Kinetic studies of the reactions CH3+NO2→products, CH3O+NO2→products, and OH+CH3C(O)CH3→CH3C(O)OH+CH3, over a range of temperature and pressure

Article abstract of DOI:10.1021/jp0005726

The title reactions were investigated using pulsed laser photolysis combined with pulsed laser induced fluorescence detection of CH₃O to determine the rate coefficients for CH₃+NO₂→products (3) and CH₃O+NO₂→products (5) as a function of temperature and pressure, and to estimate the yield of CH₃ (and thus the yield of CH₃C(O)OH) from the reaction of OH with CH₃C(O)CH₃ (2) at two different temperatures. Reaction 3 has both bimolecular and termolecular components: a simplified falloff parametrization with F_cent = 0.6 gives k_3b⁰ = (3.2±1.3)×10^-28(T/297)^-0.3 cm⁶ s^-1 and k_3b^∞ = (4.3±0.4)×10^-11(T/297)^-1.2 cm³ s^-1 with CH₃NO₂ the likely product. The rate constant for the bimolecular reaction pathway to form CH₃O+NO (3a) was found to be 1.9×10^-11 cm^-3 s^-1. The low- and high-pressure limiting rate coefficients for reaction between CH₃O and NO₂ to form CH₃ONO₂ (5b) were derived as k_5b⁰ = (5.3±0.3)×10^-29(T/297)^-4.4 cm⁶ s^-1 and k_5b^∞ = (1.9±0.05)×10^-11(T/297)^-1.9 cm³ s^-1, respectively. Although the final result is associated with some experimental uncertainty, we find that CH₃ is formed in the reaction between OH and CH₃C(O)CH₃ at ≈50% yield at room temperature and 30% at 233 K.

Full text of DOI:10.1021/jp0005726

J. Phys. Chem. A 2000, 104, 6429-6438
6429
Kinetic Studies of the Reactions CH3 + NO2 f Products, CH3O + NO2 f Products, and
OH + CH3C(O)CH3 f CH3C(O)OH + CH3, over a Range of Temperature and Pressure
Matthias Wollenhaupt and John N. Crowley*
Max-Planck-Institut f u¨ r Chemie, DiVision of Atmospheric Chemistry, Postfach 3060, 55020 Mainz, Germany
ReceiVed: February 11, 2000; In Final Form: April 5, 2000
The title reactions were investigated using pulsed laser photolysis combined with pulsed laser induced
fluorescence detection of CH
NO f products (5) as a function of temperature and pressure, and to estimate the yield of CH
the yield of CH C(O)OH) from the reaction of OH with CH C(O)CH (2) at two different temperatures.
3
O to determine the rate coefficients for CH
3
+ NO
2
f products (3) and CH
3
O
+
2
3
(and thus
3
3
3
Reaction 3 has both bimolecular and termolecular components: a simplified falloff parametrization with Fcent
0
-28
-0.3
6
-1
∞
3b
-11
-1.2
3
-1
)
0.6 gives k3b ) (3.2 ( 1.3) × 10 (T/297)
with CH NO
NO (3a) was found to be 1.9 × 10
reaction between CH O and NO to form CH
cm s and k ) (4.3 ( 0.4) × 10 (T/297)
cm s
3
2
the likely product. The rate constant for the bimolecular reaction pathway to form CH
3
O +
-
11
-3 -1
cm s . The low- and high-pressure limiting rate coefficients for
0
-29
3
∞
2
3
-11
ONO
2
(5b) were derived as k ) (5.3 ( 0.3) × 10 (T/
5
b
-
4.4
6
-1
-1.9 3 -1
2
97)
associated with some experimental uncertainty, we find that CH
CH C(O)CH
at ≈50% yield at room temperature and 30% at 233 K.
