Stereodynamics of the vibrational channel O(1D) + H2O→OH(v′ = 2) + OH

Article abstract of DOI:10.1063/1.478682

The state-selected differential cross section (DCS) and rotational angular momentum polarization for the reaction O(¹D) + H₂O→OH+OH have been measured by utilizing the polarized Doppler-resolved laser-induced fluorescence probing technique. Stereodynamics of the reaction channel forming the newly formed OH in the specific vibrational level ν′ = 2 is discussed on the basis of the vector properties. A nearly isotropic DCS for the product OH(²Π_3/2, ν′ = 2, j′ = 5.5) most probably indicates that the reaction is dominated by an insertion mechanism involving a collisional HOOH complex with a lifetime comparable to its rotational period. The extremely asymmetrical energy partitioning between the two OH fragments, therefore, suggests that the redistribution of the available energy does not occur on a time scale comparable to the rotational period of the complex. Furthermore, it has been found that the product rotational angular momentum vector j′ is predominantly perpendicular to the collision plane spanned by k and k′ (the relative velocity vectors of the reactants and products, respectively) both for the forward- and backward-scattered products. It suggests that the initially excited bending motion of the H-O-O moiety in the collisional HOOH complex primarily contributes to the product rotation.

Full text of DOI:10.1063/1.478682

Stereodynamics of the vibrational channel O ( 1 D)+ H 2 O→OH (v ′ =2)+ OH
Hiroshi Tsurumaki, Yo Fujimura, and Okitsugu Kajimoto
Citation: The Journal of Chemical Physics 110, 7707 (1999); doi: 10.1063/1.478682
View online: http://dx.doi.org/10.1063/1.478682
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/110/16?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Relaxation of H 2 O from its |04 − vibrational state in collisions with H 2 O , Ar, H 2 , N 2 , and O 2
J. Chem. Phys. 120, 5592 (2004); 10.1063/1.1649726
The branching ratio between reaction and relaxation in the removal of H 2 O from its |04 − vibrational state in
collisions with H atoms
J. Chem. Phys. 115, 4586 (2001); 10.1063/1.1389304
The dynamics of the reactions H+H 2 O→OH+H 2 and H+D 2 O→OD+HD at 1.4 eV
J. Chem. Phys. 114, 6690 (2001); 10.1063/1.1356008
The product rovibrational and spin–orbit state dependent dynamics of the complex reaction H+CO 2 →OH( 2
Π;ν,N,Ω, f)+ CO : Memories of a lifetime
J. Chem. Phys. 112, 4557 (2000); 10.1063/1.481081
Translational to vibrational and rotational (T→V,R) energy transfer and reactive exchange collisions of H ( D )+
HF ( DF ) in the energy range from 1 to 2 eV by time-resolved Fourier transform spectroscopy
J. Chem. Phys. 106, 2265 (1997); 10.1063/1.473149
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

JOURNAL OF CHEMICAL PHYSICS
VOLUME 110, NUMBER 16
22 APRIL 1999
1
Stereodynamics of the vibrational channel O„ D…؉H O˜OH„v ؍2…؉OH
Ј
2
Hiroshi Tsurumaki, Yo Fujimura, and Okitsugu Kajimoto
Department of Chemistry, Graduate School of Science, Kyoto University, Kitashirakawa-Oiwakecho,
Sakyo-ku, Kyoto 606-8502, Japan
͑Received 30 October 1998; accepted 25 January 1999͒
The state-selected differential cross section ͑DCS͒ and rotational angular momentum polarization
for the reaction O(1D)ϩH2O→OHϩOH have been measured by utilizing the polarized
Doppler-resolved laser-induced fluorescence probing technique. Stereodynamics of the reaction
channel forming the newly formed OH in the specific vibrational level ϭ2 is discussed on the
v
Ј
basis of the vector properties. A nearly isotropic DCS for the product OH(2⌸3/2
,
ϭ2, j
v
Ј Ј
ϭ5.5) most probably indicates that the reaction is dominated by an insertion mechanism involving
a collisional HOOH complex with a lifetime comparable to its rotational period. The extremely
asymmetrical energy partitioning between the two OH fragments, therefore, suggests that the
redistribution of the available energy does not occur on a time scale comparable to the rotational
period of the complex. Furthermore, it has been found that the product rotational angular momentum
vector j is predominantly perpendicular to the collision plane spanned by k and k ͑the relative
Ј
Ј
velocity vectors of the reactants and products, respectively͒ both for the forward- and
backward-scattered products. It suggests that the initially excited bending motion of the H–O–O
moiety in the collisional HOOH complex primarily contributes to the product rotation. © 1999
American Institute of Physics. ͓S0021-9606͑99͒00816-8͔
I. INTRODUCTION
plex breaking up prior to the redistribution of the available
energy among its internal modes.5
The reaction of electronically excited oxygen atom,
O(1D) with H2O plays an important role as a source of OH
radicals in the stratosphere. The two OH product molecules
produced from this reaction are chemically identical but their
origins are distinctly different, i.e., ‘‘new’’ OH formed in the
reaction and ‘‘old’’ OH present in the H2O reagent from the
beginning. To distinguish these two types of OH products for
the detailed analysis of the reaction dynamics, the reaction
with isotopically labeled reagent H218O,
To determine whether the reaction proceeds via direct
abstraction, stripping or insertion pathways, the measure-
ment of product differential cross section ͑DCS͒ is most de-
sirable. Along this line, King et al. measured the LAB frame
speeds of both the new and old OH products by utilizing
polarized Doppler spectroscopy.7 They fit the Doppler pro-
files representing the product LAB speed distribution with a
single parameter of average LAB kinetic energy and esti-
mated a representative average center-of-mass ͑CM͒ scatter-
ing angle under several assumptions. Their observation of
broader linewidth of the new OH than that of the old OH was
explained by the strong scattering anisotropy in the CM
frame. Based on such analysis, they concluded that the frag-
mentation of the collision complex proceeds promptly with
the lifetime less than 100 fs. The difference of the linewidth
between the new and old OH molecules shown by King et al.
is indeed very obvious. However, it has also been shown that
the internal energy content of the new OH is far larger than
that of the old OH.5,6 Therefore, the difference of the line-
width between the new and old OH molecules can arise from
this asymmetric energy partitioning between the two OH
products. Thus, the product LAB speed cannot be uniquely
related to the CM scattering angle. Consequently, the addi-
tional information such as the product LAB recoil anisotropy
is required for uniquely determining the product DCS with
the translational energy distribution.
1
16
16O D͒ϩH 18O→ OHϩ18OH ⌬H0 0͒ϭϪ118 kJ/mol
͑
͑
2
͑1͒
has been studied.1–11
Pioneering measurements using the H218O reagents in
early 1980s, revealed that the newly formed 16OH is vibra-
tionally and rotationally much more excited than the old
18OH.1–4 This unequal energy partitioning was explained by
a direct process such as abstraction or stripping.
