Collision energy dependence of the O(1D) + HCl -OH + Cl( 2P) reaction studied by crossed beam scattering and quasiclassical trajectory calculations on ab initio potential energy surfaces

Article abstract of DOI:10.1021/jp076581x

The dynamics of the O(¹D) + HCl → OH + Cl(²P) reaction are investigated by a crossed molecular beam ion-imaging method and quasiclassical trajectory calculations on the three ab initio potential energy surfaces, the ground 1¹A′ and two excited (1 ¹A″ and 2¹A′) states. The scattering experiment was carried out at collision energies of 4.2, 4.5, and 6.4 kcal/mol. The observed doubly differential cross sections (DCSs) for the Cl(²P) product exhibit almost no collision energy dependence over this inspected energy range. The nearly forward-backward symmetric DCS indicates that the reaction proceeds predominantly on the ground-state potential energy surface at these energies. Variation of the forward-backward asymmetry with collision energy is interpreted using an osculating complex model. Although the potential energy surfaces obtained by CASSCF-MRCI ab initio calculations exhibit relatively low potential barriers of 1.6 and 6.5 kcal/mol for 1 ¹A″ and 2¹A′, respectively, the dynamics calculations indicate that contributions of these excited states are small at the collision energies lower than 15.0 kcal/mol. Theoretical DCSs calculated for the ground-state reaction pathway agree well with the observed ones. These experimental and theoretical results suggest that the titled reaction at collision energies less than 6.5 kcal/mol is predominantly via the ground electronic state.

Full text of DOI:10.1021/jp076581x

818
J. Phys. Chem. A 2008, 112, 818-825
Collision Energy Dependence of the O(1D) + HCl f OH + Cl(2P) Reaction Studied by
Crossed Beam Scattering and Quasiclassical Trajectory Calculations on Ab Initio Potential
Energy Surfaces
Hiroshi Kohguchi and Toshinori Suzuki*
Chemical Dynamics Laboratory, RIKEN, Hirosawa 2-1, Wako 351-0198, Japan
Shinkoh Nanbu
Research Institute for Information Technology, Kyusyu UniVersity, Hakozaki, Higashiku,
Fukuoka 812-8581, Japan
Toshimasa Ishida
Fukui Institute for Fundamental Chemistry, Kyoto UniVersity, Takano-nishihirakicho,
Sakyoku, Kyoto 606-8103, Japan
Gennady V. Mil’nikov,† Ponmile Oloyede,‡ and Hiroki Nakamura
Institute for Molecular Science, Myodaiji, Okazaki 444-8585, Japan
ReceiVed: August 15, 2007; In Final Form: October 31, 2007
The dynamics of the O(1D) + HCl f OH + Cl(2P) reaction are investigated by a crossed molecular beam
ion-imaging method and quasiclassical trajectory calculations on the three ab initio potential energy surfaces,
the ground 11A′ and two excited (11A′′ and 21A′) states. The scattering experiment was carried out at collision
energies of 4.2, 4.5, and 6.4 kcal/mol. The observed doubly differential cross sections (DCSs) for the Cl(2P)
product exhibit almost no collision energy dependence over this inspected energy range. The nearly forward-
backward symmetric DCS indicates that the reaction proceeds predominantly on the ground-state potential
energy surface at these energies. Variation of the forward-backward asymmetry with collision energy is
interpreted using an osculating complex model. Although the potential energy surfaces obtained by CASSCF-
MRCI ab initio calculations exhibit relatively low potential barriers of 1.6 and 6.5 kcal/mol for 11A′′ and
21A′, respectively, the dynamics calculations indicate that contributions of these excited states are small at
the collision energies lower than 15.0 kcal/mol. Theoretical DCSs calculated for the ground-state reaction
pathway agree well with the observed ones. These experimental and theoretical results suggest that the titled
reaction at collision energies less than 6.5 kcal/mol is predominantly via the ground electronic state.
Introduction
reactions. Opening of the excited-state reaction pathway at an
elevated collision energy has been identified by the crossed
molecular beam study and the theoretical calculations for the
O(1D) + H2 (and its isotope variants) reaction.6-8 The excitation
function observed by Hsu et al. declined with Ecol up to ∼1.8
kcal/mol and then gradually increased above this energy.8 The
declining reactive cross section, σreac(Ecol), is a typical attribute
to attractive PESs with no energy barrier, while the slow rise is
attributed to a PES with an energy barrier. The energy threshold
of the latter channel is in good agreement with the barrier height
of 2.3 kcal/mol at the H-H-O collinear configuration obtained
by the ab initio calculation by Schatz et al.9 Their calculations
predicted a comparable barrier height and reactivity for the 11A′′
and 21A′ states. The energy dependence of the angle-specific
translational energy release, or doubly differential cross section
(DCS),6,7 unambiguously showed the contribution of the excited
state PES(s); the forward-backward symmetric DCS below the
energy threshold is ascribed to the pure insertion mechanism
on the 11A′ PES,10 while the increasing backward components
at higher Ecol is interpreted as a manifestation of abstraction
mechanism due to the excited 11A′′ and/or 21A′ state PESs.6,7
The interaction of atomic oxygen in the first excited singlet
state [O(1D)] with a closed-shell hydride (MH) gives rise to
five electronic states at close distance. Among them, the
minimum-energy path leads to a deep potential well at the
M-O-H geometry, which corresponds to stable intermediates
such as H2O(X A1) and CH3OH(X A′) in the O(1D) + H2
and O(1D) + CH4 reactions, respectively. Meanwhile, the
excited-state potential energy surfaces (PESs) of these inter-
mediates have energy barriers for bimolecular reactions.1-5 The
large reactivity of O(1D) results from the attractive interactions
on the ground-state PES, while the repulsive excited-state PESs
may contribute at higher collision energy (Ecol). These are
qualitatively common features of the PESs for the O(1D) + MH
1
1
* To whom correspondence should be addressed. E-mail:
toshisuzuki@riken.jp.