cm s and k5b ) (1.9 ( 0.05) × 10 (T/297)
cm s , respectively. Although the final result is
3
is formed in the reaction between OH and
3
3
9
1
. Introduction
A combination of field measurements and modeling studies
has recently highlighted the important role of acetone (CH3C-
O)CH3) in the chemistry of the upper troposphere, where it
south Atlantic. We note also that even at room temperature
vo iu ar ca hn aa nl yn se i ls 2i na .d icated that ≈40% of the reaction may proceed
The estimation of the branching ratio to CH3C(O)OH based
on the shape of the temperature dependence of the overall rate
coefficient is clearly not unambiguous. Therefore, in the present
study, we have carried out a series of experiments to confirm
the presence of reaction pathway 2a and to quantify its
contribution to the overall reaction of OH with CH3C(O)CH3
at different temperatures. As neither CH3C(O)OH or CH3 can
be detected by laser induced fluorescence (see below), we chose
to scavenge any CH3 formed in reaction 2 with NO2 to generate
CH3O (which can be detected by pulsed laser induced fluores-
cence (PLIF)). Detailed experiments of the temperature and
(
1-7
can contribute significantly to the concentration of HOx. The
major sink of CH3C(O)CH3 in the upper troposphere is
photolysis, though the reaction with OH contributes signifi-
cantly; in the lower troposphere this is reversed, and reaction
with OH is the major sink:
CH C(O)CH + hν f CH C(O) + CH
3
(1)
(2)
3
3
3
CH C(O)CH + OHf products
3
3
8
3 3
pressure dependence of the reactions of CH and CH O with
In a recent publication we have shown that the overall rate
coefficient, k2, for reaction of OH with CH3C(O)CH3 does not
display normal Arrhenius behavior, but that the temperature
NO2 were carried out to provide kinetic and mechanistic
information necessary to analyze the CH3O formation and decay
in this system. The reaction of CH O with NO has recently
-
12
-14
dependence, k2 ) 8.8 × 10
exp (-1320/T) + 1.7 × 10
3
2
3
-1
3 2
been identified as an important source of CH ONO in the lower
exp (423/T) cm s , can be explained if the reaction proceeds
via two pathways, one of which has a positive activation barrier
and the second an apparent negative activation barrier. We
hypothesized that this was due to a change in reaction
mechanism at low temperatures, where the initial step is
electrophillic addition of OH to form an association complex,
AC, of formula (CH3)2C(OH)O, that can dissociate to form
acetic acid and the methyl radical (2a). At higher temperatures
the reaction proceeds via H-atom abstraction (2b).
1
0
stratosphere, although accurate estimation of the source
strength was made difficult owing to the lack of kinetic data
for this termolecular reaction at the appropriate pressure and
temperature, which is remedied in the present experiments. In
addition, we note that radical-radical reactions of the type CH3
NO2 and CH3O + NO2 which have more than one reactive
channel are of particular interest from both kinetic and dynamic
standpoints. An improved coverage of experimentally acces-
sible parameter space (in this case, rate coefficients as a function
of temperature and pressure) is needed to aid and exploit
theoretical advances in the description of such reactions.
+
1
1
CH C(O)CH + OH f AC f CH C(O)OH + CH
(2a)
f CH C(O)CH + H O (2b)
3
3
3
3
3
2
2
2
. Experimental Section
We further speculated that reaction 2a could represent a
significant source of CH3C(O)OH in the upper troposphere
The experiments were carried out using the method of pulsed
(where cold temperatures favor its formation), and help explain
laser photolysis combined with pulsed laser induced fluorescence
recent observations of this species at high altitudes over the
(PLP-PLIF). Details of the experimental setup have been
1
0.1021/jp0005726 CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/16/2000

6430 J. Phys. Chem. A, Vol. 104, No. 27, 2000
Wollenhaupt and Crowley
8
presented in a recent publication. Briefly, CH3O, generated in
suitable precursor photochemistry initiated by the pulsed 193
or 351 nm radiation of an excimer laser, was excited at 292.8
photolysis reactor. The quantum yield for OH formation at 351
8
nm was taken to be unity; the absorption cross section used
-
19
2 14,15
was 1.83 × 10
cm .
-
1
2
2
12
nm (34 152 cm , A A1 r X E, V3′ ) 4) and its red-shifted
emission was detected with a photomultiplier (PMT) screened
by an interference filter with maximum transmission at 330 nm
and a bandwidth at half-maximum of 5.5 nm. The filter
effectively shielded the PMT from scattered light from both
the excitation laser (292.8 nm) and the photolysis laser (351 or
93 nm). The tunable excitation radiation was provided by
frequency doubling the narrow-band emission from a dye laser
Rhodamine B), which was pumped at 532 nm using the second
Detection of absorption between 205 and 370 nm using a
0.5 m monochromator equipped with a 1024 pixel diode array
camera and a D2 lamp as analysis light source enabled the
concentrations of HONO, CH3C(O)CH3, and NO impurity to
be measured simultaneously. Optical density measurements
between 205 and 370 nm were converted to concentrations of
NO, HONO, and CH3C(O)CH3 by least-squares fitting to
reference spectra that had been obtained in the same setup (NO
1
8
14,15
(
and CH3C(O)CH3 ) or to a literature spectrum (HONO
). For
harmonic of a Nd:YAG laser operating at 10 Hz. The CH3O
excitation spectrum, measured by scanning the dye laser between
NO, a non-Beer-Lambert relationship between the measured
optical density and the NO concentration was observed, due to
spectrally unresolved γ-bands. This was taken into account when
calculating the NO concentration.