About 10 years later, with the development of the detec-
tion technique of laser-induced fluorescence ͑LIF͒, the titled
reaction was reexamined under collision-free conditions5,6
which were not achieved in the earlier experiments.1–4 The
new experiments free from the rotational relaxation of the
nascent products showed remarkable rotational excitation of
the new OH compared with the old OH. In addition, the
⌳-doublet state distribution measured in these studies
In recent years, precise determination of the DCS, trans-
lational energy distribution, and polarization of rotational an-
gular momentum of molecular products in bimolecular reac-
tions have been performed by several groups.12–31 The
showed some preferential formation of the ⌸(A ) compo-
Ј
nent. Based on these observation, Cleveland et al. proposed a
direct insertion mechanism with a collisional HOOH com-
0021-9606/99/110(16)/7707/10/$15.00
7707
© 1999 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

7708
J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
theory to extract such information with the Doppler probing
technique has been well established by Aoiz et al.32,33 With
this sophisticated procedure, stereodynamics of several reac-
tions of fundamental importance have been analyzed in de-
tail so far.14–22 We have re-examined the titled reaction by
using this method. Although the isotopically labeled reaction
͑1͒ is essential in distinguishing the two types of OH prod-
ucts, we used the isotopically unlabeled normal H2O reagents
in this study. However, by probing the OH products gener-
ated in the specific vibrational level ϭ2, the newly formed
v
Ј
OH can be selectively observed since the previous works
reported that the amount of the old OH in
ϭ2 was
v
Ј
negligible.5,6 We have measured the state-selected DCS and
rotational angular momentum polarization of the newly
formed OH(2⌸3/2
,
ϭ2, j ϭ5.5) by utilizing polarized
v
Ј Ј
Doppler spectroscopy. The reaction dynamics is discussed
on the basis of thus obtained relatively isotropic DCS and the
angular correlation of j which is perpendicular to the colli-
Ј
sion plane spanned by k and k , the relative velocity vectors
Ј
of the reactants and products, respectively. The results sug-
gest an insertion mechanism involving a collisional HOOH
complex with the main contribution of H–O–O bending mo-
tion to the product OH rotation. Even the HOOH intermedi-
ate complex can survive for its rotational period, it breaks up
prior to the energy randomization, resulting in the asym-
metrical energy partitioning between the two OH fragments.
II. EXPERIMENT
FIG. 1. ͑a͒ Schematic diagram of the experimental apparatus used in the
measurement of Doppler profiles. M, mirror; P, polarizer; PEM, photoelastic
modulator; Li , synthetic silica lens; IF, interference filter; and PMT, pho-
tomultiplier tube. The probe laser beam was introduced into the vacuum
chamber either collinearly or orthogonally to the photolysis laser beam. ͑b͒
Characterization of the LIF excitation-detection geometries employed in the
measurement of Doppler profiles. The vectors of k and ⑀ represent the
propagation and electric vectors of the laser light or fluorescence, respec-
tively. The subscripts p, a, and d imply the photolysis, absorbed, and de-
tected photons, respectively.
The experimental apparatus is schematically depicted in
Fig. 1͑a͒. Experiments were carried out in a stainless steel
vacuum chamber. A mixture of N2O and H2O vapor was
introduced into the reaction chamber and the flow rate was
controlled with stainless steel needle valves. Nitrous oxide
͑Showa Denko, 99.999%͒ and distilled H2O were used with-
out further purification. The sample gases in the reaction
chamber were continuously pumped by using a rotary pump.
The partial pressures of N2O and H2O vapor were about 100
and 50 mTorr at 300 K, respectively. The electronically ex-
cited oxygen atom, O(1D) was generated from the photodis-
sociation of N2O at 193 nm with an ArF excimer laser
͑Lambda Physik EMG 53MSC͒. At a time delay of 120 ns
after the photolysis laser radiation, a tunable UV light ͑TMQ,
346–355 nm͒ of a dye laser ͑Lambda Physik SCANmate 2E͒
pumped by a XeCl excimer laser ͑Lambda Physik COMPex
102͒ was introduced into the reaction chamber to probe the
the photolysis and probe lasers. Such ⌬vϭϪ1 excitation
with ⌬vϭ0 detection was essential to improve the signal to
noise ratio. The LIF signal was amplified by a fast preamp-
lifier ͑Comlinear CLC 100͒ and gated by a boxcar integrator
͑Stanford Research System SR 250͒. The gated signal was
digitized by an A/D convertor ͑Stanford Research System
SR245͒ and stored in a personal computer via a GPIB inter-
face. The pump-probe experiments were performed at a rep-
etition rate of 10 Hz. The timing of the laser system was
adjusted by a digital pulse/delay generator ͑Stanford Re-
search System DG 535͒. Each laser power was monitored by
a photodiode ͑Hamamatsu 1336-5BQ͒ for normalizing the
LIF signal to correct for the fluctuation of the power.
The present experiments utilized the photolysis of N2O
at 193 nm for generating O(1D). Different from the O3 pho-
tolysis at 266 nm ͑Refs. 1–4, 6–9͒ or 248 nm ͑Ref. 5͒ so far
used, the photolysis at 193 nm could also photolyze H2O and
produce OH fragments.34 Although a very small amount of
contamination due to the OH photofragments was observed
nascent product OH( ϭ2) by laser-induced fluorescence
v
Ј
͑LIF͒ via the OH A 2⌺ϩ –X 2⌸ ͑1,2͒ excitation in the spec-
tral region of 348–352 nm. Measurements of Doppler pro-
files were performed by using the probe laser with an intra-
Ϫ1
´
cavity etalon which yields the bandwidth of about 0.04 cm
in full-width at half-maximum ͑FWHM͒.