† Current address: Faculty of Engineering, Osaka University, Suita 565-
0871, Osaka, Japan
‡ Current address: Department of Chemistry and Environmental Science,
New Jersey Institute of Technology, University Heights, Newark, New
Jersey 07102.
10.1021/jp076581x CCC: $40.75 © 2008 American Chemical Society
Published on Web 01/12/2008

Collision Energy Dependence of O(1D) + HCl f OH + Cl(2P)
J. Phys. Chem. A, Vol. 112, No. 5, 2008 819
The O(1D) + HCl f OH + Cl system also forms a deep
insertion well on the ground-state PES (11A′) corresponding to
the hypochlorous acid (HOCl) molecule. Although a part of the
excited-state PES(s) responsible for ultraviolet (UV) photodis-
sociation of HOCl has been theoretically studied,11-14 global
PESs of the 11A′′ and 21A′ states relevant to the titled reaction
were investigated only recently.15 Previous experimental studies
have pointed only to the ground-state (11A′) dynamics. Park
and Wiesenfeld measured the vibrational and rotational state
distributions of OH products by laser-induced fluorescence (LIF)
spectroscopy, and they suggested the reaction mechanism solely
on the ground-state PES on the basis of the observed unimodal
triangle plot of energy disposal (rotational, vibrational, and
translational energy release).16 Matsumi and co-workers mea-
sured the branching ratio between the OH + Cl and ClO + H
pathways together with their isotope effects (H and D) and the
kinetic energy releases.17 They also argued that the primary
reaction pathways of both product channels are the 11A′ ground-
state PES; yet, the products are controlled by orientation of HCl
with respect to the incoming O atom; a sideway attack of O(1D)
to HCl leads to OH + Cl by an insertion/decomposition
mechanism, while the initial approach of O(1D) to the Cl end
results in ClO production due to abstraction, both on the ground-
state PES. For the ClO production, an abstraction mechanism
through the excited 11A′′ PES was proposed as another
possibility by Kruus et al.18 from their time-resolved Fourier
transform chemiluminescence spectroscopy. The doubly DCS
of the OH + Cl pathways was measured for each of the
Cl(2PJ)3/2) and Cl*(2PJ)1/2) products by our previous crossed
beam study at Ecol ) 6 kcal/mol.19 The angular distributions
exhibit pronounced preference in the forward direction. Al-
though any evidence of the contribution of the excited-state PESs
was not seen in the experimental observables, further investiga-
tion at different collision energies has been performed to
elucidate the influence of the excited states.
The detailed structure of the ground-state (11A′) PES has been
examined in theoretical studies. In addition to the potential well
at the O-insertion geometry (HOCl), another shallower potential
well associated with the HClO configuration has been confirmed
by the ab initio studies for the 11A′ PES.11,12,15,20-22 These ab
initio data for the 11A′ PES have been applied to the dynamics
calculations.23-35 Although the global PESs of the excited states
(11A′′ and 21A′) have been calculated by Nanbu et al. with the
same accuracy as that of the ground state,15 the dynamics
calculations on these PESs were limited only to J (total angular
momentum) ) 0.24,25 The ab initio calculations predicted
comparable or even lower reaction barriers of the excited-state
PESs than the Ecol applied in our previous study. Here, we
present the experimental and theoretical studies to explore the
role of the excited-state dynamics in the O(1D) + HCl f OH
+ Cl(2P) reaction. Scattering distributions observed by crossed
beam ion imaging at three collision energies are compared with
quasiclassical trajectory (QCT) calculations executed on the
11A′′ and 21A′ excited-state PESs as well as on the 11A′ ground
state.
determined the HCl beam speed from the difference in arrival
times between these two positions to be 1398 ( 118 m/s in
He, 729 ( 58 m/s in Ne, and 531 ( 40 m/s in Ar. These beam
speeds yield Ecol of 6.4 (1.4), 4.5 (0.8), and 4.2 (0.7) kcal/mol
for the He, Ne, and Ar carrier gases, respectively. The collision
energy spread (∆Ecol: full width at half-maximum) in the
parentheses is evaluated from the velocity width of each HCl
beam and the O(1D) beam. The Newton diagrams at each Ecol
,
which are constructed from the beam characterization, are
employed for image data analysis.
The O(1D) beam was generated by F2 laser photolysis at 157
nm (LPF 220, Lambda Physik) of the O2 precursor (1% O2 in
He). The scattering Cl(2P3/2) signal was detected with the UV
probe laser pulse of 200 µJ/pulse. The lower probe laser intensity
than that in our previous study (400 µJ/pulse) suppressed the
one-color dissociation/ionization noise from the HCl target
beam; the spot-like HCl+ noise image is less than 5% of the
Cl+ signal from the titled reaction, which is sufficiently small
to obviate subtraction of this particular noise image from the
signal image. As shown later, we still needed subtracting of
another type of noise arising from the O atom beam source.
