The Atmospheric Chemistry of the Acetonoxy Radical

Article abstract of DOI:10.1021/jp0026991

The chemistry of the acetonoxy radical, CH₃C(O)CH₂O, formed in the atmospheric degradation of acetone, was studied by a combination of experimental and theoretical methods. In an environmental chamber study conducted over the temperature range 225-298 K, acetonoxy radical chemistry was shown to be dominated by decomposition, CH₃C(O)CH₂O → CH₃C(O) + CH₃O. No evidence was found for a reaction of this species with O₂, CH₃C(O)CH₂O + O₂ → CH₃C(O)CHO + HO₂, even at 225 K in the presence of 1 atm O₂. In a theoretical ab initio and statistical kinetics investigation, the barrier to CH₃C(O)CH₂O decomposition was found to be 6-7 kcal/mol. Using SSE theory and RRKM-based master equation analysis, it was determined that about 80% of the CH₃C(O)CH₂O radicals formed in the CH₃C(O)CH₂O₂ + NO reaction have sufficient energy to decompose promptly under tropospheric conditions. On the basis of TST theory and allowing for falloff, the dissociation rate of thermalized CH₃C(O)CH₂O radicals was found to be on the order of 5 × 10⁷ s^-1 at 1 atm and 300 K and 5 × 10⁵ s^-1 at 0.2 atm and 220 K. The results confirm that the acetonoxy reaction with O₂ is always negligible in the troposphere, consistent with the experimental observations. As part of this study, the rate coefficient for reaction of Cl with acetone (k₂) was measured by a relative rate technique, and a value of k₂ = (3.1 ± 0.5) × 10^-11 exp(-815 ± 150/T) cm³ molecule^-1 s^-1 is reported.

Full text of DOI:10.1021/jp0026991

11578
J. Phys. Chem. A 2000, 104, 11578-11588
The Atmospheric Chemistry of the Acetonoxy Radical
John J. Orlando* and Geoffrey S. Tyndall
Atmospheric Chemistry DiVision, National Center for Atmospheric Research, Boulder, Colorado 80303
Luc Vereecken* and Jozef Peeters
Department of Chemistry, KU LeuVen, Celestijnenlaan 200F, B-3001, HeVerlee, Belgium
ReceiVed: July 27, 2000; In Final Form: October 12, 2000
The chemistry of the acetonoxy radical, CH3C(O)CH2O, formed in the atmospheric degradation of acetone,
was studied by a combination of experimental and theoretical methods. In an environmental chamber study
conducted over the temperature range 225-298 K, acetonoxy radical chemistry was shown to be dominated
by decomposition, CH3C(O)CH2O f CH3C(O) + CH2O. No evidence was found for a reaction of this species
with O2, CH3C(O)CH2O + O2 f CH3C(O)CHO + HO2, even at 225 K in the presence of 1 atm O2. In a
theoretical ab initio and statistical kinetics investigation, the barrier to CH3C(O)CH2O decomposition was
found to be 6-7 kcal/mol. Using SSE theory and RRKM-based master equation analysis, it was determined
that about 80% of the CH3C(O)CH2O radicals formed in the CH3C(O)CH2O2 + NO reaction have sufficient
energy to decompose “promptly” under tropospheric conditions. On the basis of TST theory and allowing for
falloff, the dissociation rate of thermalized CH3C(O)CH2O radicals was found to be on the order of 5 × 107
s-1 at 1 atm and 300 K and 5 × 105 s-1 at 0.2 atm and 220 K. The results confirm that the acetonoxy reaction
with O2 is always negligible in the troposphere, consistent with the experimental observations. As part of this
study, the rate coefficient for reaction of Cl with acetone (k2) was measured by a relative rate technique, and
a value of k2 ) (3.1 ( 0.5) × 10-11 exp(-815 ( 150/T) cm3 molecule-1 s-1 is reported.
Introduction
is formed throughout the entire troposphere. Its subsequent
chemistry appears to be of particular importance in the low-
temperature and moderate-NOx conditions of the UTLS, where
it can affect to some extent the overall HOx yield from acetone.
Acetone is one of the most abundant oxygenated organic
compounds in the remote troposphere, with surface mixing ratios
ranging from 500 to over 2000 pptv.1 In the upper troposphere
and the lower stratosphere, mixing ratios between 100 and 3000
pptv have been measured.2-6 The global source strength of
acetone has been estimated at 40 to 60 Tg/yr, about half of this
originating from the atmospheric oxidation of hydrocarbons such
as propane, isobutane, isobutene, and terpenes.7,8 The estimated
average atmospheric lifetime is on the order of 16 days, the
major sinks being photolysis (estimated at ≈64% globally),
reaction with OH radicals (≈24%), and deposition (≈12%).4,7
In marine environments, especially during polar sunrise events,9
the reaction with Cl atoms can also be of some importance.
As confirmed by direct measurements of OH and HO2,
photolysis of acetone constitutes an important source of HOx
in the dry upper troposphere and lower stratosphere (UTLS)
via reactions of the products CH3 + CH3CO;10,11 recent model
calculations suggest a contribution of 20 to 40% to total HOx
in the dry UT.12 Yet, in a recent determination of the rate
coefficient of (R1) at low temperatures, Wollenhaupt et al. found
that k1 in the range 200 < T < 300 K depends only slightly on
temperature, decreasing from 1.8 × 10-13 cm3 s-1 at T ) 298
K to a minimum of 1.35 × 10-13 cm3 s-1 at 240 K and
increasing again to 1.5 × 10-13 cm3 s-1 at T ) 200 K. They
conclude that, contrary to earlier assumptions, even in the UTLS
(T ≈ 200 K) the OH reaction should still be an important acetone
sink, accounting for about 30% of the total loss.13 These authors
also report14 the formation of CH3 radicals, (R1b), occurring in
parallel with the abstraction channel (R1a). From the above
discussion, it is clear that the acetonyl radical, CH3C(O)CH2,
CH3C(O)CH3 + OH f CH3C(O)CH2 + H2O (1a)
CH3C(O)CH3 + OH f CH3 + CH3COOH
(1b)
The Cl-atom initiated oxidation of acetone has previously
been investigated at 295-298 K in O2-N2 mixtures at pressures
near atmospheric, both in the presence and the absence of
NOx.15,16 These studies have shown that at room temperature
the CH3C(O)CH2O acetonoxy radicals resulting from the
reaction of acetonylperoxy radicals with NO or from their self-
reaction predominantly undergo decomposition, whereas the
reaction with O2 to form methylglyoxal is negligible.
CH3C(O)CH3 + Cl f CH3C(O)CH2 + HCl
(2a)
CH3C(O)CH2 + O2 + MfCH3C(O)CH2O2 + M (2b)
CH3C(O)CH2O2 + NOfCH3C(O)CH2O(*) + NO2 (3)
2 CH3C(O)CH2O2 f 2 CH3C(O)CH2O + O2 (4a)
f CH3C(O)CHO + CH3C(O)CH2OH + O2
CH3C(O)CH2O f CH3C(O) + CH2O
(4b)
(5)
CH3C(O)CH2O + O2 f CH3C(O)CHO + HO2 (6)
On the basis of these findings,15,16 on an estimated rate
* To whom correspondence should be addressed.
coefficient for (R6) of ≈10-14 cm3 molecule-1 s-1, and on an
10.1021/jp0026991 CCC: $19.00 © 2000 American Chemical Society
Published on Web 11/18/2000

Atmospheric Chemistry of the Acetonoxy Radical
J. Phys. Chem. A, Vol. 104, No. 49, 2000 11579
Experimental Section
estimated A-factor for (R5) of ≈1014 s-1, one can estimate an
upper limit of ≈10 kcal/mol for the barrier to decomposition
of acetonoxy. The kth(300 K) rate constant of about 106 s-1
recommended by Atkinson and Carter17 for exothermic oxy
radical decompositions implies a ≈9.5 kcal/mol barrier, con-
sistent with the above. However, a barrier of this magnitude
implies that at T ) 200-220 K in the UTLS the thermal
decomposition should become slower than reaction with O2.
Therefore, mechanistic/kinetic studies of the oxidation of
acetonyl, both in the presence and the absence of NO, have to
be extended to much lower temperatures, down to 200 K.
All experiments were conducted using an environmental
chamber/Fourier transform spectrometer (FTS) system, that has
been described previously.25 The stainless steel chamber is 2
m in length and has a volume of 47 L, and is interfaced to a
Bomem DA3.01 FTS via a set of Hanst-type multipass optics,
which provided an IR observational path length of 32.6 m. FTIR
spectra were measured over the range 800-3900 cm-1 at a
spectral resolution of 1 cm-1 and were typically obtained from
the co-addition of 150-200 scans (≈3-4 min acquisition time).
Experiments were conducted at temperatures ranging from 215
to 298 K. Chilled ethanol was circulated around the cell to
control the gas temperature. Photolysis of the gas mixtures was
conducted using a cw Xe-arc lamp, filtered to provide radiation
between 235 and 400 nm. Minor components of gas mixtures
were added to the chamber from smaller calibrated volumes,
using a flow of N2.
Furthermore, the chemistry of acetonoxy radicals resulting
from (R3) is likely to be strongly affected or even dominated
by a chemical activation effect, particularly at low temperature.