The fluorescence was collected by a synthetic silica lens
and focused by another silica lens on a photomultiplier tube
͑Hamamatsu R928͒ mounted perpendicular to both of the
photolysis and probe laser beam axes. An interference filter
͑Nippon Sinku Kogaku 6205–5989͒ centered at 310 nm with
13.5 nm FWHM was placed in front of the photomultiplier
tube to transmit only OH A 2⌺ϩ –X 2⌸ ͑1,1͒ emission by
blocking the strong scattered light and emission arising from
in v ϭ0, we confirmed that such contamination was abso-
Ј
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
7709
III. ANALYSIS
A. Doppler-resolved line profile
lutely negligible in ϭ2 as consistent with the extremely
v
Ј
cold vibrational distribution of the OH photofragments
(Pv ϭ1 /Pv ϭ0Ͻ0.03) reported by Grunewald et al.34
Ј
Ј
In recent years, based on the Dixon’s formulation for the
Doppler analysis of the vector correlations in photodissocia-
tion processes,36 Aoiz et al. have developed the theoretical
basis to analyze the mutual angular correlations among the
LAB velocity vector of the atomic reagent ͑v͒, LAB velocity
The effect of collisional relaxation was also carefully
examined since it has been found that the OH(A 2⌺ϩ) and
OH(X 2⌸) are effectively deactivated by the collisions with
H2O.5,35 The Doppler line shapes did not show any appre-
ciable changes within experimental errors with varying the
total pressure of 2:1 N2O/H2O mixed sample gas from 75
mTorr to 150 mTorr at a 120 ns pump–probe delay. There-
fore, the deactivation of OH(A 2⌺ϩ) and OH(X 2⌸) is neg-
ligible under the experimental conditions employed in the
present study. Moreover, we ensured that the probe laser
power was weak enough to avoid the saturation effect in the
OH A 2⌺ϩ –X 2⌸ ͑1,2͒ transition.
vector (v ), and rotational angular momentum vector (j ) of
Ј
Ј
the product molecules in photo-initiated bimolecular reac-
tions from the polarized Doppler-resolved line profiles as
described below.32,33 In case of the linearly polarized pump-
probe experiment, the 1ϩ1 LIF Doppler profile D(␯) as a
function of the displaced frequency ␯ from the line center
frequency ␯ can be written by the Legendre polynomial
0
expansion of Pn(x) up to the second order as
The Doppler-resolved spectral line profiles were mea-
sured in the three different LIF excitation-detection geom-
etries I, II, and III shown in Fig. 1͑b͒ according to a combi-
nation in the directions of the polarization and propagation
vectors of each laser. Incomplete vertical polarization of the
output of the probe laser was converted to higher vertical
linear polarization ͑degree of polarization Ͼ0.99͒ with a
Glan laser prism ͑Meiritsu LAN-10-L͒. The direction of the
polarization vector of the probe laser light denoted as ⑀a in
Fig. 1͑b͒ was always vertical in all the excitation-detection
geometries. The originally unpolarized ArF laser light was
passed through a pile of plates consisting of twelve Brew-
ster’s angle quartz plates to obtain a vertically polarized light
͑degree of polarization Ͼ0.95͒. The firing of the ArF laser
was synchronized to the stress cycle of a photoelastic modu-
lator ͑Hinds PEM 90͒. By changing the arrival timing of the
polarized 193 nm light at the photoelastic modulator, one can
control the polarization of the light passing through this
modulator. The trigger pulses for the shot-by-shot alternation
of the two appropriate timings corresponding to the vertical
and horizontal polarization were generated with home-made
electronic devices. This shot-by-shot alternation of the polar-
ization vector allows us to observe the two polarization-
dependent spectral line profiles in the LIF geometries of Ib
and III in a single scan. In order to improve the signal to
noise ratio, 50–100 laser shots were averaged at each step of
a frequency scan and the spectra obtained from 9 to 25 scans
were summed up.
Polarization measurements were performed for the rota-
tional transitions of R1(5.5) and Q1(5.5). Unfortunately, the
signal intensities of their respective associated satellites were
too weak to separately probe the dynamics of the two
⌳-doublet states. We discuss the adequacy of the analysis
using only the main branches of different ⌳-doublet states in
Sec. IV. The rotational alignment A(02) used in the Doppler
profile analysis was determined by the ratio of the area of the
Doppler profile obtained in the geometry Ib to that obtained
in the geometry III at a single scan with the shot-by-shot
polarization alternation method described above. The aver-
age value over about 25 scans was 0.00Ϯ0.03 for both of
R1(5.5) and Q1(5.5).
ϱ
1
vp
D ␯͒ϭ
g
͒ϩg
͒P
2
d
,
v
Ј
͑
͑v
͑v
Ј
Ј
͵
ͫ
ͩ ͪ
ͬ
0
2
2
v ϭ ͑␯/␯ ͒c
͉ ͉
0
v
v
Ј
Ј
p
͑2͒
where pϭ
͉
(␯/␯0)c
͉ ͑c, the speed of light͒ represents the
v
minimum product LAB speed giving the displaced frequency
␯ as the Doppler shift according to the projection of v along
the propagation direction of the probe laser, ka. The coeffi-
cients g ( ) and g ( ) in Eq. ͑2͒ are expressed as
Ј
v
v
Ј
Ј
0
2
␤
photo
2
0
g0
͒ϭb ␤0 0,0; ͒ϩb
␤
0,2;
͒
͒
͑3͒
͑v
͑
v
͑
v
v
Ј
Ј
Ј
Ј
Ј
0
1
0
2
and
␤
photo
2
0
g2
͒ϭb
␤
2,0; ͒ϩb ␤0 2,2;
͑v
͑
v
͑
0
Ј
2
3
2
␤
photo
2
0
ϩb4
␤
2,2; ͒,
vЈ
͑4͒
͑
2
where ␤photo is the translational anisotropy parameter of the
atomic reagent produced from the photodissociation of a mo-
lecular precursor. The ␤K(k ,k ; ) represents the bipolar
moments averaged over all the possible velocities of the re-
agents and reflects the vector correlations as a function of
v
2
Ј
1
0
product LAB speed . The multipliers b of bipolar mo-
v
Ј
i
ments in g0 and g2 depend on the LIF excitation-detection
geometries shown in Fig. 1͑b͒ and the types of rotational
transitions. As shown by Docker,37 an appropriate linear
combination of the Doppler profiles measured in several LIF
geometries for two different types of rotational transitions ͑P
or R vs Q͒ gives the composite Doppler profile which de-
pends on a single bipolar moment ␤K(k ,k ; ) in g or g .
v
Ј
1
2
0
2
0
The composite Doppler profile DK0 (k1 ,k2 ;␯) reflecting each
bipolar moment ␤K(k ,k ; ) is expressed as33
v
Ј
1
2
0
DK0 k ,k ;␯͒
͑
1
2
ϱ
1
vp
K
0
ϭ
␤
k ,k ; ͒P
k
1
d .
v
Ј
͑5͒
͑
v
Ј
͵
ͩ ͪ
1
2
2
v ϭ ͑␯/␯ ͒ c
͉ ͉
0
v
v
Ј
Ј
p
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

7710
J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
The composite Doppler profiles D00(0,0;␯) and D02(2,0;␯)
corresponding to the leading terms in g0 and g2 are related to
DK0 k ,k;␯͒ϭ
c
f ,cos ␪ ͒
͑
kq T t
͑
͚ ͚ ͚
1
f
cos ␪
q
T
t
the LAB speed distribution ͑population͒ and v؊v correla-
Ј
ϫBK0 k ,k;␯;f ,cos ␪ ;q͒.
