Detection of O (3PJ) atoms formed by reaction, Al+ O2 → AlO+O under crossed-beam condition

Article abstract of DOI:10.1063/1.1924387

The vacuum ultraviolet laser-induced fluorescence technique was employed to detect the oxygen atoms formed by the reaction, Al+ O2 → AlO+O. The measurements were carried out under the crossed-beam condition at 12.2 kJmol of collision energy. The relative

Full text of DOI:10.1063/1.1924387

Detection of O ( P J 3 ) atoms formed by reaction, Al + O 2 → Al O + O under crossed-
beam condition
Masayuki Ishida, Tomohiko Higashiyama, Yoshiteru Matsumoto, and Kenji Honma
Citation: The Journal of Chemical Physics 122, 204312 (2005); doi: 10.1063/1.1924387
View online: http://dx.doi.org/10.1063/1.1924387
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/122/20?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Reaction dynamics of Al + O2 → AlO + O studied by a crossed-beam velocity map imaging technique: Vib-
rotational state selected angular-kinetic energy distribution
J. Chem. Phys. 140, 214304 (2014); 10.1063/1.4879616
Crossed-beam radical-radical reaction dynamics of O ( P 3 ) + C 3 H 3 → H ( S 2 ) + C 3 H 2 O
J. Chem. Phys. 124, 204320 (2006); 10.1063/1.2201996
Product spin–orbit state resolved dynamics of the H + H 2 O and H + D 2 O abstraction reactions
J. Chem. Phys. 121, 10426 (2004); 10.1063/1.1809578
Combined crossed-beam studies of C ( 3 P J )+ C 2 H 4 → C 3 H 3 + H reaction dynamics between 0.49 and
30.8 kJmol−1
J. Chem. Phys. 119, 10607 (2003); 10.1063/1.1619374
Reaction dynamics of Al + O 2 → AlO + O studied by the crossed-beam laser-induced fluorescence technique
J. Chem. Phys. 119, 3641 (2003); 10.1063/1.1591177
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:
193.61.135.80 On: Mon, 15 Dec 2014 17:56:07

THE JOURNAL OF CHEMICAL PHYSICS 122, 204312 ͑2005͒
3
Detection of O„ PJ… atoms formed by reaction, Al+O2\ AlO+O under
crossed-beam condition
Masayuki Ishida, Tomohiko Higashiyama, Yoshiteru Matsumoto, and Kenji Honmaa͒
Department of Material Science, University of Hyogo, 3-2-1 Kohto, Kamigori, Hyogo 678-1297, Japan
͑Received 9 February 2005; accepted 29 March 2005; published online 24 May 2005͒
The vacuum ultraviolet laser-induced fluorescence technique was employed to detect the oxygen
atoms formed by the reaction, Al+O2→ AlO+O. The measurements were carried out under the
crossed-beam condition at 12.2 kJ/mol of collision energy. The relative populations of three
spin-orbit states of O͑3PJ͒ were determined to be 3.8, 1.0, and 0.2 for J=2, 1, and 0, respectively.
They show nonstatistical populations, i.e., more population in O͑3P2͒ and less population in O͑3P0͒
than the statistical expectation. These populations were almost identical for two Al beam conditions
where the relative concentrations of two spin-orbit states of Al, 2P1/2, and 2P3/2, are different. These
results suggest that the reaction of Al with O2 proceeds via an intermediate complex where the
memory of the initial spin-orbit state is lost. Deviation from the statistical population of O͑3PJ͒
implies the occurrence of the interaction among potential surfaces in the exit channel. © 2005
American Institute of Physics. ͓DOI: 10.1063/1.1924387͔
I. INTRODUCTION
fluorescence technique and determined the rotational and vi-
brational distributions of AlO for each spin-orbit state.15
While our results again confirmed the lower reactivity of the
excited state, Al͑2P3/2͒, the rotational and vibrational distri-
butions of AlO corresponding to two spin-orbit levels were
quite similar to each other. These results, different reactivi-
ties and similar energy partition, suggest that reaction ͑1͒
proceeds via intermediate complexes which are supported by
recent theoretical study.16 Because both spin-orbit states lead
to the intermediate complexes, AlO2, in which the energy
randomizes, the rotational and vibrational state distributions
become similar for both reactant states.
