A reinvestigation of the kinetics and the mechanism of the CH 3C(O)O2 + HO2 reaction using both experimental and theoretical approaches

Article abstract of DOI:10.1039/b518321a

The kinetics and the mechanism of the reaction CH₃C(O)O ₂ + HO₂ were reinvestigated at room temperature using two complementary approaches: one experimental, using flash photolysis/UV absorption technique and one theoretical, with quantum chemistry calculations performed using the density functional theory (DFT) method with the three-parameter hybrid functional B3LYP associated with the 6-31G(d,p) basis set. According to a recent paper reported by Hasson et al., [J. Phys. Chem., 2004, 108, 5979-5989] this reaction may proceed by three different channels: CH₃C(O)O ₂ + HO₂ → CH₃C(O)OOH + O₂ (1a); CH₃C(O)O₂ + HO₂ → CH₃C(O)OH + O₃ (1b); CH₃C(O)O₂ + HO₂ → CH₃C(O)O + OH + O₂ (1c). In experiments, CH ₃C(O)O₂ and HO₂ radicals were generated using Cl-initiated oxidation of acetaldehyde and methanol, respectively, in the presence of oxygen. The addition of amounts of benzene in the system, forming hydroxycyclohexadienyl radicals in the presence of OH, allowed us to answer that channel (1c) is <10%. The rate constant k₁ of reaction (1) has been finally measured at (1.50 ± 0.08) × 10^-11 cm ³ molecule^-1 s^-1 at 298 K, after having considered the combination of all the possible values for the branching ratios k_1a/k_1, k_1b/k_1, k _1c/k₁ and has been compared to previous measurements. The branching ratio k_1b/k₁, determined by measuring ozone in situ, was found to be equal to (20 ± 1)%, a value consistent with the previous values reported in the literature. DFT calculations show that channel (1c) is also of minor importance: it was deduced unambiguously that the formation of CH₃C(O)OOH + O₂ (X ³Σ-g) is the dominant product channel, followed by the second channel (1b) leading to CH₃C(O)OH and singlet O₃ and, much less importantly, channel (1c) which corresponds to OH formation. These conclusions give a reliable explanation of the experimental observations of this work. In conclusion, the present study demonstrates that the CH₃C(O)O ₂ + HO₂ is still predominantly a radical chain termination reaction in the tropospheric ozone chain formation processes. the Owner Societies 2006.

Full text of DOI:10.1039/b518321a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
A reinvestigation of the kinetics and the mechanism of the CH C(O)O2
3
þ HO reaction using both experimental and theoretical approaches
2
Jean-Paul Le Crane, Marie-Therese Rayez, Jean-Claude Rayez and Eric Villenave*
ˆ
´ `
Received 3rd January 2005, Accepted 13th March 2006
First published as an Advance Article on the web 28th March 2006
DOI: 10.1039/b518321a
The kinetics and the mechanism of the reaction CH C(O)O þ HO were reinvestigated at room
3
2
2
temperature using two complementary approaches: one experimental, using flash photolysis/UV
absorption technique and one theoretical, with quantum chemistry calculations performed using
the density functional theory (DFT) method with the three-parameter hybrid functional B3LYP
associated with the 6-31G(d,p) basis set. According to a recent paper reported by Hasson et al.,
[
J. Phys. Chem., 2004, 108, 5979–5989] this reaction may proceed by three different channels:
CH
3
C(O)O
2
þ HO
2
- CH
3
C(O)OOH þ O
2
(1a); CH
3
C(O)O
2
þ HO
2
- CH
3
C(O)OH þ O
3
(
1b); CH C(O)O þ HO - CH C(O)O þ OH þ O (1c). In experiments, CH C(O)O and HO
3
2
2
3
2
3
2
2
radicals were generated using Cl-initiated oxidation of acetaldehyde and methanol, respectively, in
the presence of oxygen. The addition of amounts of benzene in the system, forming
hydroxycyclohexadienyl radicals in the presence of OH, allowed us to answer that channel (1c) is
ꢂ11
o10%. The rate constant k of reaction (1) has been finally measured at (1.50 ꢀ 0.08) ꢁ 10
1
3
ꢂ1 ꢂ1
s
cm molecule
for the branching ratios k1a/k1,
The branching ratio k1b/k , determined by measuring ozone in situ, was found to be equal to
20 ꢀ 1)%, a value consistent with the previous values reported in the literature. DFT
calculations show that channel (1c) is also of minor importance: it was deduced unambiguously
at 298 K, after having considered the combination of all the possible values
k
1b/k1, k1c/k
1
and has been compared to previous measurements.
1
(
3
ꢂ
that the formation of CH C(O)OOH þ O (X S ) is the dominant product channel, followed by
3
2
g
3 3
the second channel (1b) leading to CH C(O)OH and singlet O and, much less importantly,
channel (1c) which corresponds to OH formation. These conclusions give a reliable explanation of
the experimental observations of this work. In conclusion, the present study demonstrates that the
CH
3
C(O)O
2
þ HO
2
is still predominantly a radical chain termination reaction in the tropospheric
ozone chain formation processes.
radical chain termination reactions and are a source of ozone
and carboxylic acids in the atmosphere.
1
Introduction
Aldehydes (RCHO) are important trace constituents of the
atmosphere. They have both natural and anthropogenic
sources, with small primary sources associated with vehicle
exhaust and industrial activity and large secondary sources
associated with the oxidation of volatile organic com-
Among all aldehydes, acetaldehyde has been one of the most
studied as it is present in the atmosphere in large concentra-
10
ꢂ3
5,6
tions (from 1.5 ꢁ 10 molecule cm in rural areas up to
11
ꢂ3
7,8
4
ꢁ 10 molecule cm in urban or industrial areas ). Several
studies on the reactivity of the acetylperoxy radical have been
1,9–12
2
,3
pounds. Abstraction of the aldehydic H-atom by OH radi-
cals gives acyl radicals, RC(O), which combine with O to give
acylperoxy radicals, RC(O)O . Acylperoxy radicals play sev-
eral important roles in atmospheric chemistry. RC(O)O
radicals react rapidly with NO to give NO which is then
photolyzed leading to ozone formation. RC(O)O radicals
react with NO to form stable peroxyacylnitrates, RC(O)O2-
performed for the past years,
using different techniques,
2
and up to recently, the kinetics and the mechanisms of the
13
2
CH C(O)O reactions were considered as well-established.