As shown in recent studies,18-23 substituted alkoxy radicals with
a low dissociation barrier are subject to such an effect when
they arise in an RO2 + NO reaction. The energy available from
the exothermic RO2 + NO f ROONO* f RO + NO2 reaction
(exoergicity ≈ 11 kcal/mol for a primary RO2 radical24) is
partitioned over the reaction products such that a fraction of
the oxy radicals is born with sufficient internal energy to
decompose “promptly” (on a sub ns time scale), before
collisional stabilization can occur. The oxy radicals that either
arise with insufficient energy for dissociation or that suffer
collisional energy loss before decomposition will be thermal-
ized and can subsequently undergo either thermal decompo-
sition or react with O2. The effect will most likely apply to
acetonoxy radicals, as their decomposition appears to involve
a low barrier. As a result, in kinetic models the acetonylperoxy
+ NO reaction (R3) should rather be represented as a 2-channel
process:
The rate coefficient for reaction of Cl atoms with acetone
(R2) was determined using standard relative rate techniques,26-30
with (R7) as the reference reaction:
Cl + CH3C(O)CH3 f CH3C(O)CH2+ HCl
Cl + CH2Cl2 f CHCl2 + HCl
(2)
(7)
Measurements were made at temperatures ranging from 215 to
298 K. For these measurements, mixtures of Cl2 (≈1 × 1016
molecule cm-3), acetone (5-10 × 1014 molecule cm-3), and
CH2Cl2 (6-14 × 1014 molecule cm-3) in 1 atm synthetic air
were photolyzed, and the disappearance of acetone and CH2Cl2
was monitored by FTIR spectroscopy. The ratio of the rate
coefficients k2/k7 was determined from the relative rates of decay
using equation (A)
CH3C(O)CH2O2 + NO f CH3C(O)CH2O* + NO2 (3a)
f CH3C(O)CH2O + NO2 (3b)
ln{[acetone]o/[acetone]t} ) k2/k7 ln{[CH2Cl2]o/[CH2Cl2]t}
(A)
where CH3C(O)CH2O* represents the sufficiently “hot” radicals
that decompose promptly and CH3C(O)CH2O represents the
acetonoxy radicals that are thermalized before they react. The
branching ratio for such a process shows some dependence on
pressure as well as on temperature,19-22 yet this T-dependence
is much weaker than that of the rate of thermal dissociation,
such that prompt dissociation may still be important or even
dominant with respect to reaction with O2 at temperatures too
low for thermal dissociation. As an important consequence of
the chemical activation effect, product data obtained for
acetonoxy radicals originating from the peroxy self-reaction (R4)
cannot be transposed to oxy radicals formed in (R3).
where [acetone]o and [CH2Cl2]o are the initial concentrations
before reaction and [acetone]t and [CH2Cl2]t are the concentra-
tions after varying reaction times.
For studies of the acetone oxidation mechanism, mixtures of
Cl2, acetone, NO (in some cases), O2, and N2 were photolyzed.
Experiments were conducted over a range of conditions, as
detailed later. On the basis of previous studies in our lab,
photolysis of acetone oxidation products (e.g., formaldehyde
and methylglyoxal) will be negligibly slow under the conditions
employed. Oxidation products were quantified by means of
spectral subtraction routines, using reference spectra previously
obtained in our laboratory, except in the case of peracetic acid
for which a separate set of calibration experiments was
conducted. The peracetic acid sample was obtained (Aldrich)
as a dilute peracetic acid/acetic acid solution in water. Vapors
obtained from the sample were found to contain about 40
molar% H2O, 35-40% acetic acid (calculated as equivalent
monomer concentration), and 20-25% peracetic acid. Using a
procedure similar to that described by Crawford et al.,30
quantitative peracetic acid spectra were obtained following a
proper accounting for the presence of acetic acid, H2O, and
acetic acid dimers. Acetic acid and water absorption cross
sections used were determined in our laboratory, while the
literature value for the acetic acid dimerization equilibrium
constant, Keq ) 2.5 ( 0.3 Torr-1 was used.30 The peracetic acid
absorption cross section of (5.3 ( 0.7) × 10-19 cm2 molecule-1
at 1295 cm-1 (base e) obtained is almost three times higher
than that reported by Crawford et al.30 The reason for the
This paper reports on experimental and theoretical studies
of the Cl-initiated oxidation of acetone in the temperature range
of 220 to 300 K at near-atmospheric pressures, focusing in
particular on the chemistry of the acetonoxy radical. The first
major aim is to obtain quantitative information on the fate of
the acetonoxy at lower temperatures from detailed oxidation
product measurements under various conditions, both in the
absence and in the presence of NOx. To distinguish between
the methylglyoxal production from the acetonoxy + O2 reaction
(R6) and from the peroxy self-reaction (R4b), product data have
been collected over a wide range of O2 pressures. Furthermore,
to complement the experimental results, a detailed quantum
chemical and theoretical kinetics study of the acetonoxy radical
was carried out to quantify its thermal dissociation as well as
its formation as a chemically activated species in (R3) and its
subsequent prompt decomposition.

11580 J. Phys. Chem. A, Vol. 104, No. 49, 2000
Orlando et al.
TABLE 1: Relative Rate Coefficient Data for Reaction of Cl
with Acetone, Obtained Using the Reaction of Cl with
CH2Cl2 as the Reference Reactiona
measured k2/k7
k7
k2
temp
(this work)
(ref 26)
(this work)
298
267
251
240
225
215
5.39
5.83
6.26
6.69
7.21
7.72
0.374
0.244
0.187
0.153
0.113
0.092
2.02
1.42
1.17
1.03
0.81
0.71
a Absolute rate coefficients are given in units of 10-12 cm3 molecule-1
s-1 and are obtained using k7 ) 1.5 × 10-11 exp(-1100/T) cm3
molecule-1 s-1
.
Figure 1. Plots of the decay of acetone versus CH2Cl2 at 295 K (solid
circles) and 225 K (open circles) in the photolysis of Cl2/acetone/
CH2Cl2/air mixtures at 700 Torr total pressure.
discrepancy is not known, but may be related to an improper
accounting for the presence of the H2O impurity in the Crawford
et al.30 study. We note that this revised peracetic acid cross
section has ramifications for the branching ratios reported by
Crawford et al. in the reaction of CH3C(O)O2 + HO2:
CH3C(O)O2 + HO2 f CH3C(O)OOH + O2
f CH3C(O)OH + O3
(8a)
(8b)
Using the new cross section would lead to a revised value for
k8a/k8 ) 0.72, compared to the ratio k8a/k8 ) 0.88 originally
reported.30 This revised value would be more in line with
previous values of k8a/k8,31-33 which range from 0.67 to 0.75.
Figure 2. Arrhenius plot for k2. Best fit to the data (solid line) is
obtained for k2 ) (3.1 ( 0.5) × 10-11 exp(-815 ( 150/T) cm3
molecule-1 s-1
.
s-1 28
,
a value of k2 ) (2.5 ( 0.2) × 10-12 cm3 molecule-1 s-1
Results and Discussion
is obtained, in reasonable agreement with our work (k2 ) 2.0
( 0.3 × 10-12 cm3 molecule-1 s-1 at 298 K). Christensen et
al.29 have recently published relative rate data for k2 using a
number of reference species and report k2 ) (2.2 ( 0.4) × 10-12
cm3 molecule-1 s-1 at 298 K, also in agreement with our
determination. The 298 K value obtained recently by Notario
et al.34 using a flash photolysis/ resonance fluorescence tech-
nique, k2 ) (3.06 ( 0.38) × 10-12 cm3 molecule-1 s-1, is
significantly higher than the recent relative rate determinations.
Finally, we note that the IUPAC review panel35 recommends a
298 K value of 3.5 × 10-12 cm3 molecule-1 s-1 for k2 based
on the work of Wallington et al.,27 but using an outdated value
for k9.
(2) Cl Atom Initiated Oxidation of Acetone. The primary
goal of this series of oxidation experiments was to study the
competition between reaction of the acetonoxy radical with O2
(R6) and its thermal decomposition (R5), particularly at low
temperature. This was achieved largely via the analysis of yields
of CH2O and methylglyoxal.
(1) Kinetics of Cl Atom Reaction with Acetone. Initial
experiments focused on the determination of the rate coefficent
for (R2),
Cl + CH3C(O)CH3 f CH3C(O)CH2+ HCl
(2)
data which were required for the quantitative interpretation of
the Cl atom initiated acetone oxidation experiments to be
discussed later. Measurements were made relative to (R7),
Cl + CH2Cl2 f CHCl2 + HCl
(7)
a reaction recently studied in our laboratory.26 In that work,
the Arrhenius expression k7 ) 1.5 × 10-11 exp(-1100/T) was
derived, based on an evaluation of our data and previous data.
Sample relative rate data are shown in Figure 1, and all relative
rate determinations and absolute values for k2 are given in Table
1. These k2 data are then plotted in Arrhenius form in Figure 2;
linear least-squares fitting of these data to the Arrhenius equation
yields k2 ) (3.1 ( 0.5) × 10-11 exp(-815 ( 150/T) cm3
molecule-1 s-1, where the uncertainties include 2σ precision
plus possible systematic errors and the uncertainty in the
A-factor reflects the uncertainty in the room-temperature value.