In the present study, we wish to present experimental
results for the other half of the products of reaction ͑1͒, i.e.,
the oxygen atom. The O atom was detected by vacuum ul-
traviolet laser-induced fluorescence ͑LIF͒ and the relative
populations of its spin-orbit states, O͑3PJ͒ ͑J=2,1,0͒, were
determined at 12.2 kJ/mol of collision energy. The Doppler
line profiles of the LIF transitions were also analyzed to ob-
tain energy partitioning into the products of reaction ͑1͒.
Combined with the previous study, more detail information
was derived for the potential energy surfaces and the mecha-
nism of reaction ͑1͒.
Effects of electronic excitation on chemical reactions
have been one of the important issues in reaction dynamics.
They are especially important for reactions of transition-
metal atoms which usually have several low-lying electroni-
cally excited states.1–5 In the electronic excitation, spin-orbit
states have unique properties from both dynamical and ki-
netical points of view. Atoms with both nonzero electronic
angular momentum and spin multiplicity have spin-orbit fine
structure and the spin-orbit effects, which describe the differ-
ences in reactivity of individual levels of different total an-
gular momentum, J, are sometime significant even in the
cases of small fine-structure splittings.6–9 The effects of spin-
orbit fine structures on reactivities have special importance
for reaction kinetics, since the energy differences of the spin-
orbit levels in many atoms are small compared with thermal
energy and excited states have significant populations and
contributions to thermal rate constants.
One of the metal-atom reaction systems in which the
spin-orbit effects have been studied is the reaction,10
Al͑2P1/2,3/2͒ + O2͑X 3⌺−g͒ → AlO͑X 2⌺+͒ + O͑3PJ͒
⌬rH00 = − 14.96 kJ/mol
͑1͒
II. EXPERIMENT
The ground-state electron configuration of the Al atom forms
two spin-orbit states, Al͑2P1/2͒ and Al͑2P3/2͒, whose splitting
is rather small, 112.04 cm−1; however, the effect is signifi-
cant and depends on collision energy.11–14 At low collision
energies, the spin-orbit excited state, Al͑2P3/2͒, is less reac-
tive with O2 than the ground state, Al͑2P1/2͒. The relative
cross section of Al͑2P3/2͒ increases with the collision energy
and becomes comparable to that of Al͑2P1/2͒. Recently, we
studied reaction ͑1͒ by using the crossed-beam laser-induced
The experiment was carried out by using a crossed-beam
apparatus described in the previous paper.15 Briefly, the ap-
paratus consisted of three chambers which were differen-
tially pumped. Al atoms were generated by laser vaporiza-
tion. The fourth harmonic of an yttrium aluminum garnet
͑YAG͒ laser ͑Spectra Physics GCR-150-10͒ was focused
onto an Al rod ͑99.999% purity, Goodfellow Cambridge
Limited͒ which was rotated and translated by a stepping mo-
tor. The vaporized Al atoms were cooled and issued as an
atomic beam by carrier gas flow expanded through a pulsed
valve ͑Jordan PSV, 1.0-mm diameter͒ A pure oxygen mo-
a͒
Author to whom correspondence should be addressed; electronic mail:
honma@sci.u-hyogo.ac.jp
0021-9606/2005/122͑20͒/204312/6/$22.50
122, 204312-1
© 2005 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:
193.61.135.80 On: Mon, 15 Dec 2014 17:56:07

204312-2
Ishida et al.
J. Chem. Phys. 122, 204312 ͑2005͒
lecular beam was delivered from another pulsed valve ͑an
Even–Lavie valve, 0.2-mm diameter͒.17 Both beams were
collimated by skimmers ͑Beam Dynamics, 2-mm diameter͒
and crossed each other at a right angle in a reaction chamber.
