Introduction

The reaction

is a rate determining step in the combustion of hydrogen and hydrocarbons. Furthermore, it accounts for approximately 80 % of the O

Dynamical studies on this reaction have mainly used three different
potential energy surfaces (PESs), the Melius-Blint surface[46],
the DMBE IV surface of Pastrana *et al*,[47] and the
Diatomics-in-Molecules (DIM) surface of Kendrick and Pack.[48] At
first, only quasi-classical calculations were
possible.[17,18,19,20,21,22,23,24,25,26,27]
Only in the last five years have full quantum mechanical calculations become
possible with most calculations performed for total angular momentum
*J*=0.[28,29,30,31,32,33,34,35,36,37,38]
Whenever *J*>0 was studied, it was done using an approximate
method.[39,40,41,42] The first rigorous
quantum mechanical study for *J*>0 (*J*=1, *J*=2, and *J*=5), as far as we
know, was published only recently by us.[43] This paper is
referred to as paper I from now on.

Quantum mechanical calculations on the H + O_{2}(
)
reaction
are quite difficult. There are two reasons for this. First, the O + OH exit
channel has a long-range *R*^{-4} character, where *R* is the H to
center-of-mass O_{2} distance. This ensures that one needs large grids for the
calculation and in case of an iterative method also long propagation times.
Second, the PES has a deep well (2.38 eV below the H + O_{2} asymptote), which
can support many metastable states of the intermediate HO_{2} complex.
Consequently, a large number of studies have been devoted to the
HO_{2} complex as
well,[49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67]
including one that treats *J*>0 rigorously.[55]

There is evidence that two different reaction mechanisms play a role in the H
+ O_{2}(
)
reaction.[12,13,14,15] At low
collision energies, the reaction goes through a long-lived reaction complex.
This is evidenced by the sharp resonance structure seen in the theoretical
*J*=0 reaction probability as a function of
energy.[28,29,30,31,43]
Moreover, the experimental differential cross sections show no preference for
forward or backward scattering at these energies.[14] At higher
energies, a second ``direct'' mechanism becomes more important. The
experimental differential cross sections at these energies show a definite
preference for forward scattering, consistent with a direct
mechanism.[11,14] The theoretical *J*=0 reaction
probability shows a sharp increase at higher energies and the resonance
structure becomes less
pronounced.[28,29,30,31,43] In
calculations for *J*>0 (*J*=1, *J*=2, and *J*=5) evidence for the existence
of the two reaction mechanisms is also found.[43]

The intermediate HO
complex is a very floppy species.[68]
This means that the substates
of the wave function for total angular
momentum quantum number *J* are probably heavily coupled,[30]
so that the Coriolis coupling between the states cannot be ignored. Our
calculations for *J*=1, *J*=2, and *J*=5 in paper I show that, if the H -
O_{2} distance *R* is taken to be the *z*-axis of the coordinate system, the
substates are indeed heavily coupled. However, our choice for the
*z*-axis in paper I is not the only choice possible. In fact, previous
(approximate) calculations for H + O_{2} (*J*>0) have used the O_{2} bond as
the *z*-axis.[40,41,42] Based on the mass
difference between H and O_{2}, one would expect this choice for the *z*-axis
to result in a calculation in which the Coriolis coupling is less
important.[42] However, given the floppiness of the
HO_{2} complex, this is by no means certain.[30]

The heavy coupling between the
substates for a wave function with
total angular momentum quantum number *J* means that the H + O_{2} reaction is
well suited to be studied with the Coriolis coupled method of Goldfield and
Gray.[69,70,71] In this method the different
substates are distributed over the processors of a parallel
computer, resulting in a highly parallel calculation with almost no
communication overhead.

In this paper we examine the importance of Coriolis coupling in the
theoretical description of the H + O_{2} (
)
reaction. To
this end we have performed (rigorous) coupled channel (CC)
calculations[72] and (approximate) helicity (*J*_{z}) conserving
(HC) calculations[73] for *J*=1, *J*=2, *J*=5, and *J*=10 with
both choices for the *z*-axis. In the HC calculations the Coriolis coupling
between the
states is neglected. The CC calculations for *J*=10 have
not been published before. We discuss these result separately in
Sec. IIIA. The CC calculations with *R* as *z*-axis for *J*=1,
*J*=2, and *J*=5 were taken from paper I.

The organization of the paper is as follows. In Sec. II we discuss
the theory needed for the CC and HC calculations, we present the
transformations between the two choices for the *z*-axis that we use and some
computational details. In Sec. IIIA we discuss the results for the
*J*=10 CC calculations. In Sec. IIIB we compare the HC calculations
to the CC calculations. Finally, in Sec. IV we give our
conclusions.