The analysis of the CH3O profiles thus obtained required
kinetic data for the reactions of both CH3 with NO2 and CH3O
with NO2 at the pressures and temperatures of these experiments.
2.2. CH3 + NO2 f Products, CH3O + NO2 f Products.
In a second set of experiments we conducted a detailed study
of the reactions of CH3 and CH3O with NO2:
2
91 and 295 nm, showed qualitative agreement with that
13
presented by Wantuck et al. Time-resolved CH3O profiles were
measured by variation of the delay between the photolysis laser
and the excitation laser, both running at 10 Hz. In practice, the
Nd:YAG laser was triggered with a constant delay, and the
excimer laser delay incremented relative to it after every laser
pulse. This way a complete profile, consisting of 200 data points,
was constructed in 20 s, and 20 scans were usually averaged to
improve the signal-to-noise ratio. Generally 10-20 data points
were also collected before the excimer laser pulse, which gives
the background 292.8 nm scattered light intensity. A few percent
of the excitation laser emission was beam split to a photodiode,
and the signal obtained was used to normalize the fluorescence
signal for pulse-to-pulse fluctuations in the intensity.
CH + NO f CH O + NO
(3a)
(3b)
(5a)
(5b)
3
2
3
CH + NO + Mf CH NO + M
3
2
3
2
CH O + NO f HCHO + HONO
3
2
The experiments were carried out at pressures between 10
and 200 Torr of Ar (measured by 10, 100, and 1000 Torr
capacitance manometers) and at temperatures between 223 and
CH O + NO + Mf CH ONO + M
3
2
3
2
In these experiments, CH3 radicals were generated by the 193
nm photolysis of acetone (6), which forms two CH3 radicals.16
356 K (determined by a J-type thermocouple inserted into the
reaction volume). In a single experiment N2 bath gas was also
used. The total flow rate was regulated using calibrated mass
CH C(O)CH + hν(193 nm) f 2CH + CO
(6)
3
3
3
3
flow controllers and varied between 40 and 450 cm (STD)
-
1
min (sccm) depending on the total pressure. Acetone was
stored as either a 5% or 5‰ mixture in Ar and introduced into
the system via a flow controller. HONO was generated in situ
by the reaction of HCl with NaNO2 crystals (see below); CH3-
ONO was introduced into the reactor from a 5% mixture in a
We note that CH3 is vibrationally excited17 in V3 when formed
from acetone photolysis at 193 nm (V3 ) 1:2:3 ) 0.73:0.14:
0
6
.13). The rate coefficient for vibrational deactivation by Ar is
.8 × 10 cm s . At the lowest pressure of these experiments
-13
3 -1
(10 Torr), this corresponds to a relaxation lifetime of ≈4 µs,
1
0 L blackened glass bulb via a Teflon needle valve. NO2 was
which is much shorter than the lifetime of CH3 due to reaction
with NO2 (see below).
prepared as described below and stored at ≈5% dilution in Ar,
in a blackened glass bulb. The dilute NO2/Ar mixture was
introduced into the reactor via mass flow controllers.
The sensitivity of the setup to CH3O was determined by the
photolysis of CH3ONO at 351 nm (7). Measurement of the laser
2
.1. OH + CH3C(O)CH3 f CH3 + CH3C(O)OH. The
experiments can be broadly divided into two sets. In one set
we probed the generation of CH3C(O)OH in reaction 2a, by
monitoring the formation of its coproduct, CH3, via a conversion
reaction to CH3O, which can be detected using PLIF.
CH ONO + hν(351 nm) f CH O + NO
(7)
3
3
fluence using a calibrated Joulemeter, and the concentration of
CH3ONO, measured by optical absorption, allowed the amount
of CH3O generated in the laser-irradiated volume to be
calculated. The CH3ONO concentration in these experiments
was derived by least-squares fitting of the optical absorption
between 300 and 360 nm to a reference spectrum measured
using the same absorption cell. The cross section at the
photolysis wavelength (351 nm) was determined to be 3.31 ×
OH + CH C(O)CH f CH C(O)OH + CH
3
(2a)
(3a)
3
3
3
CH + NO f CH O + NO
3
2
3
In these experiments, OH was formed in the 351 nm (excimer
8
laser) photolysis of HONO (4) as described previously, but
-
19
2
18
1
0
cm , which is in good agreement with literature values.