3
2
2
In contrast, the kinetics and the mechanism of the reaction of
CH C(O)O with HO radicals are still a matter of contro-
2
3
2
2
2
versy, as the four principal investigations on the rate constant
of this reaction or on its different reaction channels all differ
1,9,11,12
2
NO , which are severe irritant compounds in photochemical
2
on many different points.
smog, and may be efficient reservoirs of NOx in the tropo-
4
sphere, as they may transport NO far from its sources. In air
1
For instance, Hasson et al. recently suggested that the
x
3 2 2
mechanism of the reaction of CH C(O)O with HO radicals
masses with low [NO ], acylperoxy radicals undergo reactions
x
proceeds neither via the only two channels (1a) and (1b) but
via the three following channels:
with HO
2
and in a lesser extent with other peroxy radicals.
with HO radicals are important
Reactions of RC(O)O
2
2
CH
CH
CH C(O)O
3
C(O)O
C(O)O
þ HO
2
þ HO
2
- CH
3
C(O)OOH þ O
- CH
C(O)OH þ O
C(O)O þ OH þ O
2
(1a)
(1b)
(1c)
Laboratoire de Physico-Chimie Mole´culaire, CNRS UMR 5803,
Universite´ Bordeaux I, 33405 Talence Cedex, France. E-mail:
e.villenave@lpcm.u-bordeaux1.fr
3
2
þ HO
2
3
3
3
2
2
- CH
3
2
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 2163–2171 | 2163

View Article Online
This observation was important as the existence of the third
channel (1c), already demonstrated in fluorinated-peroxy ra-
2 Experimental section and computational details
1
7
1
dical reactions with HO2, was proposed for the first time for
4
The experimental system used is described in detail elsewhere
and is dealt with briefly here.
1
this type of peroxy radicals. According to the authors, this
channel, leading to the formation of OH and CH C(O)O
3
radicals, may reach up to 40% of reaction (1). This result
means that this reaction would not be considered any more as
a full radical chain termination reaction in the tropospheric
ozone chain formation processes, according to the following
reactions:
2
.1 Flash photolysis experiments
A conventional flash photolysis UV absorption spectrometer
was used to monitor peroxy radical absorptions at room
temperature and atmospheric pressure. The reaction cell con-
sists of a 70 cm long, 4 cm diameter Pyrex cylinder. A
continuous flow of a reactant gas mixture was irradiated
periodically by two argon flash lamps. The flash lamp energy
is approximately 250 J. The half-life for the light pulse is 5 ms,
although problems with scattered light prevent data collection
for approximately 50–100 ms after the flash. The analyzing
beam was obtained from a deuterium lamp, passed through
the cell, dispersed using a monochromator (2 nm resolution),
detected by a photomultiplier, and transferred to a PC for
averaging and analysis. About 30–40 absorption time profiles
were acquired to reach a satisfactory signal to noise ratio. The
total gas flow was adjusted so that the cell was replenished
completely between flashes thereby avoiding photolysis of
reaction products. Decay traces were simulated by numerical
integration of a set of differential equations representative of
an appropriate chemical mechanism, selected parameters
OH þ CH CHO - CH C(O) þ H O
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
3
3
2
CH
3
C(O) þ O
2
þ M - CH
3
C(O)O
2
þ M
CH
3
C(O)O
2
þ NO - CH
3
C(O)O þ NO
2
CH C(O)O þ M - CH þ CO þ M
3
3
2
CH
3
þ O
2
þ M - CH
3
O
2
þ M
CH
3
O
2
þ NO - CH
3
O þ NO
2
3
NO þ hn - NO þ O( P)
2
3
O( P) þ O
2
þ M - O
3
þ M
(
rate constant, absorption cross section, and initial radical
O
3
þ NO - O þ NO
2
2
concentration) were adjusted to give the best nonlinear least
squares fit.
As acetaldehyde is often used as the model for the reactivity of
aldehydes in chemical modeling of atmospheric chemistry, it is
Radicals were generated by photolysis of Cl
longer than the Pyrex cut-off, using Cl /CH
/N mixtures:
2
at wavelengths
CHO/CH OH/
2
3
3
important to consider if channel (1c) is existing and to confirm
1
or not the branching ratio k1c/k measured by Hasson et al.
1
O
2
2
We report here the results of a reinvestigation of the kinetics
and the mechanism of the CH C(O)O reaction, using
Cl
2
þ hn (l > 280 nm) - 2 Cl
OH
at 298 K.
C(O)
at 298 K.
(12)
(13)
3
2
þ HO
2
Cl þ CH
13 = 5.5 ꢁ 10ꢂ11 cm molecule
3
OH - HCl þ CH
2
both experimental and theoretical approaches. Experiments
were performed at room temperature and atmospheric pres-
sure, using flash photolysis technique coupled to UV absorp-
tion spectrometry. In addition, some new numerical simu-
3
ꢂ1 ꢂ1
18
18
k
k
s
Cl þ CH
3
CHO - HCl þ CH
3
(14)
1
2
lations of the older experiments performed by Tomas et al. in
our group on the same CH C(O)O þ HO reaction also allow
ꢂ11
3
ꢂ1 ꢂ1
14 = 7.2 ꢁ 10
cm molecule s
3
2
2
For acetaldehyde, it has been shown that reaction (14)
1
us to investigate the mechanism proposed by Hasson et al.
occurs predominantly (>95%) via abstraction of the aldehy-
19,20
dic hydrogen.
and to propose a new value for the rate constant k
1
. In order to
detect any presence of OH radicals in our system, formed
according to channel (1c), an original method was developed:
it consists in adding amounts of benzene in the initial gas
mixture. The choice of benzene is fully justified as it does not
2
The concentration of Cl was monitored using its well-
ꢂ19
2
known absorption at 330 nm (s = 2.55 ꢁ 10
cm
ꢂ1 21
molecule ). Liquid reactants, CH CHO and CH OH, were
3
3
introduced into the reaction cell using bubblers, by passing a
fraction of the diluent flow through the liquid maintained at
constant temperature (273 K).
initially react with chlorine atoms (k(Cl þ C
6
H
6
) = 1.3 ꢁ
and may form hydro-
6
H OH radicals in the presence of OH:
ꢂ16
3
cm molecule
ꢂ1 ꢂ1
5
1
1
0
s
at 298 k
xycyclohexadienyl C
6
Acetylperoxy and hydroperoxy radicals were formed by
adding an excess of oxygen to ensure rapid and stoichiometric
conversion of radicals formed by reactions (13) and (14) into
peroxy radicals:
C H þ OH þ M - C H OH þ M
(11)
6
6
6
6
Hydroxycyclohexadienyl radical formation has been followed
at 290 nm, a free UV spectral window where peroxy radicals
almost do not absorb.
