 nm can
safely be ignored.
E. Electronic Structure. The transition dipole moment, µ,
of a dipole-allowed electronic transition must have appropriate
symmetry properties, depending on the symmetries of the initial
and the final states. For transitions originating from the totally
symmetric Cl2O(X1A1) ground state, the symmetry property of
the transition dipole moment must be the same as that of the
target state, i.e., a1 for a transition to an A1 state, b1 for a
transition to a B1 state, and b2 for a transition to a B2 state. The
corresponding orientation of the transition dipole moments in
the molecular frame is shown in Figure 10. Consequences for
spatial distributions of photodissociation fragments and for â
anisotropy parameters, respectively, are discussed below.
The Cl2O bond length and bond angle were determined to
be 169.587 pm and 110.886°, respectively.44 The transition
dipole moment for a A1 r A1 transition (a1 symmetry) lies
parallel to the C2 symmetry axis. The angle, R(a1), of µ(a1) with
the recoil velocity, v, of a Cl fragment, taken to be parallel to
the Cl-O bond, is equal to half of the Cl-O-Cl bond angle,
and thus amounts to R(a1) ) 55°, close to the magic angle,
resulting in an anisotropy parameter of â*(a1) ) 0.0, according
Figure 9. Fits of different decay models to the experimental Cl
fragment kinetic energy distribution for Cl2O + hν(207 nm) f Cl(2P3/2
)
+ Cl*(2P1/2) + O(3P2). A good description of the experimentally
determined KED is only achieved assuming an asynchronous, concerted
three-body decay mechanism.
the dissociation energy, D0(ClO) ) 22180 cm-1, is added to
E
ClO, one obtains a ClO excitation with respect to its ground
state of ca. 33000 cm-1, which is only little more than the ClO-
(A2Π) excited state at 31480 cm-1 17
The remaining energy of
.
the order of 1500 cm-1 would be sufficient to populate three
vibrational or ca. 50 rotational levels. The lifetimes of the first
six vibrational levels of ClO(A2Π) have been determined to
vary between 0.2 and 2 ps, whereas a rotational period for ClO-
(A2Π) with one quantum of nuclear rotation lasts ca. 25 ps.
These numbers suggest that the asynchronous concerted three-
body decay might produce ClO(A2Π) + Cl(2P) in the first step,
followed by ClO(A2Π) predissociation on the picosecond time
scale.
D. 6.0 eV Anisotropy. The mean anisotropy parameters for
Cl and Cl* production at 207 nm were found to be 0.2 ( 0.2.
Again, no variation of the â parameter with fragment speed
could be observed. It must be taken into account, however, that
now each decay produces two individual fragments which are
likely to be differently distributed in space. Thus, the â
parameter determined in our experiment is an average of these
two physically different species produced by the three-body
decay. Of course, the physical properties of the final fragments
are not independent of each other, and the first fragmentation
event must necessarily induce identical spatial distribution with
respect to the electric field vector of the dissociation laser onto

7962 J. Phys. Chem. A, Vol. 108, No. 39, 2004
Roth et al.
TABLE 1: Compilation of Theoretical Results for the Excitation of Cl2O: Spectroscopic Terms, Limiting Anisotropy
Parameters, â*, for Immediate Decay from Ground-State Geometry, Vertical Excitation Energies, ∆E, and
Oscillator Strengths, fa
Tomasello et al.5
Toniolo et al.10
Del Bene et al.12
Nickolaisen et al.11
â*
∆E/eV
f
∆E/eV
f
∆E/eV
f
∆E/eV
f
1
2
3
4
5
6
7
8
9
10
11
12
13
-1
+1
11B1
11B2
11A2
21A1
21A2
21B1
31A1
21B2
31B1
41B1
31A2
41A1
31B2
2.92
4.42
4.57
4.81
5.69
6.97
7.14
7.55
7.85
8.25
8.62
8.73
9.20
0.00010
0.01060
2.68
4.39
4.39
4.47
4.88
6.41
5.35
6.77
6.58
7.70
6.44
6.42
7.04
0.00110
0.11000
2.82
4.23
4.23
4.66
5.01
0.00008
0.01017
3.42
4.98
4.97
5.45
5.88
7.35
0.00014
0.00770
0
0.00050
0.00059
0.00177
0.00009
0.00526
0.00001
-1
0
+1
-1
-1
0.00450
0.00040
0.20660
0.04530
0.00000
0.00063
0.00001
0.00900
0.13000
0.00240
7.98
0
+1
0.01030
0.03440
0.00170
0.07200
a The â*(∆E) dependence is illustrated in figure 2 by square symbols.