The metal-atomic beam source, the O2 beam source, and the
reaction chamber were pumped by 6-, 10-, and 6-in. diffu-
sion pumps ͑Edwards, Difstak͒, respectively. When both
beams were operated, typical background pressures of the
source chambers and reaction chamber were 1ϫ10−5 and
4ϫ10−6 Torr, respectively.
In the present measurements, Ne and N2 were used as
the carrier gases for the Al atom. Both carrier gases provide
the Al atomic beam of the same peak velocity with a similar
width ͑±6%͒, 1045 m/s.15 The unique properties of N2 and
Ne carrier gases are the relative populations of two spin-orbit
states of Al, Al͑2P1/2͒, and Al͑2P3/2͒. The N2 carrier provides
mostly the ground spin-orbit state, Al͑2P1/2͒, while the Ne
carrier provides both Al͑2P1/2͒ and Al͑2P3/2͒ with compa-
rable concentrations.12,15 Although the relative concentra-
tions of the spin-orbit states were not determined prior to
each measurement in this study, it could be stated from our
previous measurements that the Al͑2P1/2͒ concentration in
the beam is more than 80% for the N2 carrier and 50% for
the Ne carrier. Since the metal beam source has no extension
channel, the concentration of Al clusters is believe to be
negligible.
The VUV fluorescence from the crossing region was col-
lected by a MgF2 lens ͑f =50 mm͒ and detected by a VUV
solar blind photomultiplier tube ͑EMR 541G-08-17͒. The
output of the photomultiplier tube was amplified by a pre-
amplifier ͑SRS SR-445͒ and fed into a gated integrator ͑SRS
SR-250͒. The excitation VUV laser intensity was measured
by a Ceratron electron multiplier located near the outlet
Brewster window of the reaction chamber and its signal was
amplified and averaged by another gated integrator. The av-
eraged signals of the fluorescence and laser intensity were
stored in a computer for further analysis.
As described in the previous paper, the Al beam contains
a trace amount of AlO. We also observed the LIF signal of
the O atom even without the O2 molecular beam crossing
with the Al beam. As discussed in Sec. III, this LIF signal
originates from the O atom formed at the laser vaporization
region. In order to eliminate the contribution of the O atom
in the Al beam, the O2 beam was operated with 5 Hz while
the Al beam was operated with 10 Hz, and the 5-Hz signal
synchronized with O2 beam and that synchronized without
O2 were averaged separately by two gated integrators and
accumulated by a computer. The difference of the two sig-
nals was obtained by simple subtraction.
III. RESULTS AND DISCUSSION
A. O atom in Al beam
The velocity of the O2 beam was measured by a fast
ionization gauge ͑Beam Dynamics, FIG͒. The determined
velocity was 760 m/s͑±7%͒ and the collision energy was
12.2 kJ/mol. After the previous study, a quadrupole mass
spectrometer ͑ULVAC, MSQ-150A͒ was installed on the O2
beam axis. Since the Lavie–Even valve has been used to
generate large clusters,17 it has to be confirmed that the con-
tribution of clusters is minor in this study. Under the experi-
mental condition used here, no signal more than the noise
level was observed at m/z=64, ͑O2͒+2. Although an electron-
impact ionization may cause some fragmentation, this result
suggests that the concentration of the O2 dimer and larger
clusters is negligible.18
The O atom was detected by the vacuum ultraviolet
͑VUV͒ LIF technique by using the transition, O͑3S10-3PJ͒.