CH
3
C(O) þ O
2
þ M - CH
3
C(O)O
2
þ M
18
at 298 K.
(3)
3
ꢂ1 ꢂ1
The quantum chemistry calculations were performed using
the density functional theory (DFT) method with the three-
k
3
= 5.0 ꢁ 10ꢂ12 cm molecule
s
CH
2
OH þ O
2
- HO
2
þ CH
2
O
(15)
1
6
parameter hybrid functional B3LYP associated with the
polarized double-zeta 6-31G(d,p) basis set.
k15 = 9.6 ꢁ 10ꢂ12 cm molecule
3
ꢂ1 ꢂ1
s
at 298 K.
18
2
164 | Phys. Chem. Chem. Phys., 2006, 8, 2163–2171
This journal is ꢃc the Owner Societies 2006

View Article Online
1
6
Typical concentration ranges used were: [Cl
ꢂ3
2
] = 2–4 ꢁ 10
maintained at constant temperature (285 K), was maintained
18
ꢂ3
molecules cm
(Messer, 5% in N
1
2
, purity >99.9%),
ꢂ3
at 2 ꢁ 10 molecule cm
.
6
[
CH CHO] = 0.3–0.6 ꢁ 10 molecules cm
(Aldrich,
ꢂ3
No UV absorbing products were generated when gas mix-
3
1
6
9
9.5þ%), [CH OH] = 0.2–1.6 ꢁ 10 molecules cm (Al-
tures containing all reactants except Cl were irradiated sug-
2
3
1
9
ꢂ3
drich, 99.8þ%), [O ] = 2.3 ꢁ 10 molecules cm (Messer,
gesting that the present work is free from complications
associated with the formation of absorbing radical species
from the photolysis of the aldehydes.
2
1
8
ꢂ3
molecules cm (Messer,
9
9
9.995%), [N ] = 1 ꢁ 10
2
9.999%).
The total initial concentration of Cl atoms, and therefore of
peroxy radicals, was determined by (i) replacing acetaldehyde
and methanol under the same experimental conditions by
2.2 Computational details
The quantum chemistry calculations were performed using the
23
GAUSSIAN03 computer program package. Full geometry
methane, producing CH
cross sections of CH
acetaldehyde (or methanol), producing only CH
HO ) radicals, their cross sections being also well-known.
3
O
2
radicals alone (the UV absorption
1
3
3
O
2
being well-known), (ii) keeping only
optimization and force field analysis (vibrational wavenum-
bers) were performed using the density functional theory
3
C(O)O (or
2
1
3
2
(
DFT) method with the three-parameter hybrid functional
16
Under these experimental conditions, initial radical concen-
ꢂ3
B3LYP
associated with the polarized double-zeta 6-
1
3
trations were varied between 4 and 8 ꢁ 10 molecule cm
Considering that the rate constant ratio k /k = (1.35 ꢀ
.
3
1G(d,p) basis set. All electronic structure calculations were
1
4
13
carried out at the UHF level, with the constraint of destroying
the aꢂb and spatial symmetries of the wavefunction in the case
of overall closed shell systems separating into radical species.
Structures associated with the stationary points of the lowest
triplet and singlet (vide infra) potential energy surfaces (PES)
were checked to correspond effectively to minima or saddle
points through the analysis of the eigenvalues of the Hessian
matrix. Minima are characterized by non negative eigenvalues
and saddle points by the existence of only one negative
eigenvalue. This negative eigenvalue is associated with an
imaginary wavenumber in the normal mode analysis. Transi-
tion states of reactive processes are dividing surfaces in the
phase space corresponding to a minimum of the reactive flux.
In the configuration space, the transition states are considered
to be located at the saddle point along the reaction paths.
Minimum energy paths (MEP) determination is requested to
connect the various minima between them through the appro-
priate saddle point. The choice of the DFT-B3LYP/
1
8
0
.05) is well-established,
CH C(O)O and HO radicals can be anticipated controlling
the concentration of precursors:
initial concentrations of
3
2
2
½
CH3CðOÞO2ꢄ k14 ½CH3CHOꢄ
¼
ꢁ
ðIÞ
½
HO2ꢄ
During experiments, this ratio has been varied from 1.5 to
k13
½CH3OHꢄ
0
.375. Particular attention was given to the amount of acet-
aldehyde used during experiments, to minimize the potential
perturbations due to the equilibrium (16), demonstrated by
1
2
our group in a previous work:
CH
3
CHO þ HO
2
þ M # CH
3
CH(OH)O
2
þ M
16,ꢂ16)
(
k2
3
ꢂ1
K2 ¼
¼ 1:9 ꢁ 10ꢂ27 exp½þ6925=Tꢄ cm molecule
kꢂ2
As described above, to accurately know the initial concen-
trations of CH C(O)O and HO radicals, it was first therefore
necessary to know with a very good precision the concentra-
tion of radical precursors (CH CHO and CH OH). In this
6
-31G(d,p) method has been retained since it looks a good
3
2
2
compromise between accuracy and computer time. Moreover,
this level of theory has proven so far to give reliable transition
3
3
2
4–26
state energies.
work, it was demonstrated for the lowest concentrations of
acetaldehyde, that its gaseous concentration in the reaction
cell (controlled by a mass flow controller) was differing by a
factor of 1.7 (lower) compared to the expected value calculated
from that of the nitrogen flow circulating in the acetaldehyde
bubbler, by measuring its UV absorption at 280 nm
3
Results and discussion
In the first part, new numerical simulations of the former
1
experiments performed by Tomas et al. on the CH C(O)O
2
3
2
ꢂ20
2
ꢂ1 22
(
s(CH
3
CHO) = 4.6 ꢁ 10
cm molecule ). This informa-
þ HO reaction, and simulations of the new experiments
2
tion may be crucial when considering the initial radical con-
centration ratio in the simulations, and was attributed to
the non-linearity of the mass flow controller when using very
small flows.
performed in this work using the same technique, were carried
out using the new chemical mechanism (including the third
1
channel (1c)) proposed by Hasson et al., to suggest a new
value for the rate constant k . In the second part, as experi-
1
In order to evaluate the proportion of OH radicals formed
by channel (1c), some benzene was initially added in the gas
mixture. The choice of benzene was justified as it does not
initially react with chlorine atoms (k(Cl þ C H ) = 1.3 ꢁ
mental decays could not be well-simulated including channel
(1c) in the simulation mechanism, the branching ratio k /k
1
1
c
was investigated by adding amounts of benzene in the reaction
system, to measure the potential formation of hydroxycyclo-
hexadienyl radicals by addition of OH radicals on the benzene
ring. In the third part, the branching ratio k /k was inves-
6
6
ꢂ16
3
ꢂ1 ꢂ1
15
1
0
cm molecule
s
at 298 K) and may form hydro-
xycyclohexadienyl C
H OH radicals in the presence of OH:
6 6
1
b
1
tigated by measuring the ozone formation and a final value of
C
H
6 6
þ OH þ M - C
H
6 6
OH þ M
(11)
the rate constant k , accounting for the complete mechanism,
1
The concentration of benzene, introduced into the reaction cell
by passing a fraction of the diluent flow through the liquid
was proposed and compared to the previous recommended
1
3
value. Finally, theoretical calculations were performed to
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 2163–2171 | 2165

View Article Online
give a reliable explanation for the experimental observations
of this work.