to eq 7. The asterisk indicates that the respective value is
calculated for an immediate decay from the ground-state
geometry and does not account for any dynamical effects or
for effects due to the initial conditions of the parent, which will
be discussed below. Applying the same reasoning to a B1 r
A1 transition, the transition dipole moment of b1 symmetry is
oriented perpendicular to the molecular plane. Then, the angle,
R(b1), of µ(b1) with v is 90°, and â*(b1) ) -1.0. For a B2 r
A1 transition, the transition dipole moment of b2 symmetry is
oriented parallel to the line connecting the two Cl atoms. The
angle, R(b2), of µ(b2) with v is 34.5°, and â*(b2) ) +1.0. The
determination of anisotropy parameters is therefore a useful tool
for investigating the nature of electronically excited states in
photodissociation experiments. The dipole-forbidden A2 r A1
transition will become allowed upon a reduction of molecular
symmetry as it occurs for excitation of the anti-symmetric
Cl-O-Cl stretch vibrational mode. However, even at room
temperature, only 3% of the Cl2O molecules will be excited in
the anti-symmetric stretch due to the relatively large vibrational
quantum of 85 meV;45 for the jet-cooled molecules in our
experiment, only zero-point motion will occur. Hence, we
conclude that the A2 r A1 transition does not significantly
contribute to the optical excitation.
The observed â parameter values are a result of the
simultaneous excitation of different electronically excited states
of the parent molecule. In principle, aside from simultaneous
excitation of more than one electronically excited state, there
are other effects which might also influence the â parameter,
in addition to the relationships outlined above. However, none
of these effects can explain the experimental observations. First,
dynamic effects will be considered. If the excited potential
energy surface exhibits a gradient with respect to the bond angle,
the parent molecule might change its geometry prior to
fragmentation if its lifetime is at least of the order of a
vibrational period. Note, that for an immediate decay, the
additional motion alone that will be inferred onto the nuclei by
the gradient on the upper potential energy surface is not
sufficient to noticeably change the â parameter, except for near-
threshold conditions where the repulsive force is very small.
This latter effect can safely be ignored for our experimental
conditions. Dynamic effects might decrease or increase the value
of the limiting â* anisotropy parameter depending on whether
it increases or decreases the angle, R. Some qualitative argu-
ments on this issue can be obtained following Walsh’s rules,46
which predict a bent geometry for molecules with 20 valence
electrons and a linear geometry for 21 valence electrons, which
agrees with the bent ground-state geometry of the 20 valence
electron Cl2O molecule. Accordingly, the electronically excited
molecule should prefer a linear, or at least a more linear,
configuration with an increased bond angle. The effect on the
â parameters of the corresponding transitions would be an
increase of â(b2) for the B2 r A1 transition and a decrease of
â(a1) for the A1 r A1 transition, while the perpendicular B1 r
A1 transition would remain unaffected by the nuclear rearrange-
ment. While this behavior has in fact been predicted from
trajectory calculations based on the potential energy surfaces
of Toniolo et al.,6 it contradicts our own, as well as previously
obtained, experimental evidence.13,14,30 A second effect is the
decrease of the â parameter by initial parent molecule motion
due to translational, rotational, or vibrational excitation. This
effect can be ignored in our case for two reasons. First, the
parent molecule is cooled in the supersonic expansion to
temperatures below 15 K, and second, the dissociation threshold
is much smaller than the photon energy, so that even under bulk
conditions the fragment momentum induced by repulsion in the
dissociation coordinate would be much larger than the one
induced by parent molecule motion. A third effect influencing
the value of â that needs to be considered is connected to the
nonzero lifetime of the parent molecule for nonimmediate decay.
Previous experimental and theoretical work performed for
smaller excitation energies suggests, however, that Cl2O pho-
todissociation is direct and prompt.6,9 Therefore, we assume it
to be very unlikely that, for the larger excitation energies used
in our study, lifetime effects should be of any relevance for the
experimentally observed â parameter.