The quantum yield for CH3O formation at wavelengths close
to 350 nm has been determined as approximately unity.19
2.3. Chemicals. CH3ONO was synthesized by the dropwise
addition of 50% H2SO4 to an aqueous solution of NaNO2 and
CH3OH, and carried in a stream of He into a cold trap at -78
°C. HONO was prepared in a dynamic flow system by passing
HCl over a stirred bed of NaNO2 crystals. HONO made this
HONO + hν(351 nm) f OH + NO
(4)
1
2
was not detected. Relatively high concentrations of up to 10
-3
cm OH were generated in order to maximize the concentration
of CH3 and consequently that of CH3O. The concentration of
OH generated this way was calculated from the 351 nm laser
fluence (measured with a calibrated Joulemeter) and the
concentration of HONO, which was determined by optical
absorption in a 174 cm absorption cell located upstream of the
8
way contains high levels of NO impurity. NO2 was prepared
by mixing pure NO with a large excess of O2, which was then

Kinetics of CH3, NO2, and CH3C(O)CH3 Reactions
J. Phys. Chem. A, Vol. 104, No. 27, 2000 6431
Figure 1. CH
the presence of various excess concentrations of NO
was 233 K and the total pressure was 200 Torr Ar. A, [NO
3
O profiles in the 193 nm photolysis of CH
. The temperature
] ) 2.84
3 3
C(O)CH in
3 5 2
Figure 2. Plot of k′ and k′ versus [NO ] at 200 Torr of total pressure
and at 233 and 298 K. The data joined with the dashed line have not
2
2
1
4
-3
14
-3
14
been corrected to take into account the dimerization of NO2.
×
10 cm ; B, [NO
2
] ) 4.40 × 10 cm ; C, [NO ] ) 6.13 × 10
2
-
3
14
-3
cm ; D, [NO
2
] ) 9.24 × 10 cm
.
contributes significantly. Reactions of CH3 in the absence of
NO2 (i.e., self-reaction or with acetone) are negligibly slow,
and would in any case be expected to display a pressure and/or
temperature dependence. Also, the simultaneous formation of
O-atoms formed in the photolysis of NO2 was considered as a
removed by vacuum distillation. The NO2 prepared this way
typically had a NO impurity of 1%. NO, purchased at 99.5%
purity (Matheson), was passed through a silica gel column at
-
78 °C to remove higher oxides. Ar and N2 (Linde 99.999%)
were used without further purification.
11
potential source of error. However, although ≈1 × 10 O-atoms
-
3
1
3
cm (both O( D) and O( P)) are generated in the laser pulse
(
3
. Results and Discussion
.1. CH3 + NO2 and CH3O + NO2. The above reactions
3
for experiments with high NO2), the fate of O( P) will be
reaction with NO2 rather than with CH3 or CH3C(O)CH3. The
3
1
O( D) generated could react with CH3C(O)CH3 to form OH
were studied simultaneously by photolyzing acetone at 193 nm
in the presence of NO2. Experimental conditions were chosen
so that NO2 was in great excess over the CH3 radical. Typically,
radicals, though the fact that data obtained in Ar and N2 give
the same results suggests that any secondary chemistry related
1
14
-3
to O( D) formation has no significant influence on the CH or
acetone at concentrations of ≈1 × 10 cm was dissociated
3
2
CH O profiles. At present we have no explanation for this
with a laser fluence of ≈0.5 mJ /cm to give initial CH3
3
11
-3
apparent nonzero loss rate for CH in the absence of NO but
concentrations of [CH3]0 ≈ 3 × 10 cm , calculated using an
3
2
-
18
2
20
note that it is small, and probably does not significantly influence
the derivation of the rate coefficients for k3.
acetone cross section of 2.7 × 10
cm .
The NO2
1
5
-3
concentration was varied between ≈0.2 and 1.2 × 10 cm .
Under these conditions the self-reaction of CH3 radicals is
negligible, and reactions of CH3 or CH3O with secondary
products or impurities (e.g., NO) are insignificant. The time-
dependent CH3O concentration profile is given by
The accuracy of the rate coefficients obtained in this study
was enhanced by measuring the concentration of NO by optical
2
absorption in the 174 cm cell. Measurements of absorbance were
converted to concentrations by least-squares fitting to a reference
spectrum (310-365 nm) measured in the same setup and with
the same instrumental resolution. The reference spectrum
showed good agreement with the spectrum previously deter-
k′3a
[
CH O] )
(e-k′3t - e-k′5t)[CH3]0
(i)
3
t
k′ - k′3
5
21
mined in this laboratory. In addition, NO2 was monitored after
the reaction vessel by its absorption of the broad band emission
of a light-emitting diode (LED, 410-440 nm) over a 44 cm
path length. An effective cross section was established relative
to absorption in the 174 cm cell, which allowed us to check
that no NO2 was lost in transit through the thermostated reaction
vessel. The 174 and 44 cm absorption cells were at room
temperature. The optically determined NO2 concentration in the
where k′i is equal to ki[NO2].