The count rate of the Cl+ signal from the reactive scattering
was ∼20 counts per laser shot at most. The probe laser
wavelength was scanned over 2.7 cm-1 around the two-photon
resonant line of Cl(2P3/2) f Cl(2D3/2) at 235.326 nm. This
scanning range covers the Doppler components of (9500 m/s,
which is sufficiently wide to detect the entire range of scattering
velocities. The images of Cl+ at each Ecol are accumulated for
∼1 h. These images at the three Ecol are measured under the
same detection conditions. The signal intensities with the Ne-
seeded (Ecol ) 4.5 kcal/mol) and the Ar-seeded (Ecol ) 4.2
kcal/mol) HCl beams are ∼80 and ∼40% of that with the He-
seeded HCl sample (Ecol ) 6.4 kcal/mol), respectively. These
ratios do not correctly reflect the σreac(Ecol) but the difference
of the effective HCl beam density with the carrier gases (He,
Ne, and Ar). Resultantly, the signal-to-noise ratio of the present
data decreases in the descending order of the collision energy.
The observed image is 2D projection data of a 3D density
distribution of the Cl(2P3/2) products at the instance of laser
irradiation. The first step of image data analysis is density-flux
transformation using an apparatus function given by the temporal
and geometrical configurations of the pulsed atomic/molecular
beams and the laser beam.36 The apparatus function was
numerically simulated with the experimental parameters de-
scribing the spatial and temporal profiles of the atomic/molecular
beams and the laser beam. The degree of correction in the
density-flux transformation was rather small, as the velocity of
the Cl product (VCl) in this reaction was relatively small
(maximum VCl < 1800 m/s). The doubly (velocity-angle) DCSs
are obtained for the three Ecol values by an inverse Abel
transform of the density-flux corrected images.
Theoretical Calculation
The global PESs of the 11A′, 11A′′, and 21A′ states were
obtained in our previous study by ab initio calculation imple-
mented with the CASSCF-MRCI method by the MOLPRO suite
of programs.37 The detail of the quantum chemistry calculation
is described in the original paper.15 The MRCI calculation was
executed on 4631 molecular geometries; the ranges of the bond
lengths are R(O-H) ) 0.6-5.0 Å, R(Cl-O) ) 1.2-6.0 Å, and
R(H-Cl) ) 0.8-5.0 Å. The interpolant moving least-squares
method combined with Shepard interpolation developed by
Ishida and Schatz38 was applied to generate the fitted global
PESs, which are employed for the present QCT dynamical
Experimental Section
The details of the experimental setup are described else-
where.19 In the present experiment, we varied the speed of the
HCl molecular beam by using He, Ne, and Ar as carrier gases.
The HCl was diluted in each carrier gas by 5% in concentration.
The time profile of the pulsed HCl molecular beam was
measured by a fast ion gauge (FIG-1, Beam Dynamics, Inc.)
140 and 280 mm downstream from the collision center. We

820 J. Phys. Chem. A, Vol. 112, No. 5, 2008
Kohguchi et al.
calculations. The QCT calculations are carried out independently
on the 11A′, 11A′′, and 21A′ adiabatic PESs at Ecol’s of 3.0,
4.0, 6.0, 12.2, and 15.0 kcal/mol. The range of Ecol ) 3.0-6.0
kcal/mol is chosen for comparison with the experimental data,
while the trajectories at higher Ecol are also examined for further
surveillance of the Ecol dependence of the excited-state dynam-
ics. The couplings between these PESs are not taken into
account. Each trajectory is run with a specific Ecol for the (V, j)
) (0, 0) rovibrational state of HCl. All other variables specifying
the initial conditions are chosen randomly by Monte Carlo
sampling. The equation of motion was integrated by the fourth-
order Runge-Kutta method on each adiabatic PES. Propagation
of the trajectory was terminated when one of the three
internuclear distances exceeded 8 Bohr (4.23 Å). The number
of the trajectories was ∼20 000 for each PES. The destinations
of the trajectories are grouped into the OH + Cl, ClO + H,
and O + HCl (inelastic) pathways. Among them, the trajectories
ending with the OH + Cl pathway are compared with the
experimental results.
Results and Discussion
A. Scattering Experiment. The ion images of the Cl(2P3/2
)
product observed at the collision energies of 6.4, 4.5, and 4.2
kcal/mol are shown in Figure 1a, b, and c, respectively. The
images are placed so that the collision axis (thin black line) is
aligned upright. The Newton diagram is superimposed on each
image, where the velocity vectors of the HCl and the O(1D)
beams are depicted by thick arrows. The center-of-mass velocity
is shown by a thin arrow, whose end points at the center-of-
mass. The spot-like noise, which is due to byproducts contami-
nated in F2 laser photolysis, is seen at the end of the O(1D)
beam velocity vector. This noise image is eliminated prior to
image data analysis. The production of the Cl*(2P1/2) is a minor
spin-orbit pathway in the present collision energy region.17,39
In addition, it seems that the spin-orbit branching can occur
in the asymptotic product (OH + Cl) region;19 therefore, both
of the excited- and ground-state reaction channels can be
monitored by Cl(2P3/2). This gives us some justification for using
the data of Cl(2P3/2) to discuss the overall reaction dynamics
on the ground and excited electronic states in the rest of this
paper.