Our interpretation of simultaneous excitation is supported by
the large density of states above an excitation energy of 4 eV
which are predicted for all theoretical calculations, as compiled
in Table 1. The respective limiting â* anisotropy parameters,
which would be expected for immediate decay from ground-
state geometry following a pure excitation into a single excited
state, are shown together with experimentally determined
anisotropy parameters in Figure 2. The â parameters at 235 and
207 nm, which have been determined in this work, are marked
by filled circles. Experimentally, the anisotropy parameter
increases from -1 at the lowest excitation energies around and
below 3 eV to positive values between 3.5 and 6.5 eV, with a
maximum of about +1 at 5 eV. This â(E) dependence is firmly
established by experiments; no less than six research groups
have contributed to this result without any qualitative contradic-
tion among them. At first sight, this result may easily be
rationalized by a series of the three dipole-allowed one-electron

Photodissociation Dynamics of Cl2O
J. Phys. Chem. A, Vol. 108, No. 39, 2004 7963
V. Conclusion
excitations out of the four highest occupied molecular orbitals
(2a2)2 (9a1)2 (7b2)2 (3b1)2 into the lowest unoccupied (10a1)
molecular orbital. This series would consist of a sequence of
transitions to the 1B1, 1B2, and 2A1 states (numbers 1, 2, and
4 in Table 1) with corresponding â* parameters -1, +1, and
0. However, the excitation dynamics must be more complex
than that. The energetic separation of the four highest occupied
molecular orbital is well-known from photoelectron spectros-
copy and covers an energy range of only 2.8 eV. This
contradiction becomes apparent when individual theoretical
results are compared to experimental data. Whereas the predic-
tions of Nickolaisen et al. agree well with the observed
fragmentation dynamics for excitation energies around 5 eV,
their assignment of the first small absorption feature around 3
eV as resulting from triplet r singlet excitation must be
regarded as erroneous. On the other hand, the assignment of
this feature to the 10a1 r 3b1 excitation agrees very well the
experiments of Davis et al. and of Aures et al., but goes along
with incorrect predictions of the fragmentation dynamics around
5 eV. For this energy range, predicted zero or negative â
parameter values are in marked contrast with the experimentally
observed positive values. Considering the second series of one-
electron excitations into the second lowest unoccupied molecular
orbital (8b2) results in a sequence of transitions to 3A1, 2B2,
and 3B1 states (numbers 7, 8, and 9 in Table 1) with
corresponding â* parameters 0, +1, and -1. In fact, this
explanation extends the range where positive â parameters will
be observed well into the energy range above 6 eV and matches
the experimental observations well. In this case, energies and
oscillator strengths for the 2B2 and 3B1 states as calculated by
Tomasello et al. are in much better agreement with experimental
observations than the data published by Toniolo et al. In any
case, the influence of the 2B1 state, which both authors calculate
to lie between the two one-electron sequences into the 10a1 and
8b2 orbitals, respectively, has not been observed experimentally.
Either this state, being the first involving higher excitations than
the 8b2 molecular orbital, mixes strongly with other high-lying
excited states or the rather small oscillator strength, in combina-
tion with unfavorable Franck-Condon factors, might reduce
the intensity of this perpendicular transition to such an extent
that it is too small to be experimentally observable.
The photodissociation of Cl2O has been investigated by
resonance-enhanced multiphoton ionization in combination with
time-of-flight measurements for the excitation energies 5.3 eV
(235 nm) and 6.0 eV (207 nm) in adjacent regions of the first
and the second absorption bands. At 5.3 eV, the main primary
dissociation channel is ClO + Cl(2P1/2). A significant portion
of Cl(2P1/2) undergoes conversion to ground-state Cl(2P3/2) during
fragmentation. The decay is characterized by a positive anisot-
ropy parameter of 0.7 ( 0.2. Excitation proceeds via the
transitions 10a1 r 7b2 and 10a1 r 9a1. At 6.0 eV, the main
dissociation channel is the three-body decay into 2Cl + O. The
decay mechanism is of an asynchronous concerted type, possibly
with ClO(A2Π) as intermediate fragment. The decay is char-
acterized by a small positive anisotropy parameter of 0.2 ( 0.2.