Time-dependent CH3O profiles obtained in this manner and
for different initial [NO2] are displayed in Figure 1. The variation
in the maximum height of the CH3O signal with [NO2] is due
to quenching of the fluorescence by NO2. The CH3O profiles
were fit to eq i to obtain k′3 and k′5 for a given temperature,
pressure, and bath gas. The individual rate coefficients k3 and
k5 are obtained by plotting k′3 and k′5 versus [NO2] as shown
in Figure 2 for data obtained at 200 Torr and at 233 and 298 K.
We note that for k5 the intercept does not deviate from zero
within statistical uncertainty. However, the data for k3 display
1
74 cm cell was then corrected for pressure and temperature
differences between the absorption cell and the reaction vessel
to obtain its concentration. At temperatures of 262, 297, and
3
56 K the concentration of NO2 is known to within 5%. At the
-
1
lowest temperatures of this study 233 K, a further correction to
the NO2 concentration had to be made due to its dimerization
to N2O4.
a positive intercept of between 1100 and 2200 s , with no
systematic dependence on either the temperature or pressure of
the experiment. The lack of variation with pressure rules out
that diffusion from the reaction zone (usually a maximum of
-
1
2NO T N O
4
(8,-8)
≈
200 s ) or an effect related to vibrational excitation of CH3
2
2

6432 J. Phys. Chem. A, Vol. 104, No. 27, 2000
Wollenhaupt and Crowley
TABLE 1: Rate Coefficients for CH
3
+ NO
2
f Productsa
temperature (K)
297
pressure (Torr)
233
262
356
b
1
2
5
0
0
0
4.56
4.25 ( 0.07 3.84 ( 0.09
4.67 ( 0.12 4.41 ( 0.20
3.43 ( 0.10
3.95 ( 0.10
4.50 ( 0.09
4.72 ( 0.12
4.92 ( 0.14
c
b
c
b
c
b
c
b
c
4
.09
5.21
.72
5.97
.38
6.57
.84
6.85
.04
4
5.77 ( 0.14 5.10 ( 0.07
5.01 ( 0.11
6.25 ( 0.17 5.24 ( 0.16
d
5
1
2
00
00
5
6.12 ( 0.16 5.49 ( 0.13
6
a
Rate coefficients in units of 10-11 cm s-1. b Data corrected for
3
-
29
3 c
NO
2
dimerization using K
8
) 5.2 × 10
exp(6782/T) cm . Data
-
29
corrected for NO
2
dimerization using K
8
) 5.2 × 10 exp(6643/T)
3
22
cm . For T ) 262 K, the correction was of the order of just 1% and
d
was therefore not carried out. Value obtained in N
2
bath gas. Errors
f Productsa
356
are statistical only (2σ).
TABLE 2: Rate Coefficients for CH
3
O + NO
2
Figure 3. Falloff curves for CH
56 K were fitted simultaneously to eq ii as described in the text. The
rate constants at 233 K were corrected for the dimerization of NO to
3
2
O + NO . The data at 262, 297, and
temperature (K)
297
3
pressure (Torr) 233
262
2
-
29
form N
2
O
4
. The open squares were corrected using K
8
) 5.2 × 10
b
c
b
1
2
5
0
0
0
12.7
1.4
8.67 ( 0.02
5.56 ( 0.07 3.22 ( 0.06
7.46 ( 0.18 4.48 ( 0.06
3
exp(6643/T) cm ; the solid squares were corrected using K
8
) 5.2 ×
1
-29
3
1
0
exp(6782/T) cm .
15.9 11.6 ( 0.02
c
1
4.4
b
20.0 15.8 ( 0.02
10.8 ( 0.01
11.0 ( 0.10
13.2 ( 0.2
6.57 ( 0.08
7.64 ( 0.11
9.48 ( 0.16
c
d
1
8.1
b
1
2
00
00
22.6 18.0 ( 0.03
c
2
0.1
b
25.3 19.0 ( 0.03
15.1 ( 0.2
c
2
2.3
a
Rate coefficients in units of 10-12 cm s-1. b Data corrected for
3
-
29
3 c
NO
2
dimerization using K
8
) 5.2 × 10
exp(6782/T) cm . Data
-
29
corrected for NO
2
dimerization using K
8
) 5.2 × 10 exp(6643/T)
3
22
cm . For T ) 262 K, the correction was of the order of just 1% and
d
2
was therefore not carried out. Value obtained in N bath gas. Errors
are statistical only (2σ).