The center-of-mass translational energy releases, P(Et),
obtained from these images are shown in Figure 2a for an Ecol
of 6.4 kcal/mol, in b for 4.5 kcal/mol, and in c for 4.2
kcal/mol. The available energies (Eavl) are almost the same, (a)
17 800, (b) 17 100, and (c) 17 000 cm-1, because of a large
exothermicity of the reaction. Correspondingly, the obtained
average translational energy release [<Et> ) ∑EtP(Et)/P(Et)]
is similar for the three cases, that is, (a) 3830 (170), (b) 3820
(270), and (c) 3590 (210) cm-1 for Ecol ) 6.4, 4.5, and 4.2
kcal/mol, respectively. The values in the parentheses are the
differences between the results with/without sensitivity correc-
tion for the observed images, which were found to be unim-
portant in the following discussion. The observed P(Et) values
at the three Ecol’s yield a small energy partitioning to the
translational energy (<fT>) of 22 ( 2%, exhibiting formation
of highly internally excited OH. The strong internal excitation
of OH has been well-known both in the experimental16,18,19,40,41
and theoretical23-29,33-35,42 studies as a distinctive attribute of
the O(1D) + HCl reaction. Our P(Et) agrees very well with the
state distribution of the counterpart product OH reported by Park
and Wiesenfeld.16 For comparison, we calculated P(Et) from
the triangular plot reported by Park and Wiesenfeld, as presented
in a dashed curve in Figure 2a. Note that their experiment was
Figure 1. Observed scattering images of Cl(2P3/2) products in the O(1D)
+ HCl f OH + Cl reaction. The collision energies are (a) 6.4, (b)
4.5, and (c) 4.2 kcal/mol. The geometry of the HCl and O(1D) beams
is shown by the Newton diagram superimposed in each image. The
HCl and O(1D) beam velocity vectors, which are crossed at right angles,
are designated by the thick arrows. The center-of-mass velocity shown
by a thin arrow points to the scattering center on the collision axis,
which is represented by a vertical black line. The spot-like noise at the
end of the O(1D) beam velocity is due to byproducts in laser photolysis.
The direction of the probe laser beam bisects the included right angle
between the HCl and the O(1D) beams in all images.
carried out with a rather broad collision energy spread (average
Ecol ) 7 kcal/mol), while the present experiment was carried
out at a well-defined collision energy. Despite this difference
in conditions, the P(Et)’s distribute mostly in the low Et
(<10 000 cm-1) region in both cases and are consistent with
each other.

Collision Energy Dependence of O(1D) + HCl f OH + Cl(2P)
J. Phys. Chem. A, Vol. 112, No. 5, 2008 821
Figure 2. The center-of-mass translational energy release of the O(1D)
+ HCl f OH + Cl reaction at the collision energies (Ecol) of (a) 6.4,
(b) 4.5, and (c) 4.2 kcal/mol. The dashed curve in (a) is obtained from
the rovibrational state distribution of OH measured by Park and
Wiesenfeld,16 for which the translational energy was determined for
the average collision energy of 7.4 kcal/mol. All of the scattering
angular components are integrated. All of the graphs are normalized
to have a certain common area. The available energies (Eavl ) ∆H +
Figure 3. The differential cross section of the O(1D) + HCl f OH +
Cl reaction at the collision energies of (a) 6.4, (b) 4.5, and (c) 4.2 kcal/
mol. All of the scattering velocity components are integrated. All of
the graphs are normalized to have a certain common area for
comparison.
the relation of T(θcm ) 180°)/T(θcm ) 0°) ) exp(-τrot/2τreac),
where T(θcm) is the experimental DCS. The rotational period
of τrot ) 2πI/leff is evaluated by assuming the moment of inertia
(I) at the equilibrium geometry of the HOCl or HClO potential
Ecol) are denoted by an arrow in each plot.
The DCSs extracted from the slice (Abel inverted) image data
well and the effective angular momentum of leff ) µVrelbpeak,
at the three Ecol’s are presented in Figures 3a-c. These are the
angular distributions integrated over the whole scattering
velocity (or Et) component at Ecol ) 6.4 kcal/mol (Figure 3a),
4.5 kcal/mol (Figure 3b), and 4.2 kcal/mol (Figure 3c). As is
expected from the observed images (Figure 1), the DCSs at the
three Ecol’s are peaking in the forward and backward directions,
while the sideways components with the center-of-mass scat-
tering angle (θcm) from 30 to 150° amount to less than 50% of
the total intensity. The DCS at Ecol ) 6.4 kcal/mol (Figure 3a)
shows pronounced preference for forward scattering; the
intensity of the forward components (0 < θcm < 30°) is more
than twice the backward (150 < θcm < 180°) ones. The
preference slightly decreases in the lower Ecol (4.5 kcal/mol in
where µ is the reduced mass, Vrel is the relative velocity, and
bpeak is the impact parameter (b) at which the opacity function
[P(b)] without a factor of 2πbdb is maximized. The moments
of inertia of I(HOCl) ) 35.0 amu Å2 and I(HClO) ) 29.4 amu
Å2 are obtained from the ab initio results by Nanbu et al.15 The
representative value of bpeak ) 2.5 Å is used on the basis of the
QCT results, which are discussed later. An almost identical value
of bpeak is presented in the QCT study at Ecol ) 12.2 kcal/mol
by Mart´ınez et al.23 The geometry at the HOCl well yields the
evaluation of τrot ) 360, 430, and 440 fs for Ecol ) 6.4, 4.5,
and 4.2 kcal/mol, respectively. Consequently, the osculating
complex model provides τreac ) 340, 940, and 1070 fs for Ecol
) 6.4, 4.5, and 4.2 kcal/mol, respectively. The assumption of
the collision complex at the HClO well geometry yields shorter
τreac by ∼15% than those at the HOCl well for each Ecol
condition. The above estimation employs the moment of inertia
of a stable molecule, while a short-lived collision complex
undergoes a large-amplitude motion during chemical reaction.
Therefore, this analysis should be regarded qualitatively. Despite
such a limitation, the osculating complex model, which implic-
itly assumes the ground-state reaction, well explains the trend
of Ecol dependence of the observed DCS. Since the enhancement
Figure 3b and 4.2 kcal/mol in Figure 3c). The DCSs at Ecol
)
4.2 and 4.5 kcal/mol, which are overlapped each other within
the ∆Ecol, are almost identical, except for the lower signal
statistics of the data with the Ar buffer beam (Ecol ) 4.2
kcal/mol). The observed preference of the forward scattering
with the increasing Ecol is explained by the osculating complex
model for chemical reactions.43,44 The model provides an
estimate of the reaction time (τreac), which is a fraction of, or
comparable to, the rotational period of the complex (τrot) by

822 J. Phys. Chem. A, Vol. 112, No. 5, 2008
Kohguchi et al.