6 6
Decay traces were recorded at 290 nm where C H OH radicals
ꢂ18
2
cm molecule ), whereas
ꢂ1 27
absorb strongly (s = 8.91 ꢁ 10
absorbance of all other molecular species and peroxy radicals
is weaker. Initial concentrations of both CH C(O)O and HO
2
3
.1 Impact of the OH formation channel (1c) on the kinetics of
3
2
the CH C(O)O reaction
3
2
þ HO
2
radicals were measured at 207, 240 and 290 nm before adding
benzene. After adding benzene, the signal recorded at 290 nm
is still very weak, as it is corresponding only to the different
The necessity of investigating channel (1c) occurred when
studying the global kinetics of reaction (1) as, in our system,
this third channel leads to the formation of OH radicals which
react with methanol and acetaldehyde to produce additional
negative (Cl
and O ) contributions to the total absorbance, leading finally
to a poor signal-to-noise ratio (Fig. 2). Nevertheless, varying
the value of the branching ratio g (OH) = k1c/k (leading to
the formation of OH radicals) from 0 to 0.40, let us demon-
strate that 0.10 appears as an upper limit for g (OH) at room
temperature and that it was impossible to correctly simulate
experimental traces using g (OH) = 0.40 with the mechanism
2 3 3 2 3 2
, CH CHO) and positive (CH C(O)O , CH O
3
amounts of CH
formation of CH
compose and react with O
3
C(O)O
C(O)O radicals which instantaneously de-
to form CH radicals:
2 2
and HO radicals, and leads to the
3
1
1
2
3 2
O
1
CH C(O)O þ HO - CH C(O)OOH þ O
(1a)
(1b)
(1c)
(5)
3
2
2
3
2
-
CH
3
C(O)OH þ O
3
1
presented Tables 1 and 2 (see Fig. 2).
-
CH C(O)O þ OH þ O
3
2
CH
3
C(O)O þ M - CH
3
þ CO þ M
2
CH þ O þ M - CH O þ M
(6)
3
2
3
2
OH þ CH
3
CHO - CH
3
2
C(O) þ H O
(2)
CH
3
C(O) þ O
OH þ CH
CH
OH þ O
2
þ M - CH
OH - CH
OH þ H
- HO
þ CH
3
C(O)O
2
þ M
O
(3)
3
2
2
(17)
(15)
2
2
2
2
O
As all the quoted peroxy radicals absorb in our analysis
wavelength window, they may all change the shape of the
experimental decays and thus change the extrapolated value of
the rate constant k1.
First, reaction (1) has been studied simulating experimental
traces recorded at 207 and 240 nm (see Fig. 1), with the
mechanism presented in Table 1, taking into account the three
channels (1a) (1b) and (1c) with the different branching ratios
1
proposed by Hasson et al. and using UV absorption cross-
1
3
sections recommended by Tyndall et al. The rate constant k
was measured at 207 nm where both CH C(O)O and HO
1
2
3
2
radicals absorb strongly, resulting in a decay well-adapted to
the measurement of k . Then, the mechanism was verified at
1
2
40 nm, where CH O and O absorb strongly. Ten determi-
3 2 3
nations of k were performed, resulting in the following rate
1
constant at room temperature:
= (2.20 ꢀ 0.07) ꢁ 10ꢂ11 cm molecule
3
ꢂ1 ꢂ1
s .
k
1
Unfortunately, as can be seen in Fig. 1B and 1C, decay traces
recorded at 240 nm could not be well-simulated with this new
value of the rate constant k
the new mechanism of Hasson et al. As a result, to propose a
rate constant k allowing the simulation of all experimental
curves, and particularly those recorded at both 207 and
40 nm, it was first necessary to evaluate the proportion of
OH radicals generated according to channel (1c).
1
, extracted from the simulation of
1
1
2
3
.2 Determination of the branching ratio k /k
1c 1
Channel (1c) was investigated by initially adding benzene in
the gas mixture as it may fast form hydroxycyclohexadienyl
radicals C H OH in the presence of OH radicals.
Fig. 1 Decay traces recorded at 207 nm ([A]: 12.5 ms) and 240 nm
([B]: 12.5 ms and [C]: 500 ms) following the irradiation of Cl
CH CHO/CH OH/O /N gas mixtures at 298 K. Solid lines are results
of simulations using the chemical mechanism detailed in Table 1 with
2
/
3
3
2
2
6
6
ꢂ11
3
ꢂ1 ꢂ1
C
H
6 6
þ OH þ M - C
H
6 6
OH þ M
(11)
g (OH) = 0.40 at k = 2.2 ꢁ 10
cm molecule
s .
1
1
2
166 | Phys. Chem. Chem. Phys., 2006, 8, 2163–2171
This journal is ꢃc the Owner Societies 2006

View Article Online
Table 1 Mechanism used to simulate the flash photolysis of Cl
2
/CH CHO/CH
3
3
OH/N /O
2
2
mixtures at 298 K
a
No.
Reactions
2CH C(O)O
Rate constants
Ref.