͑9͒
tion ͑anisotropy͒, respectively. The other three composite
Doppler profiles reflect the polarization of the product rota-
tional angular momentum vector, j ; D2(0,2;␯) for v؊j ;
͑
1
T
t
Ј
Ј
0
D0(2,2;␯) for v ؊j , and D2(2,2;␯) for v؊v ؊j correla-
Ј
Ј
Ј
Ј
0
0
The basis functions B0K(k1 ,k;␯;fT ,cos ␪t ;q) represent the
contribution of the q component of the PDDCSs to
DK(k1 ,k;␯) resolved into a specific CM scattering angle ␪
and the fraction of the available energy ͑excluding the inter-
nal energy of the observed product͒, released into the prod-
uct CM translational energy, fT . The set of coefficients
ckq(fT ,cos ␪t) determined by the fitting procedure represents
the PDDCSs for each fT . In the present analysis, the fT
region ranging from 0.0 to 1.0 was divided into 7 bins while
tions.
t
0
B. Extraction of the product state-selected vector
correlations in the CM frame
The vector correlations in the LAB frame represented by
DK0 (k1 ,k2 ;␯) need to be transformed into those in the
center-of-mass ͑CM͒ frame; the angular correlations among
the relative velocity vectors of the reactants (k) and products
(k ), and the product rotational angular momentum vector
Ј
the region of cos ␪ ranging from Ϫ1.0 to 1.0 was divided
t
(j ). The angular distribution function P(␻ ,␻ ) represent-
Ј
t
r
into 10 bins, i.e., 70(ϭ7ϫ10) basis functions were used for
this fitting.
ing the mutual orientation among the three vectors k, k , and
Ј
j is expressed as the expansion in modified spherical har-
Ј
For the preparation of the basis functions, the knowledge
of the recoil velocity distribution and translational anisotropy
of O(1D) produced from the 193 nm photodissociation of
N2O is required. Felder et al. using an angle-resolved time-
of-flight mass spectroscopy method reported the recoil
speed-angular distribution of O(1D) generated from the pho-
todissociation of N2O at 193 nm.40 In order to consider all
the possible collisions, the velocity distribution of O(1D)
was convoluted with isotropic thermal motions of the precur-
sor N2O and the reactant H2O, and the angle between the
velocity vectors of the reactants O(1D) and H2O. Isotropic
thermal motions of N2O and H2O molecules were assumed
to be represented by the Maxwell–Boltzmann distributions at
300 K. The angular dependence of the signal integrated over
the recoil speed distribution was described by a single an-
isotropy parameter of ␤photoϭ0.48.40 However, Suzuki et al.
recently showed a recoil speed dependence of ␤photo for
O(1D) produced from the photodissociation of N2O at 205
nm by using a two-dimensional imaging technique.41 They
described the observed recoil energy and angular distribution
as the superposition of the two Gaussian distributions; a
broad one with ␤photoϭ0.0 and a relatively narrow one with
monics Ckq(␪r ,␾r) ͑Ref. 38͒ with a set of coefficients
(1/␴)d␴kq /d␻ which is defined as polarization-dependent
t
differential cross sections ͑PDDCSs͒,39
2kϩ1 1 d␴
kq
*
␪ ,␾ ͒ ,
͑
r r
P ␻ ,␻ ͒ϭ
C
͑6͒
͑
͚
t
r
kq
4␲ ␴ d␻
k,q
t
where ␻tϵ(␪t ,␾t) and ␻rϵ(␪r ,␾r) are the polar and azi-
muthal angles representing the directions of k and j , re-
Ј
Ј
spectively, with respect to k. Furthermore, the angular distri-
bution function P(␪r ,␾r) representing the product rotational
angular momentum polarization with reference to k and kk
Ј
plane is obtained by integrating P(␻t ,␻r) over the whole
scattering angles
P ␪ ,␾ ͒ϭ P ␻ ,␻ ͒d␻
͑
͑
͵
r
r
t
r
t
2kϩ1
4␲
*
r
ϭ
akqCkq ␪ ,␾ ͒ ,
͑7͒
͑8͒
͑
͚
r
k,q
where
␤
photoϭ1.3. In the present study, as a reasonable assumption,
1 d␴
a ϭ
kq
kq d␻t .
͵
the observation at 205 nm by Suzuki et al. is adapted for the
193 nm result. We fit the translational energy distribution
measured by Felder et al. with two components of broad and
narrow Gaussian type distributions. The average ␤photo
ϭ0.48 was reproduced by setting ␤photoϭ0.0 for the broad
component and ␤photoϭ1.03 for the narrow component. In
practice, from the comparison between the single and double
␴ d␻
t
Aoiz et al. have shown the relationship between the bipolar
moments ␤K(k ,k ; ) defined in the LAB frame and the
v
2
Ј
1
0
PDDCSs (1/␴)d␴kq /d␻ defined in the CM frame.33 Hence,
t
the PDDCSs are determined from the observed composite
Doppler profiles DK0 (k1 ,k2 ;␯) In addition to the recoil speed
and scattering angle distributions in the CM frame, the an-
gular distribution function P(␪r ,␾r) is then calculated from
Eq. ͑7͒.
␤
photo distributions, it was found that the details of the recoil
speed dependence of the translational anisotropy of the
atomic reagent does not affect the product DCS and fT dis-
tribution significantly.
The fitting procedure with nonorthogonal basis functions
as attempted in this study is one of the notorious problems
giving an unphysical solution in numerical analysis. In such
a situation, a least-squares fitting typically gives an over-
structured DCS as the best solution. To obtain a physically
C. The fitting procedure
The set of PDDCSs is determined self-consistently from
the least-squares fitting. The experimental composite Dop-
pler profiles DK0 (k1 ,k;␯) are fit with the linear combination
of the basis functions BK0 (k1 ,k;␯;fT ,cos ␪t ;q),
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
7711
reasonable DCS with a smooth shape, we used singular value
decomposition fitting42 giving a set of orthogonal basis func-
tions from a set of original nonorthogonal functions. Then,
the angular resolution of the DCS obtained with this algo-
rithm becomes not so sensitive to the number of the basis
functions. Furthermore, a maximum-entropy method42 was
utilized for eliminating physically meaningless solutions
such as negative values of c00(fT ,cos ␪t) representing the
product fT-resolved DCS which should not be negative. The
analysis was carried out with the assumption that the reaction
probability and the set of PDDCSs do not depend signifi-
cantly on the initial translational energy of the reactants ͑the
collision energy͒.
IV. RESULTS AND DISSCUSSION
A. Scalar properties
As described in Sec. I, even with the isotopically unla-
beled reagents used in this study, the observed products
OH( ϭ2) can be exclusively assigned to the new OH since
v
Ј
the population of the old 18OH in ϭ2 was far below the
v
Ј
detection level in isotopically labeled studies.5,6 The vibra-
tional state distribution of the old 18OH was reported to be
94% in ϭ0 and 6% in ϭ1.6 Concerning the new OH,
v
v
Ј
Ј
the vibrational state distribution was reported to be 39: 29:
FIG. 2. Observed composite Doppler profiles for the product
OH(2⌸3/2 ϭ2,j ϭ5.5) ͑a͒ D0(0,0;␯) ͑᭺͒ and ͑b͒ D2(2,0;␯) ͑᭹͒. The
32 for ϭ0;1;у2. These figures indicate that the
v
Ј
,
vЈ
smooth solid lines are the best fits in the Doppler analysis.