Figure 5. The schematic potential energy diagram of the O(1D) +
HCl reaction. The electronic energy for the saddle points and the
equilibrium structures are shown in the units of kcal/mol. The electronic
states correlating the saddle points and the equilibrium structures are
designated by thick black lines for 11A′, thin black lines for 11A′′, and
thick gray lines for 21A′.
along the forward/backward directions. While the Ecol-dependent
forward-backward ratio is discussed in the former paragraph,
no significant Ecol dependence is found for the sideways
components in the VCl-θcm plot. The polarized scattering
distribution is presented in the previous experimental19 and
theoretical23 studies of the O(1D) + HCl reaction; the observed
correlation between the forward θcm and the slow VCl is ascribed
to a peripheral dynamics in which a light atom is transferred
between two heavier atoms separated at a large distance, leading
to strong excitation of OH and a forward scattering. The QCT
calculations by Mart´ınez et al.23 on the 11A′ PES at Ecol ) 12.2
kcal/mol indicated the delayed emission of products in the
backward direction with respect to the forward direction. In the
osculating complex regime of an insertion reaction, the reaction
time is a fraction of a rotational period. Our result indicates
that back-scattered products have higher recoil velocity, indicat-
ing weaker internal excitation in OH.19 The results accord well
with the theoretical prediction by Mart´ınez et al.,23 if de-
excitation of the OH stretching vibration occurs due to mode
coupling within the lifetime of a collision complex that is still
shorter than its rotational period (0.3 ps). Thus, all of the
experimental results suggest that the ground-state reaction
dominates the titled reaction at the three Ecol’s, although these
exceed the potential barrier heights predicted by the CASSCF-
MRCI calculation, as is described in the next section.
Figure 4. The doubly differential cross sections of the O(1D) + HCl
f OH + Cl reaction at the collision energies of (a) 6.4, (b) 4.5, and
(c) 4.2 kcal/mol. The abscissa and ordinate represent the center-of-
mass scattering angle and the product Cl velocity, respectively. The
upward direction points to the forward component. A fixed pattern noise
seen in the low-speed region (<400 m/s) is due to numerical artifact
in the conversion of the small radius to the polar coordinate.
of the forward scattering at the higher Ecol is considered to be
a rather contrary indication to the contribution of the excited
state PESs, the observed Ecol dependence of the DCS confirms
the dynamics predominantly on the ground-state PES. In
addition, the preference of the forward components in all three
Ecol’s is in good agreement with the QCT results on the ground-
state PES, as is described in the following section.
B. Quasiclassical Trajectory Calculations on the Ab Initio
PESs. We will briefly review the ab initio PESs on which QCT
calculations were performed. The detailed discussion on the
topography of the 11A′′ and 21A′ excited-state PESs was made
by Nanbu et al.15 The schematic energy diagram is shown in
Figure 5. Both of these PESs are globally repulsive, with no
bound-state region except for a shallow van der Waals well on
the 21A′ PES. The saddle point structures on the 11A′′ and 21A′
PESs obtained by the ab initio calculations are shown in Figure
6. The 11A′′ PES has a saddle point at a bent geometry; the
two-dimensional contour plot of the electronic energy with a
fixed H-Cl bond length of 1.31 Å is shown in Figure 6a. The
saddle point is located at the OH bond length of R(O-H) )
1.70 Å and the O-H-Cl angle of 129°. The barrier height from
the reactant arrangement is 1.6 kcal/mol. The 21A′ PES also
has a saddle point but at the O-H-Cl linear configuration
shown in Figure 6b. The barrier, whose height is 6.5 kcal/mol,
is located at R(O-H) ) 1.35 Å and R(H-Cl) ) 1.34 Å. Another
saddle point on the 11A′′ PES is found at the H-Cl-O linear
The doubly (θcm and VCl) DCSs in the polar coordinate
representation are shown in Figure 4; Ecol ) 6.4 kcal/mol in
Figure 4a, Ecol ) 4.5 kcal/mol in Figure 4b, and Ecol ) 4.2
kcal/mol in Figure 4c. An inaccurate intensity in the neighbors
of θcm ) 0 and 180° is unavoidable in the numerical Abel
inversion. A fixed pattern noise, which is apparently seen in
the slow speed region (VCl < 400 m/s) arises from the Cartesian-
to-polar transformation of pixelized image data. These artifacts
due to the numerical operations are neglected in the following
discussion. The doubly DCSs in Figure 4a-c show the VCl-
θ
cm correlation (nonseparability45 of VCl and θcm), as is expected
from the polarized (elongated) feature of the observed images

Collision Energy Dependence of O(1D) + HCl f OH + Cl(2P)
J. Phys. Chem. A, Vol. 112, No. 5, 2008 823
Figure 6. The contour plots of the electronic energy of the excited-
state PESs at the fixed H-Cl bond length obtained by the ab initio
calculation15: (a) the 11A′′ state with the fixed H-Cl bond length at
1.31 Å; (b) the 21A′ state with the fixed H-Cl bond length at 1.34 Å;
and (c) the 11A′′ state with the fixed H-Cl bond length at 1.39 Å. The
saddle points are designated by × in each two-dimensional plot. The
step of the energy contour is 1 kcal/mol. The saddle points in (a) and
(b) represent the potential barriers from the O(1D) + HCl reactant to
the OH + Cl pathway. The saddle point in plot (c) shows the barrier
from the reactant to the ClO + H pathway.
geometry. This is a saddle point from the reactant to the ClO
+ H pathway with the barrier height of 16.1 kcal/mol. The elec-
tronic energy around the saddle point is shown in Figure 6c.