Initially the recommended equilibrium constant, K8 ) 5.2 ×
0 exp((6643 ( 250)/T) cm , was used. At the highest NO2
-29
3
22
1
15
-3
concentrations employed ([NO2] ) 1 × 10 cm and at 223
-
16
3
-1
K (K8 ) 1.25 × 10
cm molecule ) the correction applied
1
4
-3
is ≈25%, which reduces to ≈5% at NO2 ) 2 × 10 cm .
Using the recommended forward and backward rate coefficients
for (8,-8),23 we calculate that the relaxation time to reach
equilibrium between NO2 and N2O4 is <1 s even at the lowest
pressures (10 Torr) and the lowest NO2 concentrations. The
residence time for the flowing gas mixture in the cooled part
of the reactor is >4 s. The correction for dimerization of NO2
removes the curvature present in plots of k′ versus [NO2] at
low temperatures as seen in Figure 2.
Figure 4. Falloff curves for CH
3
2
+ NO . The data at 262, 297, and
3
56 K were fitted simultaneously to eq ii as described in the text. The
-
28
6
-1
solid lines are described by Fcent ) 0.6, k
0
) 3.18 × 10 cm s , k
∞
-11
-
11
3
-1
)
4.34 × 10 cm s , n ) 0.3, m ) 1.17, and k3a ) 1.91 × 10
3 -1
cm s . The rate constants at 233 K were corrected for the dimerization
-29
3
of NO
2
to form N
2
O
4
using 5.2 × 10 exp(6782/T) cm . Low-pressure
flow tube measurements in He are also shown: Yamada ) ref 30;
Biggs ) ref 26. To allow direct comparison, the data obtained in He
have been assigned an equivalent pressure of Ar to reflect the relative
third-body efficiencies of these bath gases.
Data obtained for k3 and k5 at various pressures between 10
and 200 Torr and at various temperatures between 223 and 356
K are listed in Tables 1 and 2. The rate coefficients obtained
for k3 and k5 were found to be independent of the excimer laser
fluence and thus initial [CH3] when varied over a factor of 3,
and independent of the flow rate and residence time of the
flowing gas mixture in the reaction vessel. The data also showed
no dependence on the time resolution used to measure the CH3O
profile, which suggested that sufficient data points were gathered
to determine the rate of the rapid part of the CH3O formation
at short times.
coefficients at each pressure and temperature) for each reaction
was fitted simultaneously to falloff curves described by
2
4
-
n
T
k0
[M]
(
)
2
97
-
p
k(M,T) )
(Fcent
-m
)
(ii)
n
T
297
T
297
(
)
1
+ k
[M]/k
(
0
(
)
∞
(
)
)
The temperature- and pressure-dependent rate coefficients are
also plotted in Figure 3 (for the reaction of CH3O with NO2)
and Figure 4 (CH3 + NO2). The complete data set (i.e., all rate
where k0 is the low-pressure termolecular rate coefficient, k∞ is
the limiting high-pressure rate coefficient, Fcent ) exp(-T/f),
which describes the temperature-dependent broadening of the

Kinetics of CH3, NO2, and CH3C(O)CH3 Reactions
falloff curves, and
J. Phys. Chem. A, Vol. 104, No. 27, 2000 6433
-
n
2 -1
T
log10 k0
[M]
(
(
)
)
)
297
p ) 1 +
-m
T
97
(
)
{
k
}
∞
(
2
3
.1.1. CH3O + NO2. For the reaction of CH3O with NO2, a
minor, bimolecular channel (5a) has been measured along with
25
the pressure-dependent association reaction 5b. In that work,
-
12
3
k5a was measured as k5a(T) ) 9.6 × 10
exp(-1150/T) cm
-
1
s , albeit with relatively large errors associated with both the
preexponential factor and the temperature dependence. For this
reason the present data was fitted to an equation of the form k5
)
k5a(T) + k5b(M,T). Note that neglecting the bimolecular
component (5a) would result in only small changes in the fits,
as its contribution to the overall rate coefficient at the higher
pressures of this experiment is small. The maximum effect is
found at high temperatures and low pressures, e.g., k5a/k5b ≈
0
.1 at 10 Torr and 356 K, and <0.01 at 10 Torr and 233 K.