Previous determinations of k2 have been made only at room
temperature. Wallington et al.27 measured k2 relative to the rate
coefficient for reaction of Cl with ethyl chloride:
CH3C(O)CH2O f CH3C(O) + CH2O
(5)
CH3C(O)CH2O + O2 f CH3C(O)CHO + HO2 (6)
Experiments were conducted over a range of conditions, both
in the presence of NOx (298 K only) and in its absence (at
temperatures between 225 and 298 K). For both the NOx-
containing and NOx-free experiments, low (≈1015 molecule
cm-3) and high (≈2 × 1016 molecule cm-3) initial acetone
concentrations were employed. The low acetone experiments
Cl + CH3CH2Cl f CH3CHCl + HCl
(9)
Using the most recent value for k9, 8.6 × 10-12 cm3 molecule-1

Atmospheric Chemistry of the Acetonoxy Radical
J. Phys. Chem. A, Vol. 104, No. 49, 2000 11581
occurrence of (R6) that would still be consistent with the
observations (i.e., the lack of detection of methylglyoxal and
the observed steady-state level of CH2O). The lack of detection
of methylglyoxal does not provide a stringent limit; as many
as 35% of the acetonoxy radicals could be reacting with O2
before the methylglyoxal would reach observable levels. CH2O
model/measurement comparisons, however, do provide a more
stringent test. It is estimated that no more than 20% of the
acetonoxy radicals could be reacting with O2 before modeled
levels of CH2O became significantly lower than those measured.
For the high acetone experiments, initial concentrations were
as follows (in molecule cm-3): Cl2, 13-22 × 1015; acetone,
(1-2) × 1016; NO, 10-20 × 1014; O2, 2.1 × 1019; N2, 1.8-
3.2 × 1018, with small amounts of NO2, (5-7) × 1013, also
present. Models show that only a small portion of the chemistry
(≈10%) is initiated by OH under these conditions. The products
observed in these experiments were the same as those observed
in the low acetone experiments above, namely, CO, PAN, CO2,
and CH2O. Again, no methylglyoxal was observed (<3 × 1013
molecule cm-3). As was the case in the low acetone experiments
just discussed, box modeling studies of these reaction conditions
allow us to place an upper limit of 20% on the fraction of
acetonoxy radicals reacting with O2 in 1 atm O2.
Figure 3. Observed (symbols) and modeled (lines) concentration
profiles for a typical Cl2/“low” acetone/NO/air experiment at 298 K.
Solid circles: measured [acetone]. Open triangles: measured [CO].
Filled triangles: measured [CO2]. Solid squares: measured [PAN].
Open circles: measured [CH2O].
had the advantage of allowing one to follow both the acetone
loss and product growth, though the rapid reaction of Cl with
many of the key products (CH2O and methylglyoxal, in
particular) limited their steady-state levels, often to immeasur-
ably small values. In the high acetone experiments, secondary
reactions consuming the reaction products were minimized, but
only immeasurably small fractional conversions of acetone
occurred, such that mass balance could not be calculated. It was
found that experiments conducted at high initial acetone levels
and in the absence of NOx, similar to those conducted by Jenkin
et al.15 at room temperature, were the most useful in the study
of the competition between (R5) and (R6).
(b) Experiments Conducted in the Absence of NO. The finding
that the decomposition of acetonoxy dominates over its reaction
with O2 at 298 K is consistent with the results of Jenkin et al.,15
who reported k5/(k6 [O2]) > 10 in 1 atm O2 and in the absence
of NOx. Our findings are also consistent with the work of
Sehested et al,37 who observed PAN as a product of the thermal
decomposition of acetonylperoxy nitrate, even at 250 K:
CH3C(O)CH2O2NO2 f CH3C(O)CH2O2 + NO2 (32)
CH3C(O)CH2O2 + NO f CH3C(O)CH2O + NO2 (3)
(a) Experiments Conducted in the Presence of NOx. For the
low acetone experiments, typical initial concentrations were as
follows (in molecule cm-3): Cl2, (8-20) × 1015; acetone,
≈1015; NO, ≈7 × 1014; O2, (0.4-2.2) × 1019; N2, (1.3-2.0) ×
CH3C(O)CH2O (+ O2) f CH3C(O)O2 + CH2O (5)
CH3C(O)O2 + NO2 f CH3C(O)O2NO2 (PAN) (15)
1019. Small amounts of NO2 ((2-3) × 1013 molecule cm-3
)
were typically present as well. Observed products in these
experiments were CO, PAN, CO2, and CH2O, which accounted
for 90 ( 15% of the acetone consumption (on a per carbon
basis). No evidence was found for the presence of acetonyl
nitrate or acetonyl peroxynitrate. Some sample data, showing
acetone loss and product appearance versus photolysis time, are
presented in Figure 3. No methylglyoxal was observed (detection
limit ≈ 1013 molecule cm-3), and hence there is no evidence
for the occurrence of (R6).
However, as will be shown later, the majority of the acetonoxy
dissociation in the presence of NOx occurs as the result of
chemical activation, making direct comparison of the NOx-free
and NOx-containing experiments complicated and hindering our
ability to study the chemistry of the thermalized acetonoxy
radicals. Thus, further experiments (with an emphasis on lower
temperatures) were carried out in the absence of NOx where
chemical activation does not occur.
Box model simulations of the reaction system were carried
out with the Acuchem software package.36 The full model
consisted of about 110 reactions, the most important of which
are given in Table 2 (stable products are shown in bold type).
Note that destruction of acetone, CH2O and CH3C(O)CHO (if
formed) leads to the production of HO2 which in the presence
of NO generates OH. However, modeling studies indicate that
only about 5% of the acetone oxidation will actually be initiated
by OH in these experiments. For initial simulations, it was
assumed that all acetonoxy radicals decompose (either via
chemical activation or following collisional thermalization), and
that reaction with O2 is negligible. The actual decomposition
mechanism is not crucial to the modeling, as even the slower
thermal decomposition is not rate-limiting. Modeled temporal
profiles of acetone and its oxidation products agree very well
(to within (10%) with those measured, as shown in Figure 3.
Further simulations were then run to determine the extent of
Experiments were conducted under conditions of low (≈1015
molecule cm-3) and high (≈1016 molecule cm-3) initial acetone
concentrations, at temperatures ranging from 225 to 298 K. A
typical low acetone experiment involved the photolysis of a
mixture of Cl2 (≈1 × 1016 molecule cm-3) and acetone (≈1 ×
1015 molecule cm-3) in 1 atm synthetic air. At 298 K, the
observed products were CO, CO2, CH2O, CH3C(O)OH,
CH3C(O)OOH, CH3OH, and HCOOH, which accounted for 70
( 15% of the acetone carbon consumed. Methylglyoxal and
hydroxyacetone, expected from the self-reaction of the acetonyl
peroxy radicals, are below the detection limit of (2-3) × 1013
molecule cm-3 but may account for a nonnegligible fraction of
the missing carbon. Furthermore, the hydroperoxides expected
to be formed in these experiments, CH3OOH and CH3C(O)CH2-
OOH, are typically very unstable in our chamber, and their loss
at the cell walls may also account for some of the missing
carbon.

11582 J. Phys. Chem. A, Vol. 104, No. 49, 2000
Orlando et al.
TABLE 2: Major Reactions Used in Simulations of Acetone Oxidationa
reaction
k298
k225
(2)
(10)
(3)
Cl + CH3C(O)CH3
CH3C(O)CH2 + O2
CH3C(O)CH2O2 + NO
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
T
f
f
CH3C(O)CH2 + HCl
CH3C(O)CH2O2
2.1e-12
7.5e-11
8e-12
8.3e-13
7.5e-11
9e-12
CH3C(O)CH2O* + NO2
CH3C(O)CH2O + NO2
CH3C(O) + CH2O
(11)
(5)
(6)
CH3C(O)CH2O*
CH3C(O)CH2O
CH3C(O)CH2O + O2
CH3C(O) + O2
CH3C(O)O2 + NO
CH3O2 + NO (+ O2)
CH3C(O)O2 + NO2
Cl + CH2O
HCO + O2
Cl + CH3C(O)CHO
HO2 + NO
rapid
rapid
variable
1e-14
5e-11
2.6e-11
1e-11
1.3e-11
7.3e-11
6.5e-12
4.8e-11
1.1e-11
1.0e-13
4.5e-12
4.5e-12
1e-11
CH3C(O) + CH2O
variable
9.7e-15
5e-11
1.8e-11
7.7e-12
9e-12
CH3C(O)CHO + HO2
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(4)
CH3C(O)O2
CH3O2 + CO2 + NO2
NO2 + HO2 + CH2O
PAN
Cl + HCO
HO2 + CO
7.3e-11
5.5e-12
4.8e-11
8.1e-12
2.2e-13
6.0e-12
2.0e-12
1.0e-11
1.0e-12
9.0e-12
3.0e-12
9.0e-12
5.6e-12
1.6e-13
3.2e-13
1.2e-11
1.2e-12
1.7e-11
1.1e-12
1.4e-12
1.4e-12
3.0e-12
1e-13,50,10
1.9e-15
3e-4 s-1
CH3C(O) + CO + HCl
OH + NO2
OH + CH3C(O)CH3
CH3C(O)CH2O2 + CH3C(O)CH2O2
CH3C(O)CH2
CH3C(O)CH2O + CH3C(O)CH2O + O2
CH3C(O)CHO + CH3C(O)CH2OH + O2
CH3C(O)CH2O + CH3 + CO2 + O2
CH3C(O)CHO + CH3C(O)OH + O2
CH3C(O)CH2OOH + O2
CH3C(O)OH + O3
(21)
CH3C(O)CH2O2 + CH3C(O)O2
1e-12
(22)
(8)
CH3C(O)CH2O2 + HO2
CH3C(O)O2 + HO2
1.5e-11
1.8e-11
1.8e-11
1.4e-11
0.7e-13
5.1e-13
2e-11
CH3C(O)OOH + O2
CH3OOH
(23)
(24)
CH3O2 + HO2
CH3O2 + CH3O2
CH3O + CH3O + O2
CH3OH + CH2O + O2
CH3O + CH3O2 + CO2 + O2
CH2O + CH3C(O)OH + O2
2 CH3O2 + 2 CO2 + O2
CH3O + CH3C(O)CH2O
CH3OH + CH3C(O)CHO + O2
CH2O + CH3C(O)CH2OH + O2
H2O2 + O2
(25)
CH3O2 + CH3C(O)O2 (+ O2)
2e-12
(26)
(27)
CH3C(O)O2 + CH3C(O)O2 (+ 2 O2)
CH3O2 + CH3C(O)CH2O2
2.7e-11
0.6e-12
2.8e-12
2.8e-12
7.5e-12
(28)
(29)
(30)
(31)
HO2 + HO2
HO2 + HCOOH
CH3O + O2
Cl2 + hν
adduct f HCOOH
CH2O + HO2
Cl + Cl
6.5e-16
3e-4 s-1
a The full model included 110 reactions. Rate coefficients in cm3 molecule-1s-1
.