Ј
0
0
(
, old)ϭ(2,0) channel is dominant compared with the
vnew
͑2, 1͒ channel. Even with the extreme assumption that the
new OHs in newϭ0 and 1 levels are exclusively paired with
v
v
ϭ0 level, the 32% occurrence of the new OH( ϭ2)
would be distributed into ͑2, 0͒ channel with 26% and ͑2, 1͒
channel with 6%. Therefore, at least more than 80% of the
vold
v
Ј
rence of the (
ϭ2, oldϭ1) channel as discussed at the
vnew
v
beginning of this section is displayed in Fig. 3͑b͒. Tanaka
et al. reported the internal state distributions of the new and
old OH products generated from the isotopically labeled re-
action ͑1͒ with the same O(1D) source as in the present
study.10,11 The Boltzmann rotational state distribution with
Trotϭ2600 K for the old 18OH( ϭ0) determined by Tanaka
et al. is also displayed in Fig. 3͑b͒. The rotational state dis-
tribution of the old OH in the ͑2, 0͒ channel obtained in the
present study is, though slightly cold, quite close to the result
reported in Ref. 10.
new OH( ϭ2) is paired with the old OH( ϭ0). This fact
v
v
Ј
Ј
means that the translational energy distribution of the new
OH( ϭ2) in a specific rotational level almost corresponds
v
Ј
to the rotational state distribution of the old OH ͑the coprod-
v
Ј
uct͒ in ϭ0.
v
Ј
Figure 2 shows the observed composite Doppler profiles
for the OH(2⌸3/2 ϭ2,j ϭ5.5) generated in the O(1D)
,
v
Ј
Ј
ϩH2O reaction, D00(0,0;␯) and D20(2,0;␯), reflecting the
LAB speed distribution and recoil anisotropy, respectively.
The smooth solid lines in Fig. 2 represent the least-squares
fitting for the experimental data. The fT distribution deter-
mined by the least-squares fitting is displayed in Fig. 3͑a͒.
From the fT distribution, it is estimated that about 60% of the
available energy excluding the internal energy of the ob-
Tanaka et al. measured the overall rotational state distri-
bution of the old 18OH( ϭ0) whose sister fragments of the
v
Ј
new OH could span all the energetically accessible vibra-
tional levels, ϭ0–3. Although we specified the rovibronic
v
Ј
level ϭ2 and j ϭ5.5 for the new OH, the rotational dis-
v
Ј
Ј
tributions of the old OH( ϭ0) deduced in both cases are
v
Ј
served new OH( ϭ2,j ϭ5.5) are channeled into the prod-
v
Ј
Ј
quite similar. That is, the rotational state distribution of the
old OH depends very weakly on the rovibrational state of the
new OH in this reaction. This trend is partly due to the small-
ness of the energy distributed to the old OH; the average
rotational energy of the old 18OH( ϭ0) of 22 kJ/mol cor-
uct CM translational energy and the residual 40% are chan-
neled into the internal energy of the coproduct old OH. Such
dominance of the energy partitioning into translational de-
grees of freedom can be recognized from the comparison
between the observed fT distribution and statistical one cal-
culated by only considering the energy conservation as
shown in Fig. 3͑a͒.
v
Ј
responds to only 14% of the total available energy.
B. Vector properties
The internal energy distribution of the old OH paired
1. Differential cross section
with the observed new OH( ϭ2) could be estimated by
v
Ј
convoluting the (1ϪfT) distribution with the collision en-
ergy distribution. The approximate rotational state distribu-
The product total DCS averaged over the whole fT re-
gion determined by the least-squares fitting procedure is dis-
played in Fig. 4. The DCS has substantial intensities over a
tion of the old OH( ϭ0) obtained by neglecting the occur-
v
Ј
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

7712
J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
pathways. However, this cannot be the case in the present
reaction based on the following consideration. In extracting
the DCS, we can obtain information about the dependence of
the DCS on the product recoil energy. The evaluated recoil
energy distributions of the products scattered in the forward
and backward hemispheres did not show any significant dif-
ference between them. If the direct stripping and abstraction
pathways actually occur, the recoil energy distributions for
the forward scattering corresponding to the stripping path-
way should be different from that for the backward scattering
corresponding to the rebound pathway. Thus, the absence of
significant dependence of the recoil energy distribution on
the scattering angles suggests the insertion pathway involv-
ing a complex which survives for a rotational period. Fur-
thermore, as will be described in the following subsection,
the observed angular momentum polarization did not also
largely depend on the scattering angles. This result also sup-
ports the insertion pathway rather than a combination of
stripping and abstraction pathways since the dynamical con-
straints arising from an abstraction pathway is expected to be
weaker than an insertion pathway and leads to a weaker an-
gular momentum polarization for the backward component.
The nearly isotropic DCS obtained in this work seems to
disagree with the anisotropic CM scattering angle distribu-
tion emphasized in the previous Doppler study performed by
King et al.7 However, they measured only the composite
Doppler profile corresponding to D00(0,0;␯) and analyzed the
data simply with a single scattering angle without allowing
for the spread of angular distribution in the scattering. They
obtained the representative scattering angles of 60°–86° for
the new OH products in various rovibrational levels. These
values do not largely deviate from 90° that corresponds to
the average scattering angle in the completely isotropic scat-
tering. Therefore, such representative values of scattering
angles previously reported by King et al. can be considered
to suggest a nearly isotropic scattering angle distribution
with a slight preference in the forward hemisphere rather
than an anisotropic scattering. In this sense, these two results
do not seriously disagree with each other.
The nearly isotropic DCS described above implies that
the reaction proceeds via a long-lived intermediate complex
which is generally expected to yield statistical internal en-
ergy distributions. However, it is reported in many
researches1–11 that this reaction yields extremely unequal in-
ternal energy partitioning between the two OH products.