Potential energy surfaces showing the energetics of the various
reaction pathways have been given by Biggs et al.26
Our initial calculations showed that the data at 356, 297, and
2
62 K could be well reproduced, but the fit overestimated the
measured rate coefficients at 233 K. This is apparent in Figure
, which displays the rate constant corrected as above for the
3
dimerization of NO2 using the recommended equilibrium
constant (open squares). By fitting to the data at 262, 297, and
Figure 5. Comparison of present data for CH
The solid lines were calculated from the present parametrization of
3
2
O + NO with literature.
-
29
6
-1
3
56 K only, we obtain the fits in Figure 3, which are described
the termolecular component (F ) 0.6, k ) 5.3 × 10 cm s , k∞
cent
0
-
29
6
-1
-11 3 -1
by f ) (571 ( 112), k0 ) (5.5 ( 1.1) × 10
cm s , k∞ )
) 1.86 × 10
cm s , n ) 4.36, m ) 1.87) plus the temperature-
-
11
3
-1
dependent bimolecular component as measured by McCaulley et al.
McCaulley ) ref 25; Biggs ) ref 26; Frost ) ref 28. To allow direct
comparison, the data obtained in He have been assigned an equivalent
pressure of Ar to reflect the relative third-body efficiencies of these
bath gases.
(
1.86 ( 0.06) × 10
cm s , n ) (3.42 ( 0.36), and m )
(1.80 ( 0.14). A small adjustment to the temperature depen-
dence of the equilibrium constant for the dimerization of NO2
-
29
3
to N2O4 (K8 ) 5.2 × 10
data back in line with this parametrization of the rate coefficient
solid squares in Figure 3). Note that this value of K8 is well
exp(6782/T) cm ) brings our 223 K
0
(
information concerning k . They obtained k
) 2.6 ×
3
5a
-
29
-4.5
6 -1
within the quoted error limits for the temperature dependence
of the equilibrium constant.
The value of f obtained converts to values of Fcent of 0.66 at
10 (T/300)
cm s , which compares well with the present
determination of n ) 4.4. A further low-pressure, flow tube
2
6
study was conducted by Biggs et al., who monitored CH3O
2
8
2
23 K, decreasing to 0.55 at 356 K, reflecting the expected
profiles to extract both k3 and k5. Frost et al. carried out flash
photolysis experiments, using the photolysis of CH3ONO as
CH3O source at pressures similar to those of the present study,
and at 295 and 390 K.
increase in the broadening factor as the temperature decreases.
Note that the temperature dependence of Fcent was not forced
in the fitting routine but was allowed to vary freely. If Fcent is
22
held at 0.6, irrespective of the temperature, we obtain k0 )
The results obtained in those experiments generally show
good agreement with the present parametrization of the rate
coefficients. The data close to room temperature are in good
agreement at all pressures; especially the flash photolysis data
-
29
6
-1
-11
3
(
5.3( 0.3) × 10
cm s , k∞ ) (1.86 ( 0.05) × 10
cm
-
1
s , n ) (4.36 ( 0.36), and m ) (1.87 ( 0.15); i.e., only n,
which describes the temperature dependence of the low-pressure,
third-order rate constant, is significantly changed, although there
is no discernible difference in the quality of the fit.
28
of Frost et al. agree particularly well. The present parametriza-
tion of the rate coefficients is in good agreement with that
2
6
In Figure 5, we compare our results with previous determina-
tions of the pressure and temperature dependence of the rate
coefficient. The solid lines are the calculated overall rate
constant k5, which was obtained using the parametrization of
the termolecular component (in Ar) as described above, plus
obtained by Biggs et al., who derived k0 ) (5.3 ( 0.2) ×
10 cm s with
cm s and k∞ ) (1.43 ( 0.03) × 10
Fcent fixed at 0.6, by analysis of their own data at room
-
29
6
-1
-11
3
-1
2
8
temperature along with those of Frost et al. and McCaulley et
2
5
al.
2
5
The low-pressure data of McCaulley et al.25 at 220, 250, 390,
and 473 K are also correctly reproduced. Some deviation is
the bimolecular component as described by McCaulley et al.
The temperatures were chosen to match the available literature
data. The data that were obtained in He have been corrected
for the relative third-body efficiencies of He and Ar (âc(He) )
2
8
observed for the data of Frost et al. at 390 K. The reason for
this is unclear, and we can only speculate that it is related to
the use of a different CH3O precursor system than in the present
experiments.