The vacuum ultraviolet light was generated by the four-wave
difference frequency mixing in Kr. Two dye lasers ͑Con-
tinuum ND-6000 and ND-60 using DCM as a dye in both
lasers͒ were pumped by a Nd:YAG laser ͑Continuum NY-
82͒. The frequency of one dye laser was tripled by KD*P and
beta barium borate ͑BBO͒ crystals and tuned with the two-
photon transition to one of the excited states of Kr,
5p͓5/2,2͔. This UV laser beam, 46 154.1 cm−1 ͑around
217 nm͒, and the output of the other dye laser were propa-
gated coaxially and focused into a conversion cell by a
quarts lens ͑f =200 mm͒. The conversion cell containing Kr
was located in the center of the reaction chamber and the
propagation axis of the laser beams was perpendicular to the
axes of atomic and molecular beams. A MgF2 plate was used
for the output window of the Kr cell. No device was used to
separate the VUV laser light from the UV and visible ones,
then all laser lights were irradiated on the two-beam crossing
region.
As mentioned in the Experiment section, the LIF signal
of the O atom was observed without the O2 beam. One pos-
sible origin is the O atom generated in the laser vaporization
region, i.e., the O atom formed by the dissociative desorption
of oxidized surface of the Al rod and/or by the reactions of
the vaporized Al with trace of oxidant in the carrier gas.
Another possibility is the photolysis of AlO in the Al beam at
the crossing region. Since the bond dissociation energy of
AlO is 5.27 eV,12 the small amount of AlO in the Al beam
could be photolyzed by the UV ͑46 154.1 cm−1 correspond-
ing to 5.72 eV͒ and/or VUV laser lights. If the second pro-
cess occurs with a certain extent, the photodissociation of
the AlO formed by reaction ͑1͒ should be also taken into
account.
Figure 1 shows the O͑3S10-3P2͒ LIF intensity as a func-
tion of the VUV laser intensity. In this measurement, the
intensity of the VUV laser was changed by various manners,
i.e., changing UV and/or visible laser intensities, and the
pressure of Kr in the conversion cell. Although the data taken
by using the UV and visible lasers of various intensities are
shown in one plot, the LIF intensity seems to depend linearly
on the VUV laser intensity. The UV laser intensity shows no
effect on the LIF intensity of the O atom. This result strongly
suggests that the photodissociation of AlO by the UV laser
͑46 154.1 cm−1͒ is negligible in this experimental condition
and the O atom observed in the Al beam originated at the
laser vaporization region. The linear dependence shown in
this figure also confirms that the O͑3S10-3PJ͒ transition is not
saturated for the VUV laser intensity used in this study.
The above conclusion is supported by the relative popu-
lations of the spin-orbit states of O͑3PJ͒ in the Al beam. We
observed that O͑3P2͒ had a much larger population com-
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:
193.61.135.80 On: Mon, 15 Dec 2014 17:56:07

O͑3PJ͒ atoms from Al+O2
J. Chem. Phys. 122, 204312 ͑2005͒
204312-3
FIG. 1. VUV laser intensity dependence of LIF intensity of O͑3P2͒ in Al
beam.
pared with the other two states, as summarized in Table I.
The observed high population of the lowest-energy spin-orbit
states can be explained by the supersonic expansion follow-
ing to the O atom formation. The O atoms formed at the laser
vaporization region are cooled by many collisions with car-
rier gas molecules.
3
B. Populations of O„ PJ… formed by reaction „1…
The LIF spectra of the O atom formed by reaction reac-
tion ͑1͒ were obtained by eliminating the contribution of the
O atom in the Al beam. The Al and O2 beams were operated
with 10 and 5 Hz, respectively. The 5-Hz signal synchro-
nized with the O2 beam and that synchronized without O2
was averaged separately by two gated integrators and accu-
mulated by a computer. The difference of the two signals was
obtained by simple subtraction. The difference was then di-
vided by the VUV laser intensity to obtain the normalized
LIF intensity, since the absorption of O͑3S10-3PJ͒ depends
linearly on the VUV laser intensity, as shown in Fig. 1. The
LIF spectra were measured several times and averaged. Fig-
ure 2 shows the averaged spectra for the N2 carrier gas.
