Dynamics calculations for the Cl¿C
2
H
6
abstraction reaction:
Thermal rate constants and kinetic isotope effects
A. Ferna
´
ndez-Ramos
a)
and E. Martı
´
nez-Nu
´
n
˜
ez
Departamento de Quı
´
mica Fı
´
sica, Universidad de Santiago de Compostela, Santiago de Compostela
E-15706, Spain
J. M. C. Marques
Departamento de Quı
´
mica, Universidade de Coimbra, P-3049 Coimbra Codex, Portugal
S. A. Va
´
zquez
Departamento de Quı
´
mica Fı
´
sica, Universidad de Santiago de Compostela, Santiago de Compostela
E-15706, Spain
͑Received 13 November 2002; accepted 14 January 2003͒
Thermal rate constants and H/D kinetic isotope effects for the ClϩC
2
H
6
→HClϩC
2
H
5
reaction
were computed by microcanonical variational transition state theory on a high-level ab initio
potential energy surface ͑PES͒. The calculations show that the reaction proceeds through a ‘‘loose’’
transition state, and so the consideration of variational effects is important. The PES presents a
van der Waals minimum in the products side. The calculations indicate that this minimum has no
effect in the forward reaction and little effect in the reverse reaction for temperatures above room
temperature. The analysis of the kinetic isotope effects shows that the contribution due to tunneling
is fairly small, but with an important role played by the variational effects. Classical trajectory
calculations were also performed on a semiempirical PES, which was parametrized from own ab
initio calculations. This method was utilized to compare the HCl relative velocity distribution with
those obtained experimentally. The results show a good agreement with experiment and reinforce
the reliability of the proposed mechanism for this reaction. © 2003 American Institute of Physics.
͓DOI: 10.1063/1.1557453͔
I. INTRODUCTION
The hydrogen atom abstraction from hydrocarbons by
chlorine is a key-step in many processes that are relevant to
both atmospheric
1,2
and combustion
3,4
chemistry. One of
such reactions involving methane has been the subject of a
great amount of experimental and theoretical work in the
past few years ͑see, for instance, Ref. 5 and references
therein͒. Several studies of this reaction in the temperature
interval 180рT/Kр800 have shown a non-Arrhenius behav-
ior for its rate constant dependence with temperature. This
has been attributed to tunneling and a small enhancement of
the rate constant due to a low-frequency bending mode of
CH
4
in the low-temperature regime, and to the excited sym-
metric and asymmetric stretching modes for TϾ300 K.
However, the rate constants for the ClϩCH
4
reaction at
stratospheric and combustion temperatures are still uncertain,
because the kinetic measurements constitute a difficult task
in these cases. Moreover, the analysis of the kinetic isotope
effect ͑KIE͒ for ClϩCH
4
/CD
4
͑at all temperatures, the ratio
of the rate constants is greater than one͒ indicates that tun-
neling enhances the reaction probability by partially relaxing
the steric restrictions for the collinear geometry of the tran-
sition state.
5
The ClϩC
2
H
6
→C
2
H
5
ϩHCl reaction constitutes a natu-
ral extension of the ClϩCH
4
one, and is currently a chal-
lenging subject for theoreticians. In contrast to ClϩCH
4
,
experiments for the analogous reaction with ethane predict
larger values of the rate constants and essentially no signifi-
cant temperature dependence. For instance, Pilgrim et al.
6
have recently measured a value of 5.51ϫ10
Ϫ11
cm
3
molecule
Ϫ1
s
Ϫ1
at room temperature with an activation
energy of only 0.2Ϯ0.2 kcal/mol.
7
In addition, it has been
observed
7
that the pre-exponential factor is greater for Cl
ϩC
2
H
6
than that for ClϩCH
4
, which suggests a reaction
proceeding via a ‘‘loosely’’ constrained transition state ͑TS͒
geometry. The measurements of state-specific angular distri-
butions carried out by Kandel et al.
8
led them to conclude
that the reaction proceeds through a ‘‘loose’’ transition state.
Those authors also indicated that the dynamics of chlorine
with ethane should be quite similar to that of chlorine with
methane, i.e., both reactions take place via direct dynamics
without the participation of complexes.
To our knowledge there are three ab initio studies
9–11
on
this system, which predict a nearly linear TS for different
levels of theory. However, the best results of Bottoni and
Poggi
9
indicate a potential barrier of 10.77 kcal mol
Ϫ1
, and
an estimated activation energy 7.65 kcal mol
Ϫ1
, which is in
total disagreement with experimental reactivity. Conversely,
Roberto-Neto and Machado
10,11
performed MP2
12
͑Mo
¨
ller-
Plesset correlation energy correction truncated at second or-
der͒ and coupled-cluster ͓CCSD͑T͔͒
13
calculations with dif-
ferent correlation-consistent basis sets ͑double-zeta,
augmented double-zeta, and triple-zeta͒ and obtained barriers
a͒
Electronic mail: qftramos@usc.es; Fax ͑ϩ34͒ 981 595 012.
JOURNAL OF CHEMICAL PHYSICS VOLUME 118, NUMBER 14 8 APRIL 2003
62800021-9606/2003/118(14)/6280/9/$20.00 © 2003 American Institute of Physics