ꢂ11
1
5
6
1
1
2
2
2
2
1
1
1
2
2
1
3
1
8
3
2
- 2CH
3
C(O)O þ O
2
k
18 = 1.4 ꢁ 10
13
35
36
13
13
36
18
13
CH
3
C(O)O þ M - CH
þ O þ M - CH
- 2CH
O þ O
OH þ CH O þ O
O þ O - CH
þ HO - H
C(O)O
þ CH
- CH
C(O)OH þ CH
CH C(O)O - CH
þ HO
C(O)OH þ O
C(O)O þ OH þ O
þ HO - CH OOH þ O
CHO þ OH - CH CO þ H
OH þ OH - CH OH þ H
þ M - CH C(O)O
- CH
O þ HO
3
þ CO
2
þ M
Fast thermal decomposition
ꢂ12
CH
3
2
3
O
2
þ M
k
k
k
k
k
k
k
k
k
k
k
k
k
k
k
6
= 1.8 ꢁ 10
ꢂ13
9a
9b
0
2CH
- CH
CH
HO
CH
3
O
2
3
2
19 = 3.5 ꢁ 10
3
2
2
19a/k19 = 0.37
20 = 1.92 ꢁ 10
ꢂ15
3
2
2
O þ HO
þ O
- CH
O þ O
2
ꢂ12
1
2
2
2
O
O
2
2
21 = 3 ꢁ 10
ꢂ11
2a
2b
a
b
c
3
2
3
2
3
C(O)O þ CH
3
O þ O
2
22a = 1.25 ꢁ 10
3
2
2
22b/k22 = 0.90
13
This work
This work
This work
18
18
18
18
18
ꢂ11
3
2
2
3
C(O)OOH þ O
2
1a = 1.50 ꢁ 10
- CH
- CH
3
3
1b/k
1c/k
1
= 0.20
o 0.10
3
2
1
ꢂ12
3
CH
CH
CH
CH
CH
3
3
3
3
2
O
2
2
3
2
23 = 5.2 ꢁ 10
ꢂ11
ꢂ13
3
2
O
2
= 1.49 ꢁ 10
7
5
2
2
O
17 = 9.4 ꢁ 10
ꢂ12
CO þ O
2
3
2
þ M
3
= 5 ꢁ 10
ꢂ12
OH þ O
2
2
2
15 = 9.6 ꢁ 10
a
3
ꢂ1 ꢂ1
s .
Unity: in cm molecule
3
.3 Determination of the rate constant k
1
and the branching
plained before, the rate constant k was measured at 207 nm
2
where both CH C(O)O and HO radicals absorb strongly
1
ratios for the reaction (1)
3
2
whereas the branching ratio leading to the formation of O
3
Considering the upper value of 0.10 for the branching ratio
g1(OH), experimental decays recorded at 207 and 240 nm
without benzene were reconsidered for simulation. As ex-
(
b
1
(O
3
) = k1b/k
1
) was measured at 240 nm on a 500 ms time
absorbs strongly at this wavelength (see Fig.
scale. Indeed, O
3
3
) and at such a time scale, only small amounts of CH
radical (generated by the decomposition of the CH C(O)O
radical arising from the CH C(O)O self-reaction, in the
3 2
O
3
3
2
presence of oxygen) remain and are easily taken into account
using the mechanism presented in Table 1. Ten determinations
of k and b (O ) were performed, resulting in the following
1
1
3
rate constant and branching ratio values at room temperature:
ꢂ11
3 ꢂ1 ꢂ1
cm molecule s
k
1
= (1.50 ꢀ 0.08) ꢁ 10
(O
) = (0.20 ꢀ 0.01)
b
1
3
The branching ratio g (OH) was finally measured at 240 nm
1
fixing the new values of k1 and b (O ) in the model. Ten
1
3
determinations of g (OH) were performed, resulting in the
1
following value at room temperature:
Fig. 2 Decay traces recorded at 290 nm, following the irradiation of
g1(OH) = (0.05 ꢀ 0.04)
CH
are results of simulations using the chemical mechanism detailed in
Tables 1 and 2, varying g (OH) from 0.40 to 0.
3 3 6 6 2 2 2
CHO/CH OH/C H /Cl /O /N gas mixtures at 298 K. Solid lines
This value of g (OH) was validated verifying that both values
of the rate constant k1 and of the branching ratio b (O ),
1
1
1
3
Table 2 Reactions added to the mechanism detailed in Table 1 to simulate experimental decays obtained following the photolysis of the Cl
CH CHO/CH OH/C /N /O mixtures at 298 K
2
/
3
3
6
H
6
2
2
a
No.
Reactions
Rate constants
Ref.
ꢂ12
ꢂ15
1
2
1
4
C
C
6
H
H
6
þ OH þ M - C
–OH þ O - adduct
–OH þ O
–OH - products
–OH þ OH - products
–OH þ HO - products
–OH þ adduct - products
–OH þ CH C(O)O - products
6
H
6
–OH þ M
k
k
k
k
k
k
k
k
k
k
k
k
11 = 1.2 ꢁ 10
24 = 1.3 ꢁ 10
37
38
38
27
27
27
27
39
27
27
39
40
6
6
2
3
ꢂ24
Adduct - C
2C
C
C
C
C
6
H
6
2
ꢂ24 = 5 ꢁ 10
25 = 2.8 ꢁ 10
ꢂ11
ꢂ10
2
2
2
2
2
3
3
3
3
a
5
6
7
8
9
0
1
2
3
6 6
H
6
H
6
H
6
H
6
H
6
6
6
6
26 = 1–5 ꢁ 10
27 = 1–10 ꢁ 10
28 = 1–10 ꢁ 10
29 = 1–10 ꢁ 10
ꢂ11
ꢂ11
ꢂ11
2
3
2
ꢂ12
2 adduct - products
- products
30 = 1–2 ꢁ 10
31 = 5–20 ꢁ 10
32 = 5–20 ꢁ 10
ꢂ12
ꢂ12
Adduct þ HO
2
CH
CH
3
C(O)O
2
þ adduct - products
ꢂ12
3
O
2
þ adduct - products
33 = 1–2 ꢁ 10
3
Unity: in cm molecule
ꢂ1 ꢂ1
ꢂ1
s
excepted kꢂ24 in s
.
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 2163–2171 | 2167

View Article Online
The main sources of systematic errors on the rate constant
1 3 2
k arise from the absorption cross sections of CH C(O)O and
HO , from the extrapolation of the fit to time zero and from
2
the branching ratio g (OH) leading to the formation of OH.
1
Uncertainties associated with s(CH C(O)O ) and s(HO ) are
3
2
2
estimated to be 10% (more variation did not allow to correctly
simulate experimental curves), resulting in approximately 13%
uncertainty in the value of k
the extrapolation of the fit to time zero is essentially due to the
self-reaction rate constant of the CH C(O)O radical (fast
1
. Uncertainties associated with
3
2
initial decay). Uncertainties associated with this rate constant
are estimated to be 30% resulting in approximately 9% of
1
errors in the value of k . Variations of the branching ratio
g1(OH) by a factor of 2 result in a variation of 5% in the value
of k . Combining the uncertainties described above, we esti-
1
mate a global systematic uncertainty of 17% in the value of k1.