Each spectrum was fitted by the Gaussian function to
obtain an area of the absorption line. In this figure, the best
fits are shown as broken lines. Since the transitions used here
have a common upper state, O͑3S10͒, and the absorption line
strengths for three O͑3PJ͒ states are almost identical,19 no
further correction was necessary to determine the popula-
tions of three states. The results of relative populations are
FIG. 2. LIF spectra of O͑3PJ͒ ͑J=0,1,2͒ formed by the reaction of Al
+O2. The broken lines are Gaussian fits for the spectra.
summarized in Table I and Fig. 3. Those determined for the
Al beam of the Ne carrier were also shown in this table.
Because the populations of the highest-energy O͑3P0͒ have
the largest errors, all relative populations are given with re-
spect to those of O͑3P1͒.
TABLE I. Relative populations of spin-orbit states of O͑3PJ͒ formed by
reaction ͑1͒.
J=2
J=1
J=0
Population of O͑3PJ͒
Al/N2
3.9±1.0
3.8±0.8
1.95
1.0±0.1
1.0±0.2
1.00
0.2±0.1
a
Al/Ne
Statistical expectation
in Al/N2 beam
0.31
FIG. 3. Relative populations of O͑3PJ͒ ͑J=0,1,2͒ formed by the reaction of
Al+O2. The open circles and solid circles are those for the Al beams of N2
carrier and Ne carrier, respectively. Also shown by solid squares are the
population expected from statistical energy partitioning into products.
23±10
1.0±0.3
0.4±0.1
aBecause of poor signal-to-noise ratio, reliable measurement could not be
accomplished.
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:
193.61.135.80 On: Mon, 15 Dec 2014 17:56:07

204312-4
Ishida et al.
J. Chem. Phys. 122, 204312 ͑2005͒
Compared with the Al beam of the N2 carrier, that of the
Ne carrier contains more spin-orbit excited state, Al͑2P3/2͒.
Because this state is less reactive with O2 than the ground
state, Al͑2P1/2͒, the O atom signal for the Al beam of the Ne
carrier was significantly weaker and reliable population for
O͑3P0͒ could not be determined. Except for O͑3P0͒, the
populations determined for two carrier gases agree very well.
These results suggest that two spin-orbit states of Al provide
no significant effect on the spin-orbit state populations of the
product O atom.
The determined populations could be compared with
those expected from statistical energy partitioning in the
products, AlO+O. The statistical populations of the spin-
orbit states of O͑3PJ͒ were estimated by
1/2
N0͑J͒ ϰ ͑2J + 1͒
͑2N + 1͒͑Eav − Ev − Er − Eso͒
͚ ͚
N
v
ϰ ͑2J + 1͒
N0͑ ,N͒,
v
͚ ͚
N
v
where Eav, Ev, Er, N, and Eso are total available energy of the
reaction, vibrational energy of AlO, its rotational energy, ro-
tational quantum number, and spin-orbit energy of the O
atom, respectively. N0͑ ,N͒’s are the statistical distribution
v
of AlO with vibrational and rotational quantum numbers,
v
and N, respectively. The estimated values were
1.95:1.00:0.31 for J=2:J=1:J=0, respectively. The present
results show that reaction ͑1͒ forms more ground spin-orbit
state, O͑3P2͒, and less excited state than the statistical expec-
tation, i.e., the observed populations deviate significantly
from the statistical ones.
FIG. 4. LIF spectra of O͑3PJ͒ ͑J=1,2͒ formed by the reaction of Al with
O2. Also shown by broken lines are convoluted spectra based on the statis-
tical energy partitioning and linewidths of VUV laser.
C. Analysis of LIF line shape
components along the VUV laser propagation axis. As men-
tioned in Sec. III B, the O atom in the Al beam consists of
mostly the ground spin-orbit state, O͑3P2͒; the O͑3S10-3P2͒
transition was measured and fitted by the Gaussian function
to determine the linewidth. The averaged linewidth was
0.93±0.12 cm−1 ͓full width at half maximum ͑FWHM͔͒.
The convoluted spectra are shown as broken lines with
the observed spectra in Fig. 4. Because the expected Doppler
shifts of the O atom are small, the line shapes are mostly
determined by the laser linewidth and insensitive to recoil
angular distribution. The convoluted line shapes represent
the observed ones reasonably well and these suggest that the
line shapes of the O atoms are consistent with the energy
partitioning of reaction ͑1͒ reported previously.