The chemistry occurring in this NOx-free system is quite
complicated, due to the occurrence of the peroxy radical self-
and cross-reactions.15,16,35,38,39 Under conditions of low acetone,
reactions of the various peroxy radicals with Cl atoms40 and
ClO41 are also expected to occur. Box model simulations of
the chemistry were conducted using the full mechanism, portions
of which are given in Table 2. Rate coefficients were taken
from IUPAC35 and NASA38 evaluations, when available, with
acetonylperoxy radical rate coefficients taken from the work of
Bridier et al.16 Rate coefficients employed for the reaction of
Cl and ClO with acetonylperoxy were based on the correspond-
ing CH3O2 reactions.38,40,41
Initial modeling efforts were conducted under the assumption
that thermal decomposition dominates the chemistry of the
acetonoxy radical. The model was able to reproduce the
observations quite well, especially when the large uncertainty
in the RO2/RO2 and Cl/RO2 reaction rates and mechanisms is
Figure 4. Observed (symbols) and modeled (lines) concentration
profiles for a typical Cl2/low acetone/air experiment at 298 K. Solid
circles: measured [acetone]. Open triangles: measured [CO]. Solid
squares: measured [CO2]. Open circles: measured [CH2O]. Filled
triangles: measured [CH3C(O)OOH].
considered. Figure 4 gives a model/measurement comparison
for one experiment, with only the major products shown. Much
like in the low acetone experiments carried out in the presence
of NOx, the inability to detect methylglyoxal is not a sensitive
diagnostic of the acetonoxy chemistry under these conditions.
Based on a comparison of the modeled and measured CH2O
concentration profiles over a range of conditions, however, a
lower limit k5/k6 > 5 × 1019 molecule cm-3 can be determined,
consistent with (but less informative than) the lower limit
Minor amounts of CH3C(O)Cl were also observed in experi-
ments conducted at low [O2]:
CH3C(O) + Cl2 f CH3C(O)Cl + Cl
(33)
reported by Jenkin et al.,15 k5/k6 > 2 × 1020 molecule cm-3
.
Similar experiments were carried out at 225 K. Products
observed were the same as at room temperature, CO, CO2,
CH2O, CH3C(O)OOH, CH3C(O)OH, CH3OH, and HCOOH.
The carbon-balance in these experiments was similar to room
temperature, 81 ( 15%, with the shortfall again likely due to
the inability to detect methylglyoxal, hydroxyacetone, and the

Atmospheric Chemistry of the Acetonoxy Radical
J. Phys. Chem. A, Vol. 104, No. 49, 2000 11583
hydroperoxide species. In addition, an as yet unidentified
absorption feature at 1290 cm-1 was detected. For low-
temperature modeling, rate coefficients and branching ratios for
the reactions of the acetonylperoxy radical with HO2 and other
peroxy radicals were estimated from available data on the
analogous CH3O2 reactions (see Table 2 for rate coefficients
used). The model/measurement agreement is quite good (quan-
tifiable products agree to within (25%) with one exception:
the model overpredicts the levels of acetic and peracetic acid
by about 50-100%. This may be the result of an unidentified
loss for the acids at low temperature, or it may be due to
incorrect parameterization of acetylperoxy radical chemistry
in the model. Nonetheless, based on the measured levels of
CH2O and the lack of any dependence on [O2], box modeling
studies of these reaction conditions suggest k5/k6 > 2 × 1019
molecule cm-3. Though a more stringent constraint on the k5/
k6 ratio will be determined below, this lower limit is sufficient
to show that decomposition (R5) will be the major fate of
acetonoxy radicals in the upper troposphere.
Experiments with high initial acetone involved the photolysis
of mixtures of Cl2 (7-20 × 1015 molecule cm-3), acetone (15-
25 × 1015 molecule cm-3), O2 (50-650 Torr) and N2 (balance)
at a pressure of 1 atm. These experiments were conducted at
temperatures ranging between 225 and 298 K. Products observed
in all cases were CO, CO2, CH3OH, HCOOH, CH2O, CH3C-
(O)OH, methylglyoxal, acetyl chloride (low [O2] only), and
hydroxyacetone, though the high acetone concentration obscures
various regions of the IR, making the quantification of acetic
acid and the detection of peracetic acid difficult. The unidentified
absorption feature at 1290 cm-1 was also present. To study the
competition between (R5) and (R6) in the most direct fashion
possible, back-to-back experiments were conducted with identi-
cal Cl2 and acetone concentrations, but with the O2 partial
pressure varied from 50 to 650 Torr. In all cases, at all
temperatures, no significant change in any product concentration
was observed with the variation of the O2 pressure (including
methylglyoxal, for which a 25% change would be detectable).
Thus, we have no evidence for a reaction of the acetonoxy
radical with O2 under any conditions, even in 1 atm O2 at 225
K.
Figure 5. Selected geometric parameters for the acetonoxy rotamers
and their transition states for C-C bond rupture. Bond lengths are in
angstroms, angles in degrees.
the conclusion that k5 > 0.9 × 106 s-1. This fact, coupled with
an A-factor of ≈1013 s-1 (typical for alkoxy radical decomposi-
tions),43-45 implies an activation energy for decomposition of
no more than 7.5 kcal/mol. As will be made apparent in the
following sections, this upper limit is entirely consistent with
the barrier height estimated by theoretical methods, 6-7 kcal/
mol.
(3) Quantum Chemical Characterization. Geometry opti-
mizations and vibration frequency calculations for the aceto-
nylperoxy radical rotamers, the acetonylperoxynitrite interme-
diate, the acetonoxy radical rotamers, and the transition states
for C-C bond rupture of the acetonoxy radical were performed
at the B3LYP-DFT level of theory as implemented in the
Gaussian 98 program suite,46 using the 6-31G(d,p) basis set.
This level of theory was already validated in earlier studies20-22
to give barrier heights for dissociation of C3-C4 alkoxy radicals
and the â-hydroxyethoxy radical within 0.5 kcal/mol of the
available direct experimental data, indicating that this level of
theory is reliable for the task at hand.
The acetonylperoxy radicals and acetonoxy radicals have
several degrees of freedom for internal rotation and can exist
in different rotameric forms differing in the O-C-C-O and,
if applicable, the C-C-O-O dihedral angles. As shown in
Figure 5 and Table 3, two rotamers with a corresponding
enantiomer each were found for the acetonoxy radicals, with
O-C-C-O dihedral angles of 150° and 25°, with the relative
energy of the latter rotamer higher by 4.0 kcal/mol. The
geometries, relative energies, and vibrational frequencies of the
rotamers are essentially unchanged when repeating the calcula-
tions using a 6-311++G(2df,2pd) basis set, indicating conver-
gence at the B3LYP-DFT level of theory with respect to the
basis set. For the acetonylperoxy radicals, we found four
To provide a lower limit for the ratio of k5/k6, box model
simulations were carried out. For room temperature, under the
assumption that all acetonoxy radicals decompose, the model
was able to reproduce the concentrations of all quantifiable
species to better than (20%. The model was then run with
various k5/k6 ratios, at high (650 Torr) and low (50 Torr) O2. It
was found that for k5/k6 ) 1.5 × 1020 molecule cm-3, a 25-
30% increase in methylglyoxal and a ≈15% decrease in CH2O
would have occurred between the high and low O2 experiments,
changes that would have been detectable. These experiments
are essentially identical to those carried out at room temperature
by Jenkin et al.,15 who also observed that the temporal profiles
of formaldehyde and methylglyoxal were independent of O2.
Their modeling study suggested a lower limit for k5/k6 of 2 ×
1020 molecule cm-3, very similar to our limit.
The same procedure was used at 225 K, running the model
at various k5/k6 at high and low O2 partial pressures. For k5/k6
) 1 × 1020 molecule cm-3, a 30% increase in methylglyoxal
(and a 15% decrease in CH2O) is expected between the 50 Torr
O2 and 650 Torr O2 experiments, which should have been
observable. This lower limit for k5/k6 allows for an estimate of
the barrier height for acetonoxy decomposition. First, we assume
a value for k6 ) 9 × 10-15 cm3 molecule-1 s-1 at 225 K, by
analogy to reaction of 1-propoxy with O2,35,42 which leads to

11584 J. Phys. Chem. A, Vol. 104, No. 49, 2000
Orlando et al.