This analysis was also conducted for the branching ratio
b
1
(O
ciated with s(CH
0%, resulting in approximately 8% of errors in the value of
(O ). Variations of 10% in the ozone absorption cross
sections s(O ) result in a variation of 10% in the value of
). Variations of the branching ratio g (OH) by a factor of
result in a variation of 5% in the value of b (O ). Combining
3
) leading to the formation of ozone. Uncertainties asso-
3
C(O)O ) and s(HO ) are estimated to be
2
2
1
b
1
3
3
b
1
(O
3
1
2
1
3
the uncertainties described above, we estimate a global sys-
tematic uncertainty of 14% in the value of b (O ).
1
3
3
.5 Theoretical investigation of the branching ratios of the
reaction (1)
The three possible reaction pathways recalled below have been
investigated:
3
ꢂ
g
) (1a)
CH
CH
CH
3
C(O)O
C(O)O
C(O)O
2
þ HO
þ HO
þ HO
2
- CH
- CH
- CH
3
C(O)OOH þ O
2
(X S
1
(X A
3
2
2
3
C(O)OH þ O
3
1
) (1b)
2
3
2
2
3
C(O)O þ OH (X P)
3
ꢂ
g
Fig. 3 Decay traces recorded at 207 nm ([A]: 12.5 ms) and 240 nm
[B]: 12.5 ms and [C]: 500 ms) following the irradiation of Cl
CH CHO/CH OH/O /N gas mixtures at 298 K. Solid lines are results
þ O (X S )
2
(1c)
(
2
/
This reaction between the two radicals CH
3
C(O)O
2
2
and HO ,
3
3
2
2
involves singlet and triplet PES. Spin conservation entails that
reaction (1a) evolves on the lowest triplet surface and reactions
of simulations using the chemical mechanism detailed in Table 1 with
3
ꢂ1 ꢂ1
g
1
(OH) = 0.05 and k
1
= 1.50 ꢁ 10ꢂ11 cm molecule
s .
(
1b) and (1c) on the lowest singlet surface. The determination
of the different stationary points on those PES coupled to
MEP determinations leads to the energy correlation diagram
depicted in Fig. 4. The energies displayed in this diagram
correspond to relative DFT results with zero point vibrational
energies included.
measured, respectively, at 207 and 240 nm, do not significantly
change maintaining the value of g (OH) at 0.05. This latter
value is in good agreement with the previous evaluation of
(OH) carried out with benzene as scavenger of OH radicals,
validating the upper limit measured for this branching ratio
g1(OH) o 0.10).
1
g
1
(
Triplet channel (1a).
3
ꢂ
g
) (1a)
CH
3
C(O)O
2
þ HO
DFT calculations show that this reaction is highly exothermic
2
- CH
3
C(O)OOH þ O
2
(X S
3
.4 Accuracy of results
ꢂ1
ꢂ1
Several factors influence the accuracy of the results obtained in
the flash photolysis experiments. The chemistry associated
with peroxy radical reactions is very complicated (see Table
(DHrxn(298 K) = ꢂ43 kcal mol , (–49.8 kcal mol according
28
to the recommended thermodynamics data and the work of
2
9
Bozelli on peracetic acid). The exploration of this pathway
shows that the first step is the barrierless formation of a pre-
reactive complex PRCT weakly stabilised with respect to the
1
1
). To quantify the sensitivity of the rate constant k and the
branching ratio leading to the formation of ozone of the
ꢂ1
CH C(O)O2 þ HO reaction to the parameters used for
reactants (–5 kcal mol ). This PRCT is characterized by a
3
2
analysis, a systematic analysis of the propagation of errors
7
was performed as described previously.
long-range interaction between the hydrogen atom of HO2
and the terminal oxygen atom of the acetylperoxy group
1
2
168 | Phys. Chem. Chem. Phys., 2006, 8, 2163–2171
This journal is ꢃc the Owner Societies 2006

View Article Online
ꢂ1
Fig. 4 Relative energetic diagram of reaction pathways (1a), (1b) and (1c). The energies E(0 K) in kcal mol include zero-point energy
corrections.
˚
) = 1.88 A) (Fig. 5). This prereactive
CH
3
C(O)O
2
(r(Hꢅ ꢅ ꢅO
3
the reactants. This PRCS is characterized by a rather long
range interaction between the H-atom of HO and the carbo-
complex evolves, via a very small barrier TS1 (only 0.6 kcal
ꢂ1
2
˚
) = 1.89 A) (Fig. 6).
mol above PRCT energy). This path corresponds to the
nyl oxygen of CH
3
C(O)O
2
(r(Hꢅ ꢅ ꢅO
1
ꢂ1
total transfer of the H atom of HO
2
leading to the triplet
ꢂ
g
) and the peracetic acid molecule
After passing through a transition state TS2 (4.4 kcal mol
below the reactants), the calculations show a minimum corre-
3
molecular oxygen O
2
(X S
CH C(O)OOH (Fig. 5).
3
sponding to a cyclic tetroxy adduct ‘‘AD’’ located almost 17
ꢂ1
kcal mol
below the reactants (the main features of the
Singlet channels (1b) and (1c). The exploration of the singlet
geometric structure are described in Fig. 6). It can be seen
˚
) = 1.80 A
˚
in TS2 and 1.78 A in AD). Moreover, it turns out that the
surface shows, first, the presence of a long distance pre-
ꢂ1
that the H-bonding interaction still exists (r(Hꢅ ꢅ ꢅO
1
reactive complex (PRCS) almost located 7 kcal mol below
˚
3
is now equal to 1.50 A, which is a rather large
distance O
2
ꢅ ꢅ ꢅO
˚
value for an O–O distance (usually around 1.40 A). We expect
that this bond is already weakened in this cyclic adduct and
will break more easily.
This cyclic adduct AD can evolve into two different ways:
Channel (1b).