Generally, the Doppler line shapes of the O atom transi-
tions contain some information about recoil energy in reac-
tion ͑1͒. Because the reaction has only three atoms, the recoil
translational energy directly correlates to the rotational and
vibrational energy partitioning into the counterproduct AlO,
which have been reported to be almost statistical at low col-
lision energies.12,15,20 The line shapes expected from the sta-
tistical energy partitioning were convoluted and compared
with the observed ones.
The distributions of the vibrational and rotational states
of AlO expected from statistical energy partitioning were es-
timated by N0͑ ,N͒ϰ͑2N+1͒͑E −E −E −E ͒1/2, as de-
v
av
v
r
so
scribed in Sec. III B. The recoil energy, Eav−Ev −Er−Eso,
was converted to the recoil velocity and the Doppler shift of
the O atom absorption. Since the propagation axis of the
VUV laser is perpendicular to the scattering plane defined by
atomic and molecular beams, the conversion of the center-
of-mass ͑c.m.͒ recoil energy to the velocity component along
the laser propagation axis is straightforward. Since no infor-
mation about the differential cross section of reaction ͑1͒ is
available, an isotropic c.m. angular distribution was assumed
for the convolution of the Doppler profile.
D. Reaction mechanism
The results observed in this study are summarized as the
following:
͑A͒ The populations of the spin-orbit states, O͑3PJ͒, show
significant deviation from the statistical expectation.
͑B͒ These populations do not show the dependence on the
spin-orbit states of reactant, Al͑2P1/2,3/2͒.
The linewidths of the O atom transitions observed in the
Al beam were used for the linewidths of the VUV laser.
These O atoms formed in the source region are cooled by the
supersonic expansion and expected to have narrow velocity
The dynamics of reaction ͑1͒ has been studied by mea-
suring the rotational and vibrational distributions of AlO.
Most studies have consistently reported that the rotational
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:
193.61.135.80 On: Mon, 15 Dec 2014 17:56:07

O͑3PJ͒ atoms from Al+O2
J. Chem. Phys. 122, 204312 ͑2005͒
204312-5
and vibrational distributions of AlO are almost statistical and
suggested a mechanism forming a long-lived intermediate
complex.11,12,20–22 Recently, we determined the rotational and
vibrational distributions for the specific spin-orbit state,
Al͑2P1/2͒ or Al͑2P3/2͒, and observed that the rotational dis-
tributions were quite similar to each other.15 The different
reactivities and the similar rotational distributions of AlO
were explained by the formation of an intermediate complex
in the course of reaction ͑1͒. At low collision energies, it has
been suggested that the long-range interaction potentials be-
tween two spin-orbit states and O2 are different, i.e., electro-
static quadrupole-quadrupole interaction for Al͑2P3/2͒ and
dispersion force for Al͑2P1/2͒.13,14 Because the quadrupole-
quadrupole interaction depends on the orientation of the ap-
proach of Al to O2, Al͑2P3/2͒ could form the intermediate
only at a limited range of orientations while Al͑2P1/2͒ forms
the intermediate complex at any orientation of approach. Al-
though the steric factor for Al͑2P3/2͒ leads to lower reactiv-
ity, the system loses memory of the initial spin-orbit states in
the intermediate complex, and provides the same rotational
distribution for both spin-orbit states of Al. The complex
formation for both spin-orbit states of Al may be consistent
with one of the results observed in this study, i.e., no differ-
ence in the population of O͑3PJ͒ was observed for two spin-
orbit states, Al͑2P1/2,3/2͒. However, the deviation from the
statistical populations needs further consideration about
mechanism.
the degenerated states of Al͑2P1/2͒+O2͑X 3⌺g−͒ can adiabati-
2
cally correlate with the A2 intermediate which is energeti-
cally accessible in the collision energy of 12.2 kJ/mol. Since
most of the experimental results have suggested the forma-
tion of intermediate complexes for both spin-orbit states,
there must be an interaction between a potential surface
evolving from Al͑2P3/2͒+O2͑X 3⌺−g͒ and the surface connect-
ing Al͑2P1/2͒+O2͑X 3⌺g−͒ and the A2 intermediate.