TABLE 3: Zero-Point Energy Corrected Relative Energies for the Rotamers of the Acetonoxy Radicals and Their Transition
States for CsC Bond Rupture, the Acetonylperoxy Radicals and the 3-Hydroperoxyacetonyl Radicals at Various Levels of
Theory
energy
hartree
ZPE
kcal/mol
OCCO
angle
CCOO
angle
Erel
kcal/mol
structure
CH3COCH2O
B3LYP-DFT/6-31G(d,p)
trans
TS_trans
cis
-267.6985901
-267.6860115
-267.6915636
-267.6792748
45.3a
43.7a
46.7a
43.7a
150.5°
180.0°
25.3°
0.00
6.37
4.03
TS_cis
16.3°
10.53
B3LYP-DFT/6-311++G(2df,2pd)
trans
TS_trans
cis
-267.7995099
-267.7904611
-267.7926983
-267.7841703
46.8
45.3
46.5
45.2
149.3°
180.0°
28.5°
0.00
4.22
4.00
8.22
TS_cis
14.0°
CCSD(T)/6-31+G(d,p) // B3LYP-DFT/6-311++G(2df,2pd)
trans
TS_trans
-266.9847109
-266.9713680
46.8b
45.3b
0.00
6.91
CH3COCH2OO
B3LYP-DFT/6-31G(d,p)
tp
tt
mp
mt
-342.8616577
-342.8596117
-342.8584381
-342.8566964
48.6a
48.4a
48.5a
48.3a
178.8°
179.7°
-3.9°
-4.4°
79.1°
179.5°
75.8°
0.00
1.06
1.93
2.81
178.8°
CH2COCH2OOH
B3LYP-DFT/6-31G(d,p)
CH2COCH2OOH
H-shift TS
-342.8402339
-342.8129358
47.8a
45.1a
187.8°
187.3°
69.9°
54.2°
12.57
26.89
a Scaled by 0.9614 (ref 47). b Values obtained at B3LYP-DFT/6-311++G(2df,2pd) level.
different rotamers with relative energies and geometries in
agreement with earlier calculations on the â-hydroxyalkoxy
radical.20-22
As discussed in more detail later, the acetonylperoxynitrite
intermediate (ROONO) has a fairly short lifetime, and the
characteristics of this intermediate, as well as those of the
transition states for its dissociation to NO + RO2 and NO2 +
RO, have only a small effect on the RO decomposition results
obtained in the next section. We therefore chose not to perform
a full microvariational treatment of the barrierless ROONO
dissociation channels, and instead approximated these transition
states by a constrained optimization at a bond length of 2.4 Å
for the breaking bond. The lifetime of the acetonylperoxy-
nitrite intermediate is sufficiently long to ensure a steady state
distribution between its different internal rotamers, making it
unnecessary to explicitly take the different rotamers of the
acetonylperoxy radicals and of the acetonylperoxynitrite inter-
mediate into account.
Finally, we also examined the possibility of a 1,5-H shift of
a methyl-hydrogen in the acetonylperoxy radicals, forming
3-hydroperoxyacetonyl radicals (CH2C(O)CH2OOH, “QOOH”)
which might react subsequently with O2 to form a new peroxy
radical. At the B3LYP-DFT/6-31G(d,p) level of theory, this
rearrangement, for the most stable rotamer, is found to be
endoergic by 12.5 kcal/mol and to face an energy barrier of
26.9 kcal/mol. According to these results, the thermal isomer-
ization is expected to be negligibly slow under all atmospheric
conditions (rate < 10-7 s-1) compared to the usual peroxy
radical reactions with NO, HO2 etc. Yet, one also needs to
consider the “prompt” isomerization of the nascent peroxy
radicals, which contain an excess energy of some 30 to 35 kcal/
mol,24 and thus lie above the isomerization barrier. An RRKM
estimate puts the rate of the prompt isomerization RO2* f
QOOH* at the initial energy content at only ≈106 s-1, which
is much slower than the rate of collisional stabilization of ≈109
s-1. Therefore, isomerization to QOOH should be negligible.
(4) Kinetics of “Prompt” and Thermal Decomposition of
Acetonoxy Radicals: Theoretical Quantification. The first
part of this section describes the theoretical quantification of
the fraction of acetonoxy radicals formed in (R3) that dissociate
The kinetic calculations described later are most sensitive to
the characteristics of the transition state (TS) for â CsC bond
rupture of the acetonoxy radicals, shown in Figure 5 and listed
in Table 3. Again, the B3LYP-DFT/6-31G(d,p) barrier height,
6.4 kcal/mol, is in keeping with the experimentally derived upper
limit of 7.5 kcal/mol (see earlier). As was already found for
other â-substituted alkoxy radicals, the different rotamers have
comparable transition state barriers to dissociation. The opti-
mizations of the transition states were repeated using the
6-311++G(2df,2pd) basis set, but it is known that at this level
of theory one tends to find too low a barrier height for
(substituted) alkoxy radicals compared to the experimental
data,48 such that the barrier height derived at that level of theory,
4.2 kcal/mol, should be considered a lower limit. The reason
for this trend is unknown; geometries and vibrational frequencies
are not affected. We also performed single point CCSD(T)/6-
31+G(d,p) calculations on the DFT/6-311++G(2df,2pd) ge-
ometries, finding a barrier height of 6.9 kcal/mol, in fair
agreement with the DFT/6-31G(d,p) result. Because the basis
set used is somewhat small for use in CCSD(T) calculations,
we consider this value to be an upper limit estimate for the
barrier height. We also attempted to perform G2 and G3
calculations, but found that the MP2 geometries are too sensitive
to the basis set used: e.g., for the lowest acetonoxy rotamer
we find an O-C-C-O angle of 179.9° for the 6-31G(d) basis
set, 165° for 6-31G(d,p), and 161° for 6-31G(2df,p). Obviously,
this invalidates the use of single-point calculations at varying
levels of theory and basis sets based on any MP2 geometry, as
is done in the G2 and G3 methods. The geometry found using
MP2 tends toward the B3LYP-DFT geometry with increasing
basis set size, indicating that the DFT geometries are good
estimates. The difference in basis set superposition error (BSSE)
between minima and their transition states is usually small; we
therefore did not make explicit correction for BSSE.

Atmospheric Chemistry of the Acetonoxy Radical
J. Phys. Chem. A, Vol. 104, No. 49, 2000 11585
There is no barrier in the exit channel for this process.51 The
various quantities involved are not critical because collisional
energy losses are small here; the vibrational frequencies of the
exit TS were again obtained in a quantum chemical character-
ization at a breaking bond length of 2.4 Å. The ROONO f
RO + NO2 dissociation energy Ed was taken equal to 14 kcal/
mol as for the analogous 2-butyl case;51 for the overall
exoergicity ∆E3(0 K) of the RO2 + NO f RO + NO2 reaction,
a value of 10.5 kcal/mol was adopted, between the values for
the analogous ethyl- and isopropylperoxy reactions.24 Collisional
energy transfer was described in the ME equation by Troe’s
biexponential model,52 with an average energy transferred per
collision ∆Etot of -150 cm-1. The RRKM-ME analysis
revealed collisional energy losses before dissociation to RO +
NO2 to be negligible. The energy distribution at dissociation,
Figure 6. Reaction scheme for acetonoxy radicals formed in the
reaction of acetonylperoxy radicals with NO. Energies are in kcal/mol,
with -∆E3 ) 10.5 kcal/mol, Ed ) 14 kcal/mol, Eb ) 6.9 kcal/mol
(upper limit). Fform(E) is the nascent energy distribution for the
acetonylperoxynitrite radicals, F(Edisp) is their energy distribution at
dissociation. Pform(EA) is the nascent energy distribution for the
acetonoxy radicals formed in (R3).
Fdiss(E), resulting from the ME analysis is also depicted in Figure
6, for T ) 300 K and 1 atm pressure. The average ROONO*
lifetime for reactants at 300 K is about 10-10 s, long enough to
ensure complete statistical distribution of the total internal energy
over all internal modes. Formation of organic nitrate appears
to be negligible here and need not be considered.
“promptly”, i.e., before collisional stabilization. The quantum
statistics based methodologies, as implemented in our
URESAM-3 program suite,49 have been described at length in
earlier work, where they were applied to similar processes of
â-hydroxyethoxy and â-hydroxypropoxy radicals formed in the
oxidation of ethene and propene, respectively, resulting in
theoretical predictions in excellent agreement with the observa-
tions.20-22 Here, only the outlines of the methodologies will be
given, together with the results.
The sequence of events leading to prompt dissociation of the
oxy radicals is illustrated schematically by the potential energy
and energy distribution diagram of Figure 6. Four steps can be
discerned in this sequence: (i) reaction of CH3C(O)CH2O2 with
NO to form an activated peroxynitrite ROONO* with an internal
energy distribution at formation, Fform(Eth), inherited from the
thermal reactants; (ii) collisional energy loss of the ROONO*
prior to its dissociation into RO + NO2, which modifies its
internal energy distribution to Fdiss(E) at dissociation; (iii)
dissociation of the ROONO* and partitioning of the disposable
excess energy, Edisp, over the two fragments and over the degrees
of freedom of their relative motion, resulting in a given internal
energy distribution function Pform(EA) of the RO fragment; (iv)
prompt unimolecular dissociation of the resulting chemically
activated oxy radicals, in competition with their collisional
thermalization. Each of the steps above was treated theoretically
and quantified, as briefly discussed below.