Excitation proceeds mainly via the transition 8b2 r 7b2.
Acknowledgment. The authors thank Professor Kawasaki
for fruitful discussions. The work was supported by the Deutsche
Forschungsgemeinschaft. M.R. thanks the Fonds der Chemis-
chen Industrie for fellowship support.
References and Notes
(1) Tomasello, P.; Ehara, M.; Nakatsuji, H. J. Chem. Phys. 2003, 118,
5811.
(2) Cooksey, C. C.; Reid, P. J. J. Phys. Chem. A 2003, 107, 5508.
(3) Chau, F.-T.; Dyke, J. M.; Lee, E. P. F.; Mok, D. K. W. J. Chem.
Phys. 2003, 118, 4025.
(4) Aures, R.; Gericke, K.-H.; Maul, C.; Trott-Kriegeskorte, G.;
Kawasaki, M.; Nakano, Y. J. Chem. Phys. 2002, 117, 2141.
(5) Tomasello, P.; Ehara, M.; Nakatsuji, H. J. Chem. Phys. 2002, 116,
2425.
(6) Collaveri, C.; Granucci, G.; Persico, M.; Toniolo, A. J. Chem. Phys.
2001, 115, 1251.
(7) Smith, G. D.; Tablas, F. M. G.; Molina, L. T.; Molina, M. J. J.
Phys. Chem. A 2001, 105, 8658.
(8) Roth, M.; Einfeld, T.; Gericke, K.-H.; Maul, C. Phys. Chem. Earth
2001, C26, 513.
(9) Aures, R.; Gericke, K.-H.; Kawasaki, M.; Maul, C.; Nakano, Y.;
Trott-Kriegeskorte, G.; Wang, Z. Phys. Chem. Commun. 2001, 22, 1.
(10) Toniolo, A.; Persico, M.; Pitea, D. J. Phys. Chem. A 2000, 104,
7278.
(11) Nickolaisen, S. L.; Miller, C. E.; Sander, S. P.; Hand, M. R.;
Williams, I. H.; Francisco, J. S. J. Chem. Phys. 1996, 104, 2857.
(12) Del Bene, J. E.; Watts, J. D.; Bartlett, R. J. Chem. Phys. Lett. 1995,
246, 541.
Hence, we conclude that at 235 nm Cl2O excitation mainly
occurs via the transitions 10a1 r 7b2 and 10a1 r 9a1. The
observed â parameter of +0.7 suggests that the former makes
the dominant contribution. Its large oscillator strength might
easily compensate for the somewhat larger deviation of its
calculated excitation energy from the photon energy. The
situation can largely be rationalized by the assignment depicted
in Figure 1. According to Tanaka et al., the 1B2 state correlates
with a ClO(X2Π) + Cl*(2P1/2), and ground-state Cl(2P3/2) will
be produced by nonadiabatic long-range interactions, which play
a role only in a restricted range of the initially excited potential
energy surface.14 Therefore, the difference in the KEDs for
Cl*(2P1/2) and Cl(2P3/2) atoms reflects the limited interaction
region of the corresponding PES. The excitation at 207 nm, in
the next absorption band on the short-wavelength side, is likely
to promote the 7b2 electron into the second unoccupied 8b2
orbital. Although suggested by theory, no evidence is found
for electronic excitation into higher orbitals. Similarly to the
situation at 235 nm, a strong, dominant 8b2 r 9a1 transition
might also be responsible for masking the influence of higher
excitations. In any case, theoretical results must further be
improved in order to satisfactorily describe the dissociation
dynamics of Cl2O for excitation energies between 3 and 7 eV.
(13) Nelson, C. M.; Moore, T. A.; Okumura, M.; Minton, T. K. J. Chem.
Phys. 1994, 100, 8055.
(14) Tanaka, Y.; Kawasaki, M.; Matsumi, Y.; Fujiwara, H.; Ishiwata,
T.; Rogers, L. J.; Dixon, R. N.; Ashfold, M. N. R. J. Chem. Phys. 1998,
109, 1315.
(15) Maul, C.; Gericke, K.-H. Int. ReV. Phys. Chem. 1997, 16, 1.
(16) Maul, C.; Gericke, K.-H. J. Phys. Chem. A 2000, 104, 2531.