0
.07 and âc(Ar) ) 0.12) as determined in the experiments of
2
7
Frost et al.
McCaulley et al.25 measured the rate coefficients between
Although by far the most important loss process for CH3O
in the atmosphere is reaction with O2, a combination of
observations of CH3ONO2 in the lower stratosphere and upper
0.6 and 4.0 Torr of He and T ) 220-473 K using a low-pressure
flow tube and detection of CH3O with LIF. They also used the
reaction CH3 + NO2 f CH3O to generate CH3O, but did not
attempt to analyze the rise of CH3O at short time to gain
1
0
troposphere and modeling studies have led to the conclusion
that ≈50% of the observed CH3ONO2 is formed in reaction

6
434 J. Phys. Chem. A, Vol. 104, No. 27, 2000
Wollenhaupt and Crowley
-
5
5
b, with a further 50% provided by a minor channel (≈10 of
(for 1100-1400 K). For the termolecular channel they quote a
∞
-11
the overall rate coefficient) in the reaction of methylperoxy
CH3O2) with NO. The model used evaluated kinetic data on
high-pressure second-order limit of k ) 1.66 × 10 (T/
3
-
0.6
6 -1
(
1000)
cm s between 300 and 1400 K, which yields a value
22
∞
the reaction CH3O + NO2, which returns a value of k5b ) 2
-11
3
-1
of k3 ) 3.4 × 10
cm
s
at 298 K, in reasonable
-
11
3 -1
×
10 cm s for T ) 215 K and 100 Torr. The value of k5b
agreement with the present result. For the low-pressure, third-
at the same temperature and pressure from the present data is
0
3
-31
-6
3 -1
order limit they obtained k ) 6.9 × 10 (T/1000) cm s ,
-11
3 -1
3
× 10 cm s . Use of the present data to assess the sources
-28
6
-1
which gives a value of 9.9 × 10
cm s at 298 K.
of CH3ONO2 would therefore significantly reduce the extracted
branching ratio for an additional channel to CH3ONO2 formation
in the CH3O2 + NO reaction.
The present data indicate that the reaction between CH3 and
NO2 displays a pronounced pressure dependence, and that the
termolecular reaction pathway is important even at low pressures
and dominant at pressures greater than ≈20 Torr of Ar. Note
that, in the present experiments, the product(s) of the termo-
lecular channel was (were) not detected. By analogy with
previous low-pressure studies of the reactions of alkyl radicals
with NO2, we assume that CH3NO2 is formed, though we cannot
rule out stabilization of CH3ONO* to form CH3ONO at the
higher pressures and low temperatures used in this study.
3
.1.2. CH3 + NO2. The data for this reaction were treated in
a manner identical to that described above for k5. That is, we
fit the data at 233, 262, 297, and 356 K to eq ii with an
additional bimolecular component, k3a. In this case we initially
assumed that the bimolecular channel (3a) does not have a strong
temperature dependence,29 so that k3 ) k3a + k3b(M,T). Potential
energy surfaces showing the energetics of the various reaction
pathways have been given by Biggs et al.26 The present data
are displayed in Figure 4 (note that the data at 233 K have been
corrected for NO2 dimerization as described above). The
simultaneous fit to all the data yields f ) (682 ( 468), k0 )
We derive a bimolecular, pressure-independent rate coefficient
-
11
3
-1
at 298 K of k3a ) 1.9 × 10
cm s and note that our
parametrization of the reaction reproduces the low-pressure data
30
26
-28
6
-1
-11
3
of Yamada et al. and Biggs et al. If we extrapolate our results
to 1 Torr, we obtain a ratio k3b/k3 ) 0.23 at 298 K, increasing
to 0.31 at 223 K. These values are significantly higher than the
(
2.64 ( 1.7) × 10 cm s , k∞ ) (4.24 ( 0.36) × 10 cm
s , n ) -(0.47 ( 0.64), m ) (1.12 ( 0.12), and k3a ) 1.96 ×
cm s . Fcent is thus seen to decrease from 0.71 at 233
K to 0.59 at 356 K. If Fcent is held at 0.6, irrespective of the
-
1
-
11
3
-1
1
0
3
1
measurements of McCaulley et al., whose yields of CH3NO2
were between 4% and 7%. Some of this discrepancy will be
related to the use of He as bath gas in the experiments of
McCaulley et al., which is a less efficient third-body collisional
quencher (by a factor of ≈0.6 for CH3O + NO2) than Ar.
2
2
-29
6
-1
temperature, we obtain k0