(iii) When the ROONO* dissociates, the disposable excess
internal energy Edisp above the dissociation limit is imparted to
the separating fragments. The Edisp can be seen as the overall
exoergicity -∆E3(0 K) ) 10.5 kcal/mol, increased by the
thermal energy of formation of the peroxynitrite and decreased
by the collisional energy losses during the ROONO* lifetime.
The distribution F(Edisp) is identical to the Fdiss(E) function
derived above (sub ii), but shifted downward by the ROONO
f RO + NO2 dissociation energy Ed of 14 kcal/mol. As shown
earlier,20-22 the contribution of rotational energy of the initial
ROONO* to the internal energy available to the reaction
products can be neglected. As the lifetime of the ROONO* is
long enough (≈10-10 s) for a statistical distribution of its internal
energy over all its internal degrees of freedom, the excess energy
Edisp should likewise be partitioned fully statistically over the
separating fragments and their degrees of freedom of relative
motion (i.e., the “transition” modes). The energy partitioning
was quantified using the extended separate statistical ensemble
(SSE) theory,53 suitable for a dissociation process through a
barrierless exit channel and involving a late and loose transition
state. The theory and its extension as well as the conditions for
its applicability have already been discussed in depth earlier.20-22
In essence, SSE theory evaluates the partitioning of Edisp by
adopting equal probabilities for each discrete internal quantum
state of the (nearly) dissociated molecule. In this way, the
available energy is distributed over the acetonoxy radical, the
NO2 fragment and the six degrees of freedom (d.f.) of relative
motion in accordance with the respective relative increases of
the densities of states N(E) with energy. The nonnormalized
(i) The thermal internal energy distribution Fform(Eth) of
formation of the chemically activated peroxynitrite CH3C(O)-
CH2OONO* from CH3C(O)CH2O2 + NO was derived from
the sum of states G*(E - E0) of the (variational) entrance
transition state,50 based on the vibration frequencies obtained
in a quantum chemical characterization at a separation of 2.4
Å, as discussed earlier. The form of the distribution function
probability PEA (EA) of forming an acetonoxy fragment A in
disp
i
Fform(Eth) is shown in Figure 6. The thermal energy distribution,
rotameric form i with energy EA (the energy above the zero
point level of the most stable rotamer) for a given total energy
Edisp available to the products can be expressed as:
with an average value Eth of 4.9 kcal/mol at 300 K, is
superimposed on the chemical activation energy, i.e., the well
depth with respect to the reactants.
(ii) The internal energy distribution function of the ROONO*
at dissociation, Fdiss(E), as modified by collisions during its
lifetime, was obtained by an RRKM-based master equation
(ME) analysis of the dissociation of ROONO* in competition
with collisional energy transfer to the bath gas. The ROONO*
dissociation rate and hence also the collisional energy losses
depend on the internal energy E (see sub i), on the energy Ed
of dissociation to RO + NO2, and on the vibrational charac-
teristics of the (variational) transition state (TS) for dissociation.
Etot-EA
PEA (E ) ) N (E ) ×
[NNO (ENO ) ×
disp
∫
A
Ai
A
0
i
2
2
Nrel.mot.(Etot - EA - E )]dE
NO2
NO2
Normalization requires dividing by the sum of this expression
over all acetonoxy rotamers integrated over all energies EA from
0 to Edisp. For the acetonoxy and NO2 fragments the N(E) were
obtained by exact count using the DFT rovibrational data,

11586 J. Phys. Chem. A, Vol. 104, No. 49, 2000
Orlando et al.
The fate of the thermalized acetonoxy radicals, formed either
by collisional deactivation of hot acetonoxy radicals from the
reaction of acetonylperoxy radicals with NO, or formed in the
self-reaction of RO2 radicals, depends on the competition
between thermal dissociation (R5) and reaction with O2 (R6).
The rate of thermal dissociation can be estimated using transition
state theory (TST), Troe’s low-pressure theory, and Troe’s falloff
formalism which derives52,54,55 a temperature- and pressure-
dependent rate constant from the high- and low-pressure rate
coefficients. In thermal conditions, only the lowest rotamer
contributes significantly, with the higher energy rotamer ac-
counting for only 0.16% of the total population even at 300 K.
The TST expression for thermal high-pressure rate constants is
given by
Figure 7. Predicted nascent vibrational energy distribution Pform(EA)
for the acetonoxy radicals formed in (R3) at 300 K (energy grain size
) 0.239 kcal/mol). Predicted probability f(EA) of prompt dissociation
at 300 K and 1 atm.
*
*
*
kT Q*
‚
kT
h
Qrot Qvib Qel
QrotQvibQel
‚e-E /RT
)
‚
‚e-E /RT
)
0
0
k∞ )
h Q
A∞(T)‚e-E /RT
0
TABLE 4: Predicted Prompt Dissociation Fraction for the
Acetonoxy Radicals Formed in (R3), at Different
Temperatures and Pressures, Calculated Using a Barrier
Height to Dissociation of 6.9 Kcal/Mola
where the barrier height E0 ()Eb in Figure 6) and partition
functions are derived from the DFT quantum chemical calcula-
tions. At 300 K, a partition function ratio Q*/Q of 2.927 was
found, yielding a theoretical preexponential factor A∞(300 K)
of 1.8 × 1013 s-1. Combined with a barrier height of 6.9 kcal/
mol, our upper limit estimate, this leads to a high-pressure rate
coefficient k∞(300 K) ) 1.7 × 108 s-1. At 220 K, we find Q*/Q
) 2.413, A∞(220 K) ) 1.1 × 1013 s-1, and finally k∞(220 K) )
1.5 × 106 s-1. These k∞ (T) values can be expressed by a formal
P(atm)
200 K
220 K
250 K
275 K
300 K
5.00
2.00
1.00
0.75
0.50
0.20
0.10
0.515597 0.571346 0.648050 0.704735 0.754530
0.615503 0.664443 0.728942 0.774795 0.814022
0.675785 0.717619 0.772054 0.810475 0.843277
0.695680 0.734781 0.785661 0.821616 0.852360
0.718969 0.754744 0.801423 0.834516 0.862887
0.755734 0.786230 0.826338 0.854971 0.879636
0.773444 0.801417 0.838374 0.864858 0.887729
Arrhenius expression: k∞ ) 6.72 × 1013 exp(-3870/T) s-1
.
a Figures in bold indicate the approximate temperature-pressure
relationship as a function of altitude in the troposphere.
The low-pressure limit is calculated using the Troe equation:
N(Eo)e-E /kT
0
whereas the six d.f. of relative motion were treated as unhindered
ko ) kT‚ZLJ[M]â
FEFanhFrotFint.rot
particle motions with combined Nrel.mot.(E) accordingly varying
Qvib
as E(6/2)-1
.
The probability PA (EA) of formation of acetonoxy rotamer i
The Lennard-Jones collision frequency is based on estimated56,57
collision parameters for the oxy radicals: ꢀA-A ≈ 400 K and
σA ) 5 Å. At 300 K, we find ZLJ[M] ) 1.0 × 1010 s-1 for N2
as bath gas. The adopted ∆Etot of -130 cm-1 can be converted
into a collision efficiency â ) 0.22. The vibrational partition
function Qvib ) 12.8 is calculated from the DFT rovibrational
data; the vibrational density of states N(E0) equals 1.2 × 104
states per kcal/mol. The correction factors are all derived from
the formulas given by Troe:52,54,55 FE ) 1.48; from the moments
of inertia of the minimum and the transition state, I*/I ) 1.22,
i
with internal energy EA is obtained by integrating the expression
above over the Edisp distribution F(Edisp); summing over all
rotamers i gives the probability Pform(EA) of formation of
acetonoxy radicals with internal energy EA. Figures 6 and 7
show the combined Pform(EA) of the acetonoxy radicals for initial
reactants at 300 K.
(iv) The resulting acetonoxy radical will either decompose
promptly or will be collisionally thermalized. The fraction of
prompt dissociation was determined by RRKM-ME analysis,
with the PA (EA) obtained above as initial internal energy
we derive Frot ) 1.17; and the anharmonicity correction Fanh )
i
distribution, using RRKM microcanonical dissociation rates for
the oxy rotamers based on the B3LYP-DFT/6-31G(d,p) relative
energy and rovibrational data of the transition states and
equilibrium structures, and adopting an average energy trans-
ferred per collision ∆Etot of -130 cm-1. Isomerization is
sufficiently fast to maintain microcanonical equilibrium between
the various rotamers during dissociation and thermalization.
Figure 7 shows the prompt decomposition probability f(EA) of
oxy radicals born with internal energy EA, for reaction conditions
of 300 K and 1 atm. The overall fraction of oxy radicals
decomposing promptly is obtained by integrating the product
f(EA) × Pform(EA) over all energies. For atmospheric conditions,
we predict a nearly constant prompt dissociation fraction of
about 80% (see Table 4), indicating that the large majority of
the acetonoxy radicals formed in (R3) immediately decompose
to CH3CO and CH2O. A typical population-averaged rate
coefficient for prompt decomposition is ≈1.3 × 1010 s-1. The
remainder of the acetonoxy radicals, about 20%, are collisionally
thermalized.