2
More important for the branching of O͑3PJ͒ is the cor-
relation between the intermediates and products. As Pak and
Gordon indicate, the symmetry changes from C2v to Cϱv; it is
reasonable to use the Cs symmetry group to connect them. In
the Cs symmetry group, two intermediates and AlO͑X 2⌺+͒
2
+O͑3PJ͒ belong to the same E1/2 symmetry. Since the A2
͑2AЉ in Cs symmetry group͒ intermediate and the lowest-
energy AlO͑X 2⌺+͒+O͑3P2͒ belong to the same symmetry,
they may correlate adiabatically, while the surface crossing is
necessary for the potential surfaces without taking into ac-
count the spin-orbit interaction. This adiabatic correlation is
consistent with the high population of O͑3P2͒ observed in
this study. On the other hand, an interaction with other po-
tential surfaces is necessary for the O͑3P1,0͒ formation, since
the potential surfaces for these products could not correlate
2
with the intermediate A2. Further discussion requires detail
information about other electronic states at the intermediate
region and the interaction among the surfaces from them and
2
the A2 intermediate.
Recent theoretical study by Pak and Gordon16 has pro-
vided useful information about the mechanism of reaction
͑1͒. They studied the lowest two doublet potential surfaces.
IV. SUMMARY
2
2
Both surfaces, A2 and A1 in C2v symmetry, have deep po-
tential wells. They also observed that there is no direct path-
way from Al+O2 to AlO+O. These results strongly suggest
that reaction ͑1͒ proceeds via the intermediate complex,
The relative populations of three spin-orbit states,
O͑3PJ͒, were determined for the reaction, Al+O2→AlO+O,
under crossed-beam condition. The measurements were
achieved for two experimental conditions where the relative
populations of two spin-orbit states of Al, Al͑2P1/2,3/2͒, are
different. The observed relative populations were almost
identical for two conditions and suggested the absence of the
effect of the initial spin-orbit states. The lowest-energy
O͑3P2͒ had the highest population which is twice as high as
that expected from the statistical energy partitioning, while
the highest-energy O͑3P0͒ populated less than the statistical
expectation. Combined with the previous results, the results
obtained in this study suggest that the reaction proceeds via
the intermediate complex and the branching of the spin-orbit
states is determined at the exit channel.
2
AlO2. Since the surface leading to the A1 intermediate has
an energy barrier of around 20 kcal/mol, the reaction is
2
likely to proceed via the A2 intermediate in the collision
energy studied here. From the intermediate to the product,
AlO͑2⌺+͒+O͑3PJ͒, the 2A2 intermediate which belongs to
the A symmetry in C group needs the crossing to the A
Љ
Ј
s
surface.
The formation of the intermediate complex might imply
that the populations of the spin-orbit states of product,
O͑3PJ͒, become statistical as observed for rotational and vi-
brational states. However, the results observed in this study
show significant deviation from the statistical expectation.
Unfortunately, the study by Pak and Gordon does not include
the spin-orbit interaction, then no information can be derived
about the mixing of the initial spin-orbit states, Al͑2P1/2,3/2͒,
and the branching to the final spin-orbit states, O͑3PJ͒. One
simple way to derive this information is to consider the cor-
relation among reactants, intermediates, and products based
on the extended point group.23 For the reactants and interme-
diates, the C2v symmetry group can be applied. In this sym-
1D. E. Clemmer, K. Honma, and I. Koyano, J. Phys. Chem. 97, 11480
͑1993͒.
2K. Honma, M. Nakamura, D. E. Clemmer, and I. Koyano, J. Phys. Chem.
98, 13286 ͑1994͒.