The results of the QCT calculations executed on the 11A′,
11A′′, and 21A′ PESs are shown in Figures 7 and 8. The
theoretical DCSs at various Ecol are shown in Figure 7; the
results on the 11A′ are shown in Figure 7a, 11A′′ in 7b, and
21A′ PESs in 7c. Unsmooth curves of the DCS of the excited
states (Figure 7b and c) are due to the smaller number of
resultant trajectories in the OH + Cl pathway than that of the
ground state (Figure 7a). The overall feature of the QCT result
on the 11A′ PES (Figure 7a) is in good agreement with the
observed DCS. The theoretical DCS on the ground-state PES
exhibits the polarized distribution peaked at θcm ) 0 and 180°
with a discernible preference in the forward direction. The
intensity of the sideways components is considerably smaller
than that of the forward and backward ones. A qualitatively
good agreement in the DCS is seen between the present and
previous23 QCT calculations at Ecol ) 12.2 kcal/mol, despite
the fact that independent ab initio data set of the 11A′ PES was
employed. No marked Ecol dependence is found in Figure 7a,
except for the slight and nonmonotonic difference with respect
to Ecol. The variation of the calculated DCS within Ecol ) 3.0-
6.0 kcal/mol is as small as the present experimental accuracy.
Although the ground-state PES has been corroborated to have
a deep double-well structure by ab initio calculations,12,15,20-22
the fast dynamics without being trapped in the potential wells
have been revealed,19,23 resulting in the larger intensity in the
forward direction. It is expected that the reaction mechanism is
not significantly affected by the small difference of Ecol ) 3-6
Figure 7. The theoretical differential cross sections of the O(1D) +
HCl f OH + Cl reaction obtained by the QCT calculation executed
independently on (a) the 11A′ ground-state PES, (b) the 11A′′ excited-
state PES, and (c) the 21A′ excited-state PES. The collision energy for
each curve is designated with different colors in the legend, black for
3.0 kcal/mol, red for 4.0 kcal/mol, green for 6.0 kcal/mol, purple for
12.2 kcal/mol, and light blue for 15.0 kcal/mol.
kcal/mol because of the large exothermicity (∆H0 ) -44.4
kcal/mol) and the deep well structures (HOCl well depth: 101
kcal/mol) of the 11A′ PES with no barrier. It is noted that the
small Ecol dependence of the theoretical DCS, especially the
ratio between the forward and backward peaks, differs from
the monotonic change expected from the osculating complex
model. However, the present QCT results illustrate that the
ground-state PES solely causes only a small variation of the
DCS within a few kcal/mol width of Ecol
.
For discrimination between the reactions on the ground- and
excited-state PESs, it is valuable to inspect an individual feature
of the DCS on each PES prior to consideration of the coupling
between PESs. The theoretical DCS on the 11A′′ PES (Figure
7b) shows a drastic change between Ecol ) 6.0 and 12.2
kcal/mol, the small intensity without the forward components
at Ecol ) 6.0 kcal/mol and the increased intensity with a peak
around θcm ) 20° at Ecol ) 12.2 kcal/mol. The characteristic
peak of the DCS is not simply explained by the height of the
saddle point (1.6 kcal/mol) on the 11A′′ PES since it is bent. It
seems that this bent saddle point does not play any significant

824 J. Phys. Chem. A, Vol. 112, No. 5, 2008
Kohguchi et al.
by the integral of 2πbP(b) up to bpeak, where bpeak corresponds
to the effective reaction barrier. Assuming that the bpeak is given
by the saddle point, bpeak ∼ 1.0 Å for the 11A′′ PES and bpeak
∼ 0 Å for the 21A′ PES are obtained from the ab initio
calculations. The former value is obtained from the bent
structure in the reactant region [R(H-Cl) ) 1.31 Å compared
with 1.27 Å of an isolated HCl],47 and the latter are from the
collinear saddle point. The small values of bpeak of the excited-
state PESs are markedly contrasted with the bpeak ) 2.5 Å of
the 11A′ PES, resulting in the considerably smaller σreac(11A′′)
and σreac(21A′) than σreac(11A′).
Conclusion
Figure 8. The theoretical reaction cross sections of the O(1D) + HCl
f OH + Cl reaction at the collision energies (Ecol) of 3.0, 4.0, 6.0,
12.2, and 15.0 kcal/mol. These are the results obtained by QCT
calculation executed independently on the 11A′(9), 11A′′(O), and
21A′(4) PESs. The data for the excited states at Ecol ) 3.0 and 4.0
kcal/mol are omitted because of their negligibly small magnitudes. The
solid lines are guides for eyes.