1.14. Acetonoxy radicals have 2 degrees of freedom of internal
rotation; treating each rotor as an oscillator with an estimated
anharmonicity constant of 1.75 we find a correction factor for
internal rotation Fint.rot ) 3.06. Thus, at 1 atm and 300 K we
find a first-order low-pressure rate coefficient of k0(300 K) )
7.1 × 107 s-1. At 0.2 atm and 220 K, the parameter values are:
ZLJ[M] ) 2.6 × 109 s-1, â ) 0.30, Qvib ) 5.41, FE ) 1.31, and
Frot ) 1.19; this leads to a first-order low-pressure rate constant
of k0(220 K) ) 5.6 × 105 s-1. Then, on the basis of the rate
coefficients k0 and k∞, and an estimated broadening factor Fcent
) 0.5, we finally obtain a thermal dissociation rate coefficient
k
diss, listed in Table 5 for different conditions of temperature
and pressure.
The yield of CH3CO + CH2O from thermalized acetonoxy
radicals can then be determined by examining the competition
between thermal dissociation and reaction with O2. Rate
coefficients are not known for the latter reaction, but can be
estimated from the analogous 1-propoxy radical reaction: kO
2
) 1.4 × 10-14 exp(-110/T) for T ) 220-310 K.35 Table 6

Atmospheric Chemistry of the Acetonoxy Radical
J. Phys. Chem. A, Vol. 104, No. 49, 2000 11587
TABLE 5: Rate Constants for Thermal Dissociation of the
Acetonoxy Radicals, in the High-pressure Limit and the
Falloff Region, Calculated for a Barrier Height to
Dissociation of 6.9 Kcal/Mol
6.9 kcal/mol is in keeping with the upper limit of 7.5 kcal/mol
derived from the experimental data.
Atmospheric Implications
P(atm)
200 K
220 K
250 K
275 K
300 K
Both the experimental and theoretical results presented above
confirm that the rate of dissociation of the acetonoxy radical
dominates the rate of its reaction with O2 at 298 K, as has
previously been measured15,37,58 or estimated from thermody-
namic considerations.17,59 Furthermore, we clearly show that this
conclusion also applies for essentially all conditions encountered
in the troposphere, even the coldest regions of the upper
troposphere. In many regions of the atmosphere, where peroxy
radical chemistry is dominated by reaction with NO (e.g., the
continental boundary layer and the upper troposphere), this
decomposition will in fact occur largely via chemical activation.
The energetics involved in the acetonoxy dissociation are similar
to those found recently in our laboratories for the 2-hydroxy-
1-propoxy species22 (i.e., a near-thermoneutral dissociation
process, a barrier height to decomposition of 6-7 kcal/mol, and
a large propensity for prompt, chemically activated dissociation).
The occurrence of acetonoxy radical decomposition reaction,
instead of the now ruled-out alternative reaction with O2, has
some minor impacts on both the CH2O budget and on the HOx
(dOH + HO2 + RO2) production rate in the upper troposphere.
Reaction with O2, followed by the rapid photolysis of methyl-
glyoxal (lifetime ≈ 1 h at 12 km60), would lead to the formation
of CO, acetylperoxy radicals and HO2, with a net production
of HOx radicals,
∞
2.66E+5 1.51E+6 1.24E+7 5.14E+7 1.69E+8
1.83E+5 9.53E+5 6.59E+6 2.28E+7 6.12E+7
1.44E+5 7.04E+5 4.47E+6 1.47E+7 3.88E+7
1.09E+5 5.14E+5 3.17E+6 1.04E+7 2.73E+7
9.55E+4 4.45E+5 2.74E+6 8.99E+6 2.33E+7
7.81E+4 3.63E+5 2.24E+6 7.23E+6 1.83E+7
4.94E+4 2.27E+5 1.33E+6 4.03E+6 9.50E+6
3.39E+4 1.50E+5 8.21E+5 2.36E+6 5.37E+6
5.00
2.00
1.00
0.75
0.50
0.20
0.10
TABLE 6: Thermal Dissociation Fraction for the Acetonoxy
Radicals at Different Pressures and Temperatures,
Calculated Using a Barrier Height to Dissociation of 6.9
Kcal/Mol, Assuming kO2 ) k1-propoxy+O2 (see text), and 20 %
a
O2
P(atm)
200 K
220 K
250 K
275 K
300 K
5.00
2.00
1.00
0.75
0.50
0.20
0.10
0.382116 0.770951 0.961397 0.989117 0.996136
0.548344 0.861477 0.976860 0.993233 0.997558
0.648280 0.900671 0.983556 0.995206 0.998263
0.682313 0.912880 0.985732 0.995837 0.998474
0.725006 0.927687 0.988306 0.996548 0.998702
0.806431 0.952567 0.992111 0.997519 0.999002
0.851298 0.963664 0.993596 0.997886 0.999116
a Figures in bold indicate the approximate temperature-pressure
relationship as a function of altitude in the troposphere.
TABLE 7: Total Dissociation Fraction at Different
Temperatures and Pressures for Acetonoxy Radicals Formed
in (R3), Combining Prompt and Thermal Decomposition,
Determined Using a Barrier Height to Dissociation of 6.9
Kcal/Mol, Assuming kO2 ) k1-propoxy+O2 (see text), and 20 %
O2
CH3C(O)CH2O + O2 f CH3C(O)CHO + HO2 (6)
CH3C(O)CHO + hν (+ 2 O2) f CH3C(O)O2 + CO + HO2
Net:CH3C(O)CH2O + 3 O2 f CH3C(O)O2 + CO + 2 HO2
P(atm)
200 K
220 K
250 K
275 K
300 K
5.00
2.00
1.00
0.75
0.50
0.20
0.10
0.70070
0.82634
0.88597
0.90332
0.92272
0.95272
0.96631
0.90182
0.95352
0.97195
0.97689
0.98226
0.98986
0.99278
0.98641
0.99373
0.99625
0.99694
0.99768
0.99863
0.99896
0.99679
0.99848
0.99909
0.99926
0.99943
0.99964
0.99971
0.99905
0.99955
0.99973
0.99977
0.99982
0.99988
0.99990
while decomposition leads, of course, to the generation of CH2O:
CH3C(O)CH2O (+ O2) f CH3C(O)O2 + CH2O (5)
For conditions typical of the upper troposphere,10 this mecha-
nism accounts for about 10% of the total CH2O production when
acetone levels of 1-2 ppb are encountered (the majority of
CH2O production is via methane oxidation). The formation of
CH2O versus methylglyoxal also acts to suppress HOx produc-
tion, since the majority of formaldehyde destruction occurs via
non radical-generating processes (mainly photolysis to CO and
H2). However, these effects are likely to be fairly minor (about
10%), given the large rate of HOx production from the photolysis
of acetone itself.
a Figures in bold indicate the approximate temperature-pressure
relationship as a function of altitude in the troposphere.
lists the dissociation fraction for thermalized acetonoxy radicals
in different conditions, with 20% O2. It is clearly seen that
reaction with O2 is of minimal importance throughout the
troposphere, in complete agreement with the experimental
findings. The dissociation fractions listed should be considered
a lower limit, since they are based on our upper limit estimate
of 6.9 kcal/mol for the dissociation barrier height. For high-
NO conditions, Table 7 lists the total dissociation fraction by
the prompt and thermal decomposition combined; throughout
the troposphere, less than 1% of the acetonoxy radicals from
acetone are expected to react with O2 to form methylglyoxal.
No reaction with O2 was observed in the experiments, even
Conclusions
The Cl-atom-initiated oxidation of acetone was studied in an
environmental chamber over a range of temperatures, both in
the presence and absence of NOx. No evidence for a reaction
of acetonoxy with O2 was found, and an upper limit to the
activation energy for decomposition of the acetonoxy radical,
(R5), of 7.5 kcal/mol was obtained. In a theoretical ab initio
and statistical kinetics investigation of the acetonoxy decom-
position, it was found that the barrier is indeed very low, only
6-7 kcal/mol. Using appropriate and validated methodologies,
the chemical activation of acetonoxy radicals from the aceto-
nylperoxy + NO reaction was evaluated, from which it was
determined that a large majority of these radicals decompose
promptly, under all tropospheric conditions. The dissociation
of thermalized acetonoxy radicals was also quantified theoreti-
in the absence of NO at 225 K and PO ) 650 Torr (Ptot ) 700
2
Torr). To verify our estimate for the dissociation barrier height,
we redetermined the dissociation fraction theoretically for these
conditions. Using a barrier height of 6.9 kcal/mol, our upper
limit estimate, we predict 33% formation of CH3C(O)CHO
under the aforementioned conditions, which should still have
been detectable. Lowering the barrier height to 6.5 kcal/mol
reduces the calculated methylglyoxal yield to about 15%, near
the detection limit. These results support our DFT barrier height
value of 6.4 kcal/mol. Likewise, our CCSD(T) upper limit of

11588 J. Phys. Chem. A, Vol. 104, No. 49, 2000
Orlando et al.
(28) Shi, J.; Wallington, T. J.; Kaiser, E. W. J. Phys. Chem. 1993, 97,
6184.
(29) Christensen, L. K.; Ball, J. C.; Wallington T. J. J. Phys. Chem. A
2000, 104, 345.
(30) Crawford, M. A.; Wallington, T. J.; Szente, J. J.; Maricq, M. M.;
Francisco, J. S. J. Phys. Chem. A 1999, 103, 365.
(31) Niki, H.; Maker, P. D.; Savage, C. M.; Breitenbach, L. P. J. Phys.
Chem. 1985, 89, 588.
(32) Moortgat, G. K.; Veyret, B.; Lesclaux, R. Chem. Phys. Lett. 1989,
160, 443.