3K. Senba, R. Matsui, and K. Honma, J. Phys. Chem. 99, 13992 ͑1995͒.
4R. Matsui, K. Senba, and K. Honma, J. Phys. Chem. A 101, 179 ͑1997͒.
5C. L. Haynes and K. Honma, J. Chem. Soc., Faraday Trans. 94, 1171
͑1998͒.
6P. J. Dagdigian and M. L. Campbell, Chem. Rev. ͑Washington, D.C.͒ 87,
1 ͑1987͒.
metry group, both Al͑2P1/2͒+O2͑X 3⌺−g͒ and Al͑2P3/2
͒
7J. Zhou, J. J. Lin, B. Zhang, and K. Liu, J. Phys. Chem. 108, 7832 ͑2004͒.
8S.-H. Lee, L.-H. Lai, and K. Liu, J. Chem. Phys. 110, 8229 ͑1999͒.
9S. A. Nizkorodov, W. W. Harper, W. B. Chapman, B. W. Blackmon, and
D. J. Nesbitt, J. Chem. Phys. 111, 8404 ͑1999͒.
+O2͑X 3⌺g−͒ belong to the E1/2 symmetry. Since both inter-
2
2
mediates, A2 and A1, also belong to the E1/2 symmetry, the
lowest two reactant states correlate to these intermediates.
However, this correlation scheme indicates that only one of
10Derived from D00͑O–O͒=5.114 eV ͓K. P. Huber and G. Herzberg, Mo-
lecular Spectra and Molecular Structure, Constants of Diatomic Mol-
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:
193.61.135.80 On: Mon, 15 Dec 2014 17:56:07

204312-6
Ishida et al.
J. Chem. Phys. 122, 204312 ͑2005͒
ecules ͑van Nostrand, Princeton, 1979͔͒ and D00͑Al–O͒=5.27 eV ͑Ref.
12͒.
because dimer ion signals were observed for Ar or NO beams under the
same ionization condition.
19The absorption strengths of O͑3S10-3PJ͒ are given as 5.20, 5.19, and 5.19
ϫ10−2 for J=2, 1, and 0, respectively, in W. L. Wiese, J. R. Fuhr, and T.
M. Deters, J. Phys. Chem. Ref. Data Monogr. 7, 340 ͑1996͒.
20M. Costes, C. Naulin, G. Dorthe, C. Vaucamps, and G. Nouchi, Faraday
Discuss. Chem. Soc. 84, 74 ͑1987͒.
11K.-M. Chen, C.-H. Sung, J.-L. Chang, T.-H. Chung, and K.-H. Lee, Chem.
Phys. Lett. 240, 17 ͑1995͒.
12C. Naulin and M. Costes, Chem. Phys. Lett. 310, 231 ͑1999͒.
13T. Stoecklin, Faraday Discuss. 110, 231 ͑1998͒.
14D. Reignier, T. Stoecklin, S. D. Le Picard, A. Canosa, and B. R. Rowe, J.
Chem. Soc., Faraday Trans. 94, 1681 ͑1998͒.
21P. J. Dagdigian, H. W. Cruse, and R. N. Zare, J. Chem. Phys. 62, 1824
͑1975͒.
15K. Honma, J. Chem. Phys. 119, 3641 ͑2003͒.
16M. V. Pak and M. S. Gordon, J. Chem. Phys. 118, 4471 ͑2003͒.
17U. Even, J. Jortner, D. Noy, N. Lavie, and C. Cossart-Magos, J. Chem.
Phys. 112, 8068 ͑2000͒.
22L. Pasternack and P. J. Dagdigian, J. Chem. Phys. 67, 3854 ͑1977͒.
23G. Herzberg, Molecular Spectra and Molecular Structure. III Electronic
Spectra and Electronic Structure of Polyatomic Molecules ͑Van Nostrand
Reinhold, New York, 1966͒.
18We do not believe that all clusters are dissociated by the electron impact
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:
193.61.135.80 On: Mon, 15 Dec 2014 17:56:07