The collision energy dependence of the O(1D) + HCl f OH
+ Cl(2P) reaction has been investigated both experimentally and
theoretically. The doubly DCSs measured by the crossed
molecular beam experiment in the Ecol ) 4.2-6.4 kcal/mol
region exhibited nearly forward-backward symmetric distribu-
tion, with a pronounced forward component. The forward
scattering was enhanced at the higher collision energies. The
previous CASSCF-MRCI ab initio calculations predicted the
existence of the reaction barrier of the 11A′′ (1.6 kcal/mol) and
21A′ (6.5 kcal/mol) excited-state PESs. The QCT calculations
were carried out separately on the ground-state (11A′) and the
two excited-state (11A′′, 21A′) ab initio PESs. The calculated
partial reaction cross section of the 11A′ ground-state PES is
much larger than those of the 11A′′ and 21A′ states. This is
because the excited-state pathways have reaction barriers and
small cones of acceptance, while interactions in the ground-
state pathway are attractive and provide a much larger cone of
acceptance. The theoretical DCS of the ground-state 11A′ PES
is in good agreement with the experimental DCS. The present
experimental and theoretical results conclude the dominant role
of the reaction pathway via the ground-state 11A′ PES at the
collision energy below 6.5 kcal/mol.
role as a typical transition state in the dynamics. Another saddle
point at Ecol ) 16.1 kcal/mol on the 11A′′ PES (Figure 6c) does
not give any account either because it leads to the ClO + H
pathway. At higher energies, the major reaction mechanism on
the 11A′′ PES is speculated to occur over a wider interaction
region than that of the potential barrier itself. On the other hand,
the QCT results of the 21A′ PES is consistent with a na¨ıve
expectation on the topographical basis of the PES; the saddle
point associated with the collinear geometry at Ecol ) 6.5
kcal/mol yields the DCS which is characterized by backward
scattering, as shown in Figure 7c. The collinear saddle point
shown in Figure 6b is recognized to serve as an ordinary type
of transition state. We found that the sum of the DCSs of the
ground and the two excited states is almost the same as that of
the ground state, indicating that contributions from the excited
11A′′ and 21A′ states are negligible at the experimentally
investigated energies.
Acknowledgment. This work was financially supported by
Grant-in-Aid for Specially Promoted Research of the Ministry
of Education, Culture, Sports, Science and Technology of Japan
(No. 15002011). A part of the numerical calculations for the
apparatus functions was carried out with use of the computer
facilities at the RIKEN Super Combined Cluster (RSCC).
The reaction cross sections (σreac) on each PES are shown in
Figure 8. The σreac in the units of Å2 is obtained by summing
over each theoretical DCS with a multiplying factor of 2πsinθcm.
The σreac’s of the 11A′, 11A′′, and 21A′ states are independently
calculated at particular Ecol’s, which correspond to the partial
excitation functions for the specific PESs, σreac(Ecol; state). The
σ
reac(Ecol; 11A′′) and σreac(Ecol; 21A′) at Ecol ) 3.0 and 4.0 kcal/
References and Notes
mol are omitted since they are negligibly small. These theoretical
reac(Ecol; state) provide a quantitative basis for the small
(1) Cheng, B. M.; Bahou, M.; Chen, W. C.; Yui, C. H.; Lee, Y. P.;
Lee, L. C. J. Chem. Phys. 2002, 117, 1633.
(2) Dobbyn, A. J.; Knowles, P. J. Mol. Phys. 1997, 91, 1107.
(3) Kassab, E.; Gleghorn, J. T.; Evleth, E. M. J. Am. Chem. Soc. 1983,
105, 1746.
(4) Lin, J. J.; Harich, S.; Lee, Y. T.; Yang, X. J. Chem. Phys. 1999,
110, 10821.
(5) Suzuki, T.; Hirota, E. J. Chem. Phys. 1993, 98, 2387.
(6) Hermine, P.; Hsu, Y. T.; Liu, K. Phys. Chem. Chem. Phys. 2000,
2, 581.
σ
contribution of the 11A′′ and 21A′ excited states. The σreac(Ecol
;
11A′) behaves in a decreasing manner, which is typical for an
attractive PES without a barrier. The relatively large magnitude
(>10 Å2) is naturally due to the widely extended interaction
region over the double wells of the 11A′ PES. The values of
σ
reac(Ecol; 11A′) in the same Ecol region are consistent with the
previous dynamics calculations by other authors, although the
individual ab initio PES data for the 11A′ PES were em-
ployed.26,28-30,35 The QCT results predict that the contribution
of the excited states emerges in the Ecol > 12 kcal/mol region
and becomes important probably at Ecol > 20 kcal/mol. The
ratio of σreac of the two excited states to that of the ground state,
[σreac(11A′′) + σreac(21A′)]/σreac(11A′)] is 0.03 at Ecol ) 6.0 kcal/
mol and 0.23 at Ecol ) 12.2 kcal/mol. Although we also carried
out the QCT calculations including tunneling effects46 for all
of the states and energies, the contribution due to tunneling
effects was found to be negligibly small. A simple explanation
based on the ab initio results is given to the relatively small
(7) Hsu, Y. T.; Liu, K.; Pederson, L. A.; Schatz, G. C. J. Chem. Phys.
1999, 111, 7931.
(8) Hsu, Y. T.; Wang, J. H.; Liu, K. J. Chem. Phys. 1997, 107, 2351.
(9) Schatz, G. C.; Papaioannou, A.; Pederson, L. A.; Harding, L. B.;
Hollebeek, T.; Ho, T. S.; Rabitz, H. J. Chem. Phys. 1997, 107, 2340.
(10) Hsu, Y. T.; Liu, K.; Pederson, L. A.; Schatz, G. C. J. Chem. Phys.
1999, 111, 7921.
(11) Hirsch, G.; Bruna, P. J.; Peyerimhoff, S. D.; Buenker, R. J. Chem.
Phys. Lett. 1977, 52, 442.
(12) Bruna, P. J.; Hirsch, G.; Peyerimhoff, S. D.; Buenker, R. J. Can.
J. Chem. 1979, 57, 1839.
(13) Nanbu, S.; Iwata, S. J. Phys. Chem. 1992, 96, 2103.
(14) Guo, H. J. Phys. Chem. 1993, 97, 2602.