1
(X A
CH
3
C(O)O
2
þ HO
2
-AD - CH
3
C(O)OHþO
3
1
)
(
1b)
The whole exoergicity of reaction (1b) is calculated in this
ꢂ1
work DHrxn(0 K) = ꢂ19.8 kcal mol which corresponds to
an enthalpy at 298 K equal to DHrxn(298 K) = ꢂ21 kcal
ꢂ1
mol . Along this reaction path, the transition state TS3
ꢂ1
(
2 kcal mol above the molecular adduct AD and 14.8 kcal
ꢂ1
mol below the reactants) involves a concerted H-transfer
from HO
2
to the carbonyl O
1
along with an elongation of the
˚
bond (from 1.50 to 1.92 A) (Fig. 6) which will
O
2
ꢅ ꢅ ꢅO
3
eventually break to form the products ozone O and acetic
3
acid CH C(O)OH. In this transition state TS3, the C –O
1
3
1
˚
bond has partly lost its double bond character (1.21 A in AD
˚ ˚
1.26 A in TS3) to the benefit of the C –O bond (1.36 A in
1 2
-
˚
AD - 1.27 A in TS3). It is clear that we have to deal with a
highly feasible concerted process (the energy variation is small)
due to the existence of a hydrogen bond which favours the
hydrogen transfer from one oxygen atom to another.
Channel (1c).
CH
3
C(O)O
2
þ HO
2 3
- AD - CH C(O)O
Fig. 5 Selected geometric structures for the pre-reactive complex
2
3
ꢂ
þ OH (X P) þ O
2
(X S
g
)
(1c)
PRCT and the transition state TS1.
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 2163–2171 | 2169

View Article Online
Fig. 6 Selected geometric parameters for intermediate structures and transition states along singlet reaction pathways (1b) and (1c).
Contrarily to both previous reactions, this one is almost
ꢂ1
Consequences about the branching ratios
Without doing any precise rate constant calculations, these ab
initio results permit to predict that reaction CH C(O)O
HO should evolve mainly according to channels (1a) and
1b). The participation of channel (1c) is highly unlikely.
thermoneutral, DHrxn(298 K) = ꢂ2 kcal mol
(ꢂ3 kcal
ꢂ1
28
using the recommended thermodynamic data). In
mol
3
2
þ
order to explore the path, the bond O –O of the cyclic tetroxy
2
3
2
adduct AD has been extended step by step, while optimizing
all the other geometric variables. The pathway, which does
not present any barrier leads to HO þ CH C(O)O with an
(
Examination of the energetic positions of the two pre-reactive
intermediates gives a slight preference for channel (1a) with
respect to channel (1b). As a matter of fact, PRCS is slightly
3
3
ꢂ1
energy of 11.5 kcal mol above the adduct AD. The existence
of HO
have been performed on this structure.
motivation of these studies is the possible sink behaviour of
3
is still an open question. Several theoretical studies
ꢂ1
more stable than PRCT by almost 2 kcal mol and the exit
3
0–34
The atmospheric
transition states TS1 and TS2 are located at the same level
with respect to the reactants. Then, we can expect that the
recrossing flux towards the reactants coming from PRCS is
larger than the one coming from PRCT. If, ignoring the spin
multiplicity (1 for singlet, 3 for triplet), we attribute a ratio of
2
1
31
HO
3
with respect to the species OH ( P) and O
The most recent calculations lead to a weak stability less
2
g
( D ).
ꢂ134
2
3
ꢂ
than 1 kcal mol
with respect to OH ( P) þ O
2
( S
ꢂ
g
) being not
g
), a
2
3
( S
barrier separating HO
higher than 4 kcal mol
3
from OH ( P) þ O
2
2
5/20 between the rate constants of channel (1a) and channel
ꢂ1 32
.
From the method (UHF-DFT) we
ꢂ1
(
1b), respectively, the three branching ratios with respect to the
used, we found the energy of HO 2.3 kcal mol below the
3
global rate constant of disappearance of reactants are expected
to be roughly in that order: for the triplet channel (1a): 75%
2
3
ꢂ
energy of the products OH ( P) þ O ( S ) and no significant
2
g
ꢂ1
barrier (around 1 kcal mol at the very most) between them.
This situation is still puzzling. As a matter of fact, HO has
(
spin multiplicity times 25%), for the singlet channel (1b): 20%
3
and for the singlet channel (1c): negligible. Of course, we will
check in a future work if our analysis is quantitatively correct.
2
one unpaired electron whereas the products OH ( P) þ O
2
3
ꢂ
(
S
tion of HO
g
) present three unpaired electrons. Since the dissocia-
3
involves a change in the electronic configuration
–OH is elongated, an avoided
when the coordinate O
2
4
Conclusions
crossing should exist between two doublet states. Then, it
would be reasonable to find a barrier along this dissociation
coordinate.
In this study, both experimental and theoretical approaches
have been used to verify the possibility of channel (1c) for the
2
170 | Phys. Chem. Chem. Phys., 2006, 8, 2163–2171
This journal is ꢃc the Owner Societies 2006

View Article Online
reaction of CH
3
C(O)O
2
with HO
2
radicals, yielding a certain
1
16 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
17 P. D. Lightfoot, R. Lesclaux and B. Veyret, J. Phys. Chem., 1990,
4, 700–707.
amount of OH radicals, as proposed recently by Hasson et al.
The values obtained in this work for the rate constant k (k =
9
1
1
1
8 R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson, J. A. Kerr,
M. J. Rossi and J. Troe, J. Phys. Chem. Ref. Data, 1999, 28,
191–393.
ꢂ11
3
cm molecule
ꢂ1 ꢂ1
(
1.50 ꢀ 0.08) ꢁ 10
s at 298 K) and for the
branching ratio b (O ) leading to the formation of ozone
1
3
1
2
2
9 H. Niki, P. D. Maker, C. M. Savage and L. P. Breitenbach, J.
Phys. Chem., 1985, 89, 588–591.
0 M. Bartels, K. Hoyermann and U. Lange, Ber. Bunsen-Ges. Phys.
Chem., 1989, 93, 423.
1 W. B. DeMore, S. P. Sander, D. M. Golden, R. F. Hampson, M. J.
Kurylo, C. J. Howard, A. R. Ravishankara, C. E. Kolb and M. J.
Molina, NASA Conf. Publ., 1997, 97–4.
(
b (O ) = (0.20 ꢀ 0.01)) are in excellent agreement with the
12
1
3
values previously reported by Tomas et al. (k
1
= (1.42 ꢀ
ꢂ11
3
ꢂ1 ꢂ1
s
0
.07) ꢁ 10
and those recommended by the review of Tyndall et al., as
the new value of g (OH) determined in this work (g
(OH) o
.10) have only a minor effect on the determination of k and
(O ). Quantum chemistry calculations confirm the results
cm molecule
1 3
and b (O
) = (0.20 ꢀ 0.02))
1
3
1
1
0
1
2
2 B. J. Finlayson-Pitts and J. N. Pitts Jr, Atmospheric Chemistry,
Wiley and Sons, NY, 1986.
b
1
3
2
3 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N.
Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V.
Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J.
B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J.
J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M.
C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghava-
chari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S.
Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A.
Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M.
W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, J. A.
Pople, Gaussian 03, (Revision C.02), Gaussian, Inc., Wallingford
CT, 2004.
obtained experimentally. Therefore, it is difficult to explain the
difference between our observations and that reported by
1
Hasson et al. even if their system only allows the detection
and quantification of stable final products and not all the
intermediates (CH C(O)O , HO , CH O and C H OH if
3
2
2
3
2
6
6
formed) as detailed in this work. In conclusion, the present
study demonstrates that the CH C(O)O is still pre-
3
2
þ HO
2
dominantly a radical chain termination reaction in the tropo-
spheric ozone chain formation processes.
Acknowledgements
The authors wish to thank R. Lesclaux (University of
Bordeaux) for helpful discussions, and the French National
Programme for Atmospheric Chemistry for financial support.
M.T.R. and J.C.R. wish to thank the IDRIS-CNRS Center
2
4 F. Caralp, P. Devolder, C. Fittschen, N. Gomez, H. Hippler, R.
Mereau, M. T. Rayez, F. Striebel and B. Viskolcz, Phys. Chem.
´
(
Orsay, France) for computer facilities.
Chem. Phys., 1999, 1, 2935–2944.
5 L. Vereecken and J. Peeters, J. Phys. Chem. A, 1999, 103,
2
2
2
1768–1775.
6 R. Mereau, M. T. Rayez, F. Caralp and J. C. Rayez, Phys. Chem.
References
´
Chem. Phys., 2000, 2, 3765–3772.
1
2
3
4
5
6
7
8
9
A. S. Hasson, G. S. Tyndall and J. J. Orlando, J. Phys. Chem. A,
004, 108, 5979–5989.
H. Satsumabayashi, H. Kurita, Y. S. Chang, G. R. Carmichael and
H. Ueda, Atmos. Environ., 1995, 29, 255–266.
E. L. Viskari, M. Vartiainen and P. Pasanen, Atmos. Environ.,
7 D. Johnson, S. Raoult, M. T. Rayez, J. C. Rayez and R. Lesclaux,
Phys. Chem. Chem. Phys., 2002, 4, 4678–4686.
2
28 S. P. Sander, R. R. Friedl, D. M. Golden, M. J. Kurylo, R. E.
Huie, V. L. Orkin, A. R. Ravishankara, C. E. Kolb, M. J. Molina,
G. K. Moortgat and B. J. Finlayson-Pitts, Chemical Kinetics and
Photochemical Data for Use in Atmospheric Studies: Evaluation
No. 11, NASA Jet Propulsion Laboratory Publication 02-25, 2003.
2000, 34, 917–923.
H. J. Beine, D. A. Jaffe, J. A. Hering, J. A. Kelley, T. Krognes and
F. Stordal, J. Atmos. Chem., 1997, 27, 127–153.
A. Borbon, P. Coddeville, N. Locoge and J. C. Galloo, Chemo-
sphere, 2004, 57, 931–942.
I. Villanueva-Fierro, C. J. Popp and R. S. Martin, Atmos. Environ.,
2
9 J. Lee, C. J. Chen and J. W. Bozzelli, J. Phys. Chem. A, 2002, 106,
7155–7170.
30 R. J. Blint and M. D. Newton, J. Chem. Phys., 1973, 59,
6220–6228.
2004, 38, 249–260.
3
3
1 K. Broch Mathisen, O. Gropen, P. N. Skancke and U. Wahlgren,
Acta Chem. Scand. Ser. A, 1983, 37, 817–822.
2 K. Broch Mathisen and E. M. Per Siegbahn, Chem. Phys., 1984,
90, 225–230.
M. Possanzini, V. Di Palo, M. Petricca, R. Fratarcangeli and D.
Brocco, Atmos. Environ., 1996, 30, 3757–3764.
M. F. Mohamed, D. Kang and V. P. Aneja, Chemosphere, 2002,
4
7, 863–882.
G. K. Moortgat, B. Veyret and R. Lesclaux, Chem. Phys. Lett.,
989, 160, 443–447.
0 C. M. Roehl, D. Bauer and G. K. Moortgat, J. Phys. Chem., 1996,
00, 4038–4047.
33 M. A. Vincent and I. H. Hillier, J. Phys. Chem. A, 1995, 99,
3109–3113.
34 A. J. C. Varandas and H. G. Yu, Mol. Phys., 1997, 91, 301–318.
1
1
1
1
1
3
5 P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau, G. D.
Hayman, M. Jenkin, G. K. Moortgat and F. Zabel, Atmos.
Environ., Part A, 1992, 26, 1805–1964.
1
1 M. A. Crawford, T. J. Wallington, J. J. Szente and M. M. Maricq,
J. Phys. Chem. A, 1999, 103, 365–378.
2 A. Tomas, E. Villenave and R. Lesclaux, J. Phys. Chem. A, 2001,
36 R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson, J. A. Kerr,
M. J. Rossi and J. Troe, J. Phys. Chem. Ref. Data, 1997, 26,
521–1011.
105, 3505–3514.
3 G. S. Tyndall, R. A. Cox, C. Granier, R. Lesclaux, G. K.
Moortgat, M. J. Pilling, A. R. Ravishankara and T. J. Wallington,
J. Geophys. Res., 2001, 106, 12157–12182.
3
7 T. H. Lay, J. W. Bozzelli and J. H. Seinfeld, J. Phys. Chem., 1996,
00, 6543–6554.
1
3
8 S. Raoult, M. T. Rayez, J. C. Rayez and R. Lesclaux, Phys. Chem.
Chem. Phys., 2004, 6, 2245–2253.
1
4 M. P. Sulbaek Andersen, C. Stenby, O. J. Nielsen, M. D. Hurley, J.
C. Ball, T. J. Wallington, J. W. Martin, D. A. Ellis and S. A.
Mabury, J. Phys. Chem. A, 2004, 108, 6325–6330.
5 O. Sokolov, M. D. Hurley, T. J. Wallington, E. W. Kaiser, J. Platz,
O. J. Nielsen, F. Berho and R. Lesclaux, J. Phys. Chem. A, 1998,
39 E. Villenave, R. Lesclaux, S. Seefeld and W. R. Stockwell, J.
Geophys. Res., 1998, 103, 25273–25285.
40 E. Villenave and R. Lesclaux, J. Phys. Chem., 1996, 100,
1
1
4372–14382.
102, 10671–10681.
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 2163–2171 | 2171