(33) Horie, O.; Moortgat, G. K. J. Chem. Soc., Faraday Trans 1992,
88, 3305.
cally, and was likewise found to be very fast. On the basis of
these results, it can be concluded that the chemistry of the
acetonoxy radical throughout the troposphere will be dominated
by decomposition. In regions where acetonylperoxy radical
chemistry is dominated by its reaction with NO (e.g., the
continental boundary layer, and the upper troposphere), the
internal excitation provided to the acetonoxy species will result
in the majority of the decomposition occurring via chemical
activation.
Acknowledgment. The National Center for Atmospheric
Research is operated by the University Corporation for Atmo-
spheric Research under the sponsorship of the NSF. A portion
of this work was supported by the NASA Upper Atmosphere
Research Program. The research carried out at Leuven was
financially supported by the FWO (Fund for Scientific Research,
Flemish Community). L.V. acknowledges the Research Council
of the University of Leuven for his BOF Postdoctoral Mandate.
Thanks are due to Alex Guenther and Samuel Hall of NCAR
for helpful comments on the manuscript.
(34) Notario, A.; Mellouki, A.; Le Bras, G. Int. J. Chem. Kinet. 2000,
32, 62.
(35) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr,
J. A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1997, 26, 521.
(36) Braun, W.; Herron, J. T.; Kahaner, D. K. Int. J. Chem. Kinet. 1988,
20, 51.
(37) Sehested, J.; Christensen, L. K.; Nielsen, O. J.; Bilde, M.;
Wallington, T. J.; Schneider, W. F.; Orlando, J. J.; Tyndall, G. S. Int. J.
Chem. Kinet. 1998, 30, 475.
(38) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling; Evaluation #12, 1997, JPL Publication 97-04, Pasadena CA.
(39) Tyndall, G. S.; Cox, R. A.; Granier, C.; Lesclaux, R.; Moortgat,
G. K.; Pilling, M. J.; Ravishankara, A. R.; Wallington, T. J. J. Geophys.
Res. 2000, submitted.
(40) Maricq, M. M.; Szente, J. J.; Kaiser, E. W.; Shi, J. J. Phys. Chem.
1994, 98, 2083.
(41) Helleis, F.; Crowley, J. N.; Moortgat, G. K. J. Phys. Chem. 1993,
97, 11464.
(42) Atkinson, R. Int. J. Chem. Kinet. 1997, 29, 99.
(43) Blitz, M.; Pilling, M. J.; Robertson, S. H.; Seakins, P. W. Phys.
Chem. Chem. Phys 1999, 1, 73.
(44) Caralp, F.; Devolder, P.; Fittschen, C.; Gomez, N.; Hippler, H.;
Me´reau, R.; Rayez, M. T.; Striebel, F.; Viskolcz, B. Phys. Chem. Chem.
Phys. 1999, 1, 2935.
References and Notes
(1) Duce, R. A.; Mohnen, V. A.; Zimmerman, P. R.; Grosjean, D.;
Cautreels, W.; Chatfield, R.; Jaenicke, R.; Ogren, J. A.; Pellizari, E. D.;
Wallace, G. T. ReV. Geophys. 1983, 21, 921.
(2) Arnold, F.; Knop, G.; Zeiereis, H. Nature 1986, 321, 505.
(3) Mo¨hler, O.; Reiner, Th.; Arnold, F. ReV. Sci. Instrum. 1993, 64,
1199.
(4) Singh, H. B.; Kanakidou, M.; Crutzen P. J.; Jacob, D. J. Nature
1995, 378, 50.
(5) Arnold, F.; Schneider, J.; Gollinger, K.; Schlager, H.; Schulte, P.;
Hagen, D. E.; Whitefield, P. D.; Van Velthoven, P. Geophys. Res. Lett.
1997, 24, 57.
(6) Arnold, F.; Bu¨rger, V.; Droste-Fanke, B.; Grimm, F.; Krieger, A.;
Schneider, J.; Stilp, T. Geophys. Res. Lett. 1997, 24, 3017.
(7) Singh, H. B.; O’Hara, D.; Herlth, D. Sachse, G. W.; Blake, D. R.;
Bradshaw, J. D.; Kanakidou, M.; Crutzen P. J. J. Geophys. Res. 1994, 99,
1805.
(8) Orlando, J. J.; Nozie`re, B.; Tyndall, G. S.; Paulson, S. E.;
Orzechowska, G.; Rudich, Y. J. Geophys. Res. 2000, 105, 11561.
(9) Jobson, B. T.; Niki, H.; Yokouchi, Y.; Bottenheim, J.; Hopper, F.;
Leaitch, R., J. Geophys. Res. 1994, 99, 25355.
(10) McKeen, S. A.; Gierczak, T.; Burkholder, J. B.; Wennberg, P. O.;
Hanisco, T. F.; Keim, E. R.; Gao, R. S.; Liu, S. C.; Ravishankara, A. R.;
Fahey, D. W. Geophys. Res. Lett. 1997, 24, 3177.
(11) Wennberg, P. O.; et al. Science 1998, 279, 49.
(12) Mu¨ller, J.-F.; Brasseur, G. J. Geophys. Res. 1999, 104, 1705.
(13) Wollenhaupt, M.; Carl, S. A.; Horowitz, A.; Crowley, J. N. J. Phys.
Chem. A 2000, 104, 2695.
(14) Wollenhaupt, M.; Crowley, J. N. J. Phys. Chem. A 2000, 104, 6429.
(15) Jenkin, M. E.; Cox, R. A.; Emrich, M.; Moortgat, G. K. J. Chem.
Soc., Faraday Trans. 1993, 89, 2983.
(45) Orlando, J. J.; Iraci, L. T.; Tyndall, G. S. J. Phys. Chem. A 2000,
104, 5072.
(46) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.5; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(47) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
(48) Mereau, R.; Rayez, M. T.; Caralp, F.; Jimeno, P.; Rayez, J. C.;
Gomez, N.; Fittschen, C. 15th International Symposium on Gas Kinetics,
Bilbao, 1998, poster A32, Book of Abstracts, p 134.
(49) Vereecken, L.; Huyberechts, G.; Peeters, J. J. Chem. Phys. 1997,
106, 6564.
(50) Forst, W. Theory of Unimolecular Reactions; Academic Press: New
York, 1973.
(51) Jungkamp, T. P. W.; Smith, J. N.; Seinfeld, J. H. J. Phys. Chem.
A 1997, 101, 1973.
(16) Bridier, I.; Veyret, B.; Lesclaux, R.; Jenkin, M. E. J. Chem. Soc.,
Faraday Trans. 1993, 89, 2993.
(17) Atkinson, R.; Carter, W. P. L. J. Atmos. Chem. 1991, 13, 195.
(18) Sehested, J.; Christensen, L. K.; Nielsen, O. J.; Bilde, M.;
Wallington, T. J.; Schneider, W. F.; Orlando, J. J.; Tyndall, G. S. Int. J.
Chem. Kinet. 1998, 30, 475.
(19) Wallington, T. J.; Hurley, M. D.; Fracheboud, J. M.; Orlando, J.
J.; Tyndall, G. S.; Sehested, J.; Møgelberg, T. E.; Nielsen O. J. J. Phys.
Chem. 1996, 100, 18116.
(20) Orlando, J. J.; Tyndall, G. S.; Bilde, M.; Ferronato, C.; Wallington,
T. J.; Vereecken, L.; Peeters, J. J. Phys. Chem. A 1998, 102, 8116.
(21) Vereecken, L.; Peeters, J. J. Phys. Chem. A 1999, 103, 1768.
(22) Vereecken, L.; Peeters, J.; Orlando, J. J.; Tyndall, G. S.; Ferronato,
C. J. Phys. Chem. A 1999, 103, 4693.
(23) Schneider, W. F.; Wallington, T. J.; Barker, J. R.; Stahlberg, E. A.
Ber. Bunsen-Ges. Phys. Chem. 1998, 102, 1850.
(24) Lightfoot, P. D.; Cox, R. A.; Crowley, J. N.; Destriau, M.; Hayman,
G. D.; Jenkin, M. E.; Moortgat, G. K.; Zabel, F. Atmos. EnViron. 1992,
26A, 1805.
(25) Shetter, R. E.; Davidson, J. A.; Cantrell, C. A.; Calvert, J. G. ReV.
Sci. Instrum. 1987, 58, 1427.
(26) Orlando, J. J. Int. J. Chem. Kinet. 1999, 31, 515.
(27) Wallington, T. J.; Andino, J. M.; Ball, J. C.; Japar, S. M. J. Atmos.
Chem. 1990, 10, 301.
(52) Troe, J. J. Chem. Phys. 1977, 66, 4745.
(53) Wittig, C.; Nadler, I.; Reisler, H.; Noble, M.; Catanzarite, J.;
Radhakrishnan, G. J. Chem. Phys. 1985, 83, 5581.
(54) Troe, J. J. Chem. Phys. 1977, 66, 4758.
(55) Troe, J. J. Phys. Chem. 1979, 83, 114.
(56) Oref, I.; Tardy, D. C. Chem. ReV. 1990, 90, 1407.
(57) Hirschfelder, J. O.; Curtiss, C. F.; Bird, R. B. Molecular Theory of
Gases and Liquids; John Wiley & Sons: New York, 1967.
(58) Cox, R. A.; Patrick, K. F.; Chant, S. A. EnViron. Sci. Technol.
1981, 15, 587.
(59) Baldwin, A. C.; Barker, J. R.; Golden, D. M.; Hendry, D. G. J.
Phys. Chem. 1977, 81, 2483.
(60) Koch, S.; Moortgat, G. K. J. Phys. Chem. A 1998, 102, 9142.