These conflicting findings suggest that the redistribution of
the available energy among the internal modes of the inter-
mediate complex does not occur on a time scale comparable
to the rotational period of the complex. Similar nonstatistical
energy distributions for the complex-mode reactions involv-
ing long-lived complexes have been found in the reactions
O(1D)ϩCH4→OHϩCH3 ͑Refs. 43, 44͒ and H(2S)ϩCO2
→OHϩCO.45–49 The real time pump–probe experiments by
utilizing ultrafast lasers were carried out for these reactions
and the results indicated that each reaction involved a long-
lived complex with lifetime of picosecond order ͓3 ps for the
O(1D)ϩCH4 reaction and 0.5–5 ps for the H(2S)ϩCO2 re-
action͔. However, as opposed to a traditional expectation for
complex-mode reactions, the energy distributions for these
FIG. 3. ͑a͒ fT distribution of the product OH(2⌸3/2
,
ϭ2,j ϭ5.5) ͑᭹͒. A
Ј
vЈ
prior distribution ͑᭿͒. ͑b͒ Rotational state distribution of the old OH(
vЈ
ϭ0) paired with the observed new OH(2⌸3/2
,
ϭ2,j ϭ5.5) ͑᭹͒. The
vЈ
Ј
solid line is the overall rotational state distribution of the 18OH( ϭ0)
vЈ
generated from the isotopically labeled reaction of the 16O(1D) with H2 18O
͑Ref. 10͒. The error bars correspond to Ϯ1␴.
wide angular range with slightly preferred backward scatter-
ing. This result cannot be predicted either from the direct
abstraction mechanism yielding backward peaking or the di-
rect stripping mechanism yielding forward peaking. The
nearly isotropic DCS most probably indicates that the reac-
tion proceeds via an insertion pathway involving the forma-
tion of an intermediate HOOH complex which survives for a
time comparable to or longer than its rotational period.
Another possible explanation for the observed nearly
isotropic DCS is a combination of rebound and stripping
FIG. 4. Total differential cross sections averaged over the whole fT region
for the product OH(2⌸3/2
Ϯ1␴.
,
ϭ2,j ϭ5.5). The error bars correspond to
vЈ
Ј
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
7713
FIG. 6. The set of polarization-dependent differential cross sections for the
product ϭ2,j ϭ5.5). ͑a͒ (1/␴)d␴20 /d␻t , ͑b͒
OH(2⌸3/2
,
vЈ
Ј
(1/␴)d␴21Ϫ /d␻t , and ͑c͒ (1/␴)d␴22ϩ /d␻t . The error bars correspond to
Ϯ1␴.
FIG. 5. Observed composite Doppler profiles of the product OH(2⌸3/2
,
Ј
tional angular momentum vector, j , averaged over the
whole range of scattering angle derived by using Eq. ͑7͒ is
vЈ
ϭ2,j ϭ5.5). ͑᭹͒ ͑a͒ D2(0,2;␯), ͑b͒ D0(2,2;␯), and ͑c͒ D2(2,2;␯). The
Ј
0
0
0
smooth solid lines are the best fits in the Doppler analysis.
shown in Fig. 7. As noted in Sec. III, ␪ is the angle between
r
k and j , while ␾ is the dihedral angle between the kk and
Ј
Ј
r
kj planes. Thus, the two peaks near (␪ ,␾ )ϭ(90°,90°)
Ј
r
r
reactions were found to apparently deviate from the statisti-
cal predictions in both reactions. These results imply that the
absolute value of lifetime of the collisional complex is not
the unique essential factor to determine the degree of energy
redistribution. The rotational period of the collisional HOOH
complex can be roughly estimated to be about 0.5 ps by
using the rotational constants of the H2O2 molecule50 and the
initial orbital angular momentum ͑Ϸ40ប͒ of the reactants.
Thus, a nearly isotropic DCS obtained in this work suggest
that the lifetime of a collisional HOOH complex is larger
than 0.5 ps. Since this lifetime is shorter than those of the
other two reactions described above, the nonstatistical en-
ergy distributions1–11 for the reaction of O(1D)ϩH2O
→OHϩOH is not unexpected.
and ͑90°, 270°͒ imply that j is predominantly polarized per-
Ј
pendicular to the kk plane ͑i.e., collision plane͒. This pref-
Ј
erence for j perpendicular to the collision plane can be ex-
Ј
plained from a primary contribution of the bending motion of
the H–O–O moiety in the HOOH collisional complex to the
product rotation. This tendency for the product rotational an-
gular momentum polarization has been recently found in the
2. Rotational angular momentum polarization
Figure 5 shows the set of composite Doppler profiles
D20(0,2;␯), D00(2,2;␯), and D20(2,2;␯), reflecting the rota-
tional angular momentum polarization for the product
OH(2⌸3/2
,
ϭ2,j ϭ5.5). The smooth solid lines in Fig. 5
v
Ј Ј
represent the least-squares fitting for the composite Doppler
profiles. The fitting gives the polarization-dependent differ-
ential cross sections ͑PDDCSs͒ shown in Fig. 6. Further, the
angular distribution function P(␪r ,␾r) of the product rota-
FIG. 7. Polar plots of the P(␪r ,␾r) representing the angular distribution of
the rotational angular momentum vector of the product OH(2⌸3/2
,
vЈ
ϭ2,j ϭ5.5).
Ј
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

7714
J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
analogous reaction of O(1D) with H2,18 which has been con-
sidered to proceed via an insertion pathway involving the
formation of highly vibrationally excited H2O.
The angular correlation of j perpendicular to the kk
Ј
Ј
plane derived in this bimolecular reaction makes a clear con-
trast to the parallel v –j correlation reported in the photo-
Ј
Ј
dissociation studies of HOOH molecules.51–53 The equilib-
rium position of the torsional angle is the main difference in
the potential energy surfaces of HOOH molecules between
the electronically ground and excited states. Hence the
Franck–Condon excitation from the ground state to the ex-
cited states is followed by the torsional motion giving a par-
allel v –j correlation to the OH photofragments. In contrast
Ј
Ј
to such a specific condition in the HOOH photodissociation
process, the insertion of the O(1D) atom into the OH bond of
the H2O reagent in the bimolecular reaction will give rise to
the significant excitation of the stretching and bending mo-
tions of the newly formed H–O–O moiety in addition to the
excitation of the torsional motion. Furthermore, the torsional
angle dependence of the potential energy in the ground
state54 is much weaker than those in the excited states.55
Therefore, the clear difference in the angular correlation be-
tween the bimolecular reaction and photodissociation of the
HOOH system would be reasonably understood.
FIG. 8. Composite Doppler profiles corresponding to D20(0,2;␯)
ϩ(c4 /c1)D20(2,2;␯) for the rotational transitions of R1(5.5) and Q1(5.5).
The coefficients ci are defined in the text.
equal rotational angular momentum polarization for the two
⌳-doublet states, adopted in this study, is confirmed.
C. Kinematic consideration
Combining the scalar and vector properties obtained in
this study, we discuss the correlation between the two rota-
tional angular momentum vectors of the OH products in
terms of kinematic consideration. The conservation of the
total angular momentum J in this reaction will be satisfied as
follows:
As recognized from the relatively flat behavior of each
PDDCS displayed in Fig. 6, the OH products scattered in the
forward and backward hemispheres have almost the same
rotational angular momentum polarization; j is perpendicu-
Ј
lar to the collision plane. It suggests that irrespective of the
product scattering directions, the reaction is dominated by
the same mechanism, i.e., an insertion process involving a
HOOH complex with a lifetime comparable to the rotational
period.