(15) Nanbu, S.; Aoyagi, M.; Kamisaka, H.; Nakamura, H.; Bian, W.;
Tanaka, K. J. Theor. Comput. Chem. 2002, 1, 263.
(16) Park, C. R.; Wiesenfeld, J. R. Chem. Phys. Lett. 1989, 163, 230.
σ
reac(11A′′) and σreac(21A′). A classical estimate of σreac is given

Collision Energy Dependence of O(1D) + HCl f OH + Cl(2P)
J. Phys. Chem. A, Vol. 112, No. 5, 2008 825
(17) Matsumi, Y.; Tonokura, K.; Kawasaki, M.; Tsuji, K.; Obi, K. J.
Chem. Phys. 1993, 98, 8330.
(34) Lagana, A.; Deaspuru, G. O.; Garcia, E. J. Phys. Chem. 1995, 99,
17139.
(18) Kruus, E. J.; Niefer, B. I.; Sloan, J. J. J. Chem. Phys. 1988, 88,
985.
(19) Kohguchi, H.; Suzuki, T. ChemPhysChem 2006, 7, 1250.
(20) Peterson, K. A.; Skokov, S.; Bowman, J. M. J. Chem. Phys. 1999,
111, 7446.
(21) Hernandez, M. L.; Redondo, C.; Lagana, A.; deAspuru, G. O.; Rosi,
M.; Sgamellotti, A. J. Chem. Phys. 1996, 105, 2710.
(22) Lee, T. J. J. Phys. Chem. 1994, 98, 3697.
(23) Mart´ınez, T.; Hernandez, M. L.; Alvarino, J. M.; Aoiz, F. J.;
Rabanos, V. S. J. Chem. Phys. 2003, 119, 7871.
(24) Kamisaka, H.; Nakamura, H.; Nanbu, S.; Aoyagi, M.; Bian, W.;
Tanaka, K. J. Theor. Comput. Chem. 2002, 1, 285.
(25) Kamisaka, H.; Nakamura, H.; Nanbu, S.; Aoyagi, M.; Bian, W.;
Tanaka, K. J. Theor. Comput. Chem. 2002, 1, 275.
(26) Christoffel, K. M.; Bowman, J. M. J. Chem. Phys. 2002, 116, 4842.
(27) Piermarini, V.; Lagana, A.; Balint-Kurti, G. G. Phys. Chem. Chem.
Phys. 2001, 3, 4515.
(35) Schinke, R. J. Chem. Phys. 1984, 80, 5510.
(36) Kohguchi, H.; Suzuki, T.; Alexander, M. H. Science 2001, 294,
832.
(37) MOLPRO is a package of ab initio programs written by Werner,
H.-J.; Knowles, P. J.; Schu¨etz, M.; Lindh, R.; Celani, P.; Korona, T.; Rauhut,
G.; Manby, F. R.; Amos, R. D.; Bernhardsson, A.; Berning, A.; Cooper,
D. L.; Deegan, M. J. O.; Dobbyn, A. J.; Eckert, F.; Hampel, C.; Hetzer,
G.; Lloyd, A. W.; McNicholas, S. J.; Meyer, W.; Mura, M. E.; Nicklass,
A.; Palmieri, P.; Pitzer, R.; Schumann, U.; Stoll, H.; Stone, A. J.; Tarroni,
R.; Thorsteinsson, T. Version: 2006.1, URL: http://www.molpro.net.
(38) Ishida, T.; Schatz, G. C. Chem. Phys. Lett. 1999, 314, 369.
(39) Chichinin, A. I. J. Chem. Phys. 1997, 106, 1057.
(40) Luntz, A. C. J. Chem. Phys. 1980, 73, 5393.
(41) Basco, N.; Norrish, R. G. W. Proc. R. Soc. London. Ser. A 1960,
260, 293.
(42) Christoffel, K. M.; Kim, K.; Skokov, S.; Bowman, J. M.; Gray, S.
(28) Piermarini, V.; Balint-Kurti, G. G.; Gray, S. K.; Gogtas, F.; Lagana,
A.; Hernandez, M. L. J. Phys. Chem. A 2001, 105, 5743.
(29) Mart´ınez, T.; Hernandez, M. L.; Alvarino, J. M.; Lagana, A.; Aoiz,
F. J.; Menendez, M.; Verdasco, E. Phys. Chem. Chem. Phys. 2000, 2, 589.
(30) Lin, S. Y.; Han, K. L.; Zhang, J. Z. H. Phys. Chem. Chem. Phys.
2000, 2, 2529.
(31) Lin, S. Y.; Han, K. L.; Zhang, J. Z. H. Chem. Phys. Lett. 2000,
324, 122.
(32) Bittererova, M.; Bowman, J. M. J. Chem. Phys. 2000, 113, 1.
(33) Alvarino, J. M.; Rodriguez, A.; Lagana, A.; Hernandez, M. L. Chem.
Phys. Lett. 1999, 313, 299.
K. Chem. Phys. Lett. 1999, 315, 275.
(43) Alagia, M.; Balucani, N.; Casavecchia, P.; Stranges, D.; Volpi, G.
G. J. Chem. Phys. 1993, 98, 8341.
(44) Fisk, G. A.; McDonald, J. D.; Herschbach, D. R. Discuss. Faraday
Soc. 1967, 44, 228.
(45) Che, D.; Liu, K. J. Chem. Phys. 1995, 103, 5164.
(46) Oloyede, P.; Mil’nikov, G.; Nakamura, H. J. Theor. Comput. Chem.
2004, 3, 91.
(47) Herzberg, G. Molecular Spectra and Molecular Structure I. Spectra
of Diatomic Molecules, 2nd ed.; Krieger: Malabar, Florida, 1950.