(17) Howie, W. H.; Lane, I. C.; Newman, S. M.; Johnson, D. A.; Orr-
Ewing, A. J. Phys. Chem. Chem. Phys. 1999, 1, 3079.
(18) Renard, J. J.; Bolker, H. I. Chem. ReV. 1976, 76, 487.
(19) Wang, D.-C.; Lee, E. P. F.; Chau, F.-T., Mok, D. K. W.; Dyke, J.
M. J. Phys. Chem. A 2000, 104, 4936.
(20) Motte-Tollet, F.; Delwiche, J.; Heinesch, J.; Hubin-Franskin, M.-
J.; Gingell, J. M.; Jones, N. C.; Mason, N. J.; Marston, G. Chem. Phys.
Lett. 1998, 284, 452.
(21) Cornford, A. B.; Frost, D. C.; Herring, F. G.; McDowell, C. A. J.
Chem. Phys. 1971, 55, 2820.
(22) Goodeve, C. F.; J Wallace, I. Trans. Faraday Soc. 1930, 26, 254.
(23) Finkelnburg, W.; Schumacher, H. J.; Stieger, G. Z. Phys. Chem.,
Abt. B 15, 127.
(24) Lin, C.-L. J. Chem. Eng. Data 1976, 21, 411.
(25) Molina, L. T.; Molina, M. J. J. Phys. Chem. 1978, 82, 2410.
(26) Knauth, H.-D.; Alberti, H.; Clausen, H. J. Phys. Chem. 1979, 83,
1604.
(27) Nee, J. B. J. Quant. Spectrosc. Radiat. Transfer 1991, 46, 55.
(28) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric

7964 J. Phys. Chem. A, Vol. 108, No. 39, 2004
Roth et al.
Modeling, EValuation Number 12; JPL Publication 97-4; Jet Propulsion
Laboratory: Pasadena, 1997.
(37) Bamford, D. J.; Jusinski, L. E.; Bishel, W. K. Phys. ReV. A: At.,
Mol., Opt. Phys. 1986, 34, 185.
(29) Davis, H. F.; Lee, Y. T. J. Chem. Phys. 1996, 105, 8142.
(30) Moore, T. A.; Okumura, M.; Minton, T. K. J. Chem. Phys. 1997,
107, 3337.
(31) Cady, G. H. Inorg. Synth. 1957, 5, 156.
(32) Haas, T.; Maul, C.; Gericke, K.-H.; Comes, F. J. Chem. Phys. Lett.
1993, 202, 108.
(33) Roth, M.; Maul, C.; Gericke, K.-H. Phys. Chem. Chem. Phys. 2002,
4, 2932.
(34) Maul, C.; Haas, T.; Gericke, K.-H. J. Phys. Chem A 1997, 101,
6619.
(35) Arepalli, S.; Presser, N.; Robie, D.; Gordon, R. J. Chem. Phys.
Lett. 1985, 118, 88.
(36) Ascenzi, D.; Regan, P. M.; Orr-Ewing, A. J. Chem. Phys. Lett.
1999, 310, 477.
(38) Matsumi, Y.; Shamsuddin, S. M.; Sato, Y.; Kawasaki, M. J. Chem.
Phys. 1994, 101, 9610.
(39) Bersohn, R.; Lin, S. H. AdV. Chem. Phys. 1969, 16, 67.
(40) Mordaunt, D. H.; Ashfold, M. N. R.; Dixon, R. N. J. Chem. Phys.
1996, 104, 6460.
(41) Jonah, C. J. Chem. Phys. 1971, 55, 1915.
(42) Busch, G. E.; Wilson, K. R. J. Chem. Phys. 1972, 56, 3626.
(43) Busch, G. E.; Wilson, K. R. J. Chem. Phys. 1972, 56, 3638.
(44) Nakata, M.; Sugie, M.; Takeo, H.; Matsumura, C.; Fukuyama, T.;
Kuchitsu, K. J. Mol. Spectrosc. 1981, 86, 241.
(45) Xu, Y.; McKellar, A. R. W.; Burkholder, J. B.; Orlando, J. J. J.
Mol. Spectrosc. 1996, 175, 68.
(46) Walsh, A. D. J. Chem. Soc. 1953, 2288.