JϭLϩJOϩJH OϭL ϩjЈϩjЈ ,
͑10͒
Ј
1
2
2
where JO is the electronic orbital angular momentum of the
atomic reagent O(1D) and JH O and jiЈ (iϭ1,2) are the rota-
2
As briefly mentioned in the Sec. II, in the present study,
the product rotational angular momentum polarization for the
tional angular momentum vectors of the reactant H2O and
the two OH products, respectively. The orbital angular mo-
two ⌳-doublet states (A and A ) was not measured sepa-
Ј
Љ
mentum vectors, L and L are defined for the reactants and
Ј
rately and the analysis was performed by assuming the same
angular momentum polarization for these ⌳-doublet states.
Since the recent studies have shown the different rotational
angular momentum polarization for the ⌳-doublet states in
several bimolecular reactions,14,18,20,23 we will demonstrate
the validity of such an assumption by utilizing the difference
Doppler profile which is only dependent on the rotational
angular momentum polarization. The difference between the
profiles measured in the LIF excitation-detection geometries
III and Ia corresponds to D20(0,2;␯)ϩ(c4 /c1)D20(2,2;␯),
where ci is the difference between the multipliers bi for these
two geometries and the ratio c4 /c1 is independent of the
probed rotational branches. A close similarity between such
products, respectively. The value of JO equals 2ប and the
average value of JH O is estimated to be 4ប at 300 K. The
2
average value of L is estimated to be about 40ប for the av-
erage impact parameter of 1.0 Å expected from the reaction
rate constant56 (kϭ2.0ϫ10Ϫ10 cm3 moleculeϪ1 sϪ1). As
shown in Fig. 9, the average value of the product CM trans-
lational energy ͑Ϸ45 kJ/mol͒ is close to that of the initial
reagent CM translational energy, i.e., the collision energy
͑Ϸ39 kJ/mol͒. Thus, assuming that the impact parameter of
the products is almost equal to that of the reactants, the av-
erage value of L is estimated to be about 43ប. While the
Ј
rotational angular momentum of the observed new OH is
j1Јϭ5.5ប, the average rotational angular momentum of the
coproduct old OH is estimated to be about jЈ2ϭ10.5ប from
the rotational state distribution shown in Fig. 3͑b͒.
The total angular momentum vector J can be considered
to be close to the initial orbital angular momentum vector L
profiles obtained for the rotational transitions R (5.5) for A
Ј
1
and Q (5.5) for A displayed in Fig. 8 indicates the small
Љ
1
difference in the product rotational angular momentum po-
larization between the ⌳-doublet states in this reaction. To
the contrary, D20(0,2;␯)ϩ(c4 /c1)D20(2,2;␯) will be appar-
since the magnitude of JOϩJH O estimated to be 2–6ប is
ently different between the A and A states in the reactions
Ј
Љ
2
showing the different angular momentum polarization for the
⌳-doublet states, judging from the composite Doppler pro-
files D20(0,2;␯) and D20(2,2;␯) displayed in the studies for
these reactions.14,18,20 Therefore, the approximation of an
much smaller than that of L ͑Ϸ40ប͒. Consequently, the an-
gular momentum conservation could be approximated as L
ϭL ϩj
ЈϩjЈ . As described in the preceding subsection, the
Ј
1
2
preference of spatial correlation of jЈ1 perpendicular to the
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
7715
10 N. Tanaka, M. Takayanagi, and I. Hanazaki, Chem. Phys. Lett. 254, 40
͑1996͒.
11 N. Tanaka, U. Nagashima, M. Takayanagi, H. L. Kim, and I. Hanazaki, J.
Phys. Chem. A 101, 507 ͑1997͒.
12 M. Brouard, S. P. Duxon, P. A. Enriquez, R. Sayos, and J. P. Simons, J.
Phys. Chem. 95, 8169 ͑1991͒.
13 M. Brouard, S. P. Duxon, P. A. Enriquez, and J. P. Simons, J. Chem.
Phys. 97, 7414 ͑1992͒.
14 M. Brouard, H. M. Lambert, J. Short, and J. P. Simons, J. Phys. Chem. 99,
13571 ͑1995͒.
15 M. Brouard, H. M. Lambert, C. L. Russell, J. Short, and J. P. Simons,
Faraday Discuss. 102, 179 ͑1995͒.
16 A. J. Alexander, M. Brouard, S. P. Rayner, and J. P. Simons, Chem. Phys.
207, 215 ͑1996͒.
17 A. J. Alexander, F. J. Aoiz, M. Brouard, I. Burak, Y. Fujimura, J. Short,
and J. P. Simons, Chem. Phys. Lett. 262, 589 ͑1996͒.
18
˜
FIG. 9. Product CM translational energy (ET) distribution. ͑᭹͒ The solid
line is the initial translational energy ͑the collision energy͒ distribution. The
error bars represent Ϯ1␴.
A. J. Alexander, F. J. Aoiz, L. Banares, M. Brouard, J. Short, and J. P.
Simons, J. Phys. Chem. A 101, 7544 ͑1997͒.
19 A. J. Alexander, D. A. Blunt, M. Brouard, J. P. Simons, F. J. Aoiz, L.
˜
Banares, Y. Fujimura, and M. Tsubouchi, Faraday Discuss. 108, 375
͑1997͒.
20 M. Brouard, H. M. Lambert, S. P. Rayner, and J. P. Simons, Mol. Phys.
89, 403 ͑1996͒.
kk plane was observed in the whole range of the angle
Ј
21 M. Brouard, I. Burak, G. A. J. Markillie, K. McGrath, and C. Vallance,
Chem. Phys. Lett. 281, 97 ͑1997͒.
between k and k . To satisfy the conservation law and the
Ј
preferential spatial correlation of j Ќkk plane over the
Ј
Ј
22 M. Brouard, I. Burak, S. D. Gatenby, and G. A. J. Markillie, Chem. Phys.
Lett. 287, 682 ͑1998͒.
whole scattering angles at the same time, the angle between
j1Ј and j2Ј is required to be larger than at least 90° from a
simple consideration of the correlations among the set of
vectors. Such a weak correlation arises from the main con-
tribution of concerted HOOH bending motion to the torque
force which rotates the product OH molecules.
23 H. L. Kim, M. A. Wickramaaratchi, X. Zheng, and G. E. Hall, J. Chem.
Phys. 101, 2033 ͑1994͒.
24 R. Fei, X. S. Zheng, and G. E. Hall, J. Phys. Chem. A 101, 2541 ͑1997͒.
25 W. R. Simpson, A. J. Orr-Ewing, T. P. Rakitzis, S. A. Kandel, and R. N.
Zare, J. Chem. Phys. 103, 7299 ͑1995͒.
26 W. R. Simpson, T. P. Rakitzis, S. A. Kandel, A. J. Orr-Ewing, and R. N.
Zare, J. Chem. Phys. 103, 7313 ͑1995͒.
27 A. J. Orr-Ewing, W. R. Simpson, T. P. Rakitzis, S. A. Kandel, and R. N.
Zare, J. Chem. Phys. 106, 5961 ͑1997͒.
V. SUMMARY
28 T. P. Rakitzis, S. A. Kandel, and R. N. Zare, J. Chem. Phys. 107, 9382
͑1997͒.
The stereodynamics of the vibrational channel
generating the OH( ϭ2) in the reaction O(1D)
v
Ј
29 T. P. Rakitzis, S. A. Kandel, T. Lev-On, and R. N. Zare, J. Chem. Phys.
107, 9392 ͑1997͒.
ϩH2O→OHϩOH has been investigated by observing the
product state-selected differential cross section and rotational
angular momentum polarization. We have observed a nearly
isotropic scattering angle distribution and the preference for
the perpendicular direction of the rotational angular momen-
tum vector to the collision plane. These results indicate the
formation of a HOOH collisional complex with a lifetime
comparable to the rotational period and a primary contribu-
tion of the bending motion of the HOOH complex to the
product rotation. The extremely asymmetrical energy parti-
tioning between the two OH fragments, therefore, suggests
that the redistribution of the available energy does not occur
on a time scale comparable to the rotational period of the
complex.
30 F. Green, G. Hancock, and A. J. Orr-Ewing, Faraday Discuss. Chem. Soc.
91, 79 ͑1991͒.
31 M. L. Costen, G. Hancock, A. J. Orr-Ewing, and D. Summerfield, J.
Chem. Phys. 100, 2754 ͑1994͒.
32 F. J. Aoiz, M. Brouard, P. A. Enriquez, and R. Sayos, J. Chem. Soc.,
Faraday Trans. 89, 1427 ͑1993͒.
33 F. J. Aoiz, M. Brouard, and P. A. Enriquez, J. Chem. Phys. 105, 4964
͑1996͒.
34 A. U. Grunewald, K.-H. Gericke, and F. J. Comes, Chem. Phys. Lett. 133,
501 ͑1987͒.
35 C. B. Cleveland and J. R. Wiesenfeld, Chem. Phys. Lett. 144, 479 ͑1988͒.
36 R. N. Dixon, J. Chem. Phys. 85, 1866 ͑1986͒.
37 M. P. Docker, Chem. Phys. 135, 405 ͑1989͒.
38 R. N. Zare, Angular Momentum ͑Wiley–Interscience, New York, 1988͒.
39 N. E. Shafer-Ray, A. J. Orr-Ewing, and R. N. Zare, J. Phys. Chem. 99,
7591 ͑1995͒.
40 P. Felder, B.-M. Haas, and J. R. Huber, Chem. Phys. Lett. 186, 177
͑1991͒.
1 J. E. Butler, L. D. Talley, G. K. Smith, and M. C. Lin, J. Chem. Phys. 74,
41 T. Suzuki, H. Katayanagi, Y. Mo, and K. Tonokura, Chem. Phys. Lett.
256, 90 ͑1996͒.
4501 ͑1981͒.
2 K.-H. Gericke, F. J. Comes, and R. D. Levine, J. Chem. Phys. 74, 6106
42 W. H. Press, B. P. Flannery, S. A. Teukolsky, and W. T. Vetterling,
Numerical Recipes in C, 2nd ed. ͑Cambridge University Press, Cam-
bridge, 1992͒.
͑1981͒.
3 F. J. Comes, K.-H. Gericke, and J. Manz, J. Chem. Phys. 75, 2853 ͑1981͒.
4 W. A. Guillory, K.-H. Gericke, and F. J. Comes, J. Chem. Phys. 78, 5993
43 C. R. Park and J. R. Wiesenfeld, J. Chem. Phys. 95, 8166 ͑1991͒.
44 R. D. van Zee and J. C. Stephenson, J. Chem. Phys. 102, 6946 ͑1995͒.
45 K. Kleinermanns, E. Linnebach, and Wolfrum, J. Phys. Chem. 89, 2525
͑1985͒.
͑1983͒.
5 C. B. Cleveland and J. R. Wiesenfeld, J. Chem. Phys. 96, 248 ͑1992͒.
6 D. G. Sauder, J. C. Stephenson, D. S. King, and M. P. Casassa, J. Chem.
Phys. 97, 952 ͑1992͒.
46 G. Hoffmann, Y. Chen, H. Iams, and D. Oh, J. Chem. Soc., Faraday
Trans. 2 85, 1292 ͑1989͒.
7 D. S. King, D. G. Sauder, and M. P. Casassa, J. Chem. Phys. 97, 5919
͑1992͒.
47 N. F. Scherer, L. R. Khundkar, R. B. Bernstein, and A. H. Zewail, J.
Chem. Phys. 87, 1451 ͑1987͒.
8 M. P. Casassa, D. G. Sauder, and D. S. King, Proc. SPIE 1858, 256
͑1993͒.
48 N. F. Scherer, C. Sipes, R. B. Bernstein, and A. H. Zewail, J. Chem. Phys.
92, 5239 ͑1990͒.
9 D. S. King, D. G. Sauder, and M. P. Casassa, J. Chem. Phys. 100, 4200
͑1994͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45

7716
J. Chem. Phys., Vol. 110, No. 16, 22 April 1999
Tsurumaki, Fujimura, and Kajimoto
49 S. I. Ionov, G. A. Brucker, C. Jaques, L. Valachovic, and C. Wittig, J.
Chem. Phys. 99, 6553 ͑1993͒.
53 M. P. Docker, A. Hodgson, and J. P. Simons, Chem. Phys. Lett. 128, 264
͑1986͒.
50 W. C. Oelfke and W. Gordy, J. Chem. Phys. 51, 5336 ͑1969͒.
51 K.-H. Gericke, S. Klee, F. J. Comes, and R. N. Dixon, J. Chem. Phys. 85,
4463 ͑1986͒.
54 R. H. Hunt, R. A. Leacock, C. W. Peters, and K. T. Hecht, J. Chem. Phys.
42, 1931 ͑1965͒.
55 R. Schinke, Photodissociation Dynamics ͑Cambridge University Press,
Cambridge, 1993͒, p. 237.
52 A. U. Grunewald, K.-H. Gericke, and F. J. Comes, J. Chem. Phys. 89, 345
͑1988͒.
56 K.-H. Gericke and F. J. Comes, Chem. Phys. Lett. 81, 218 ͑1981͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.91.169.193 On: Sun, 23 Nov 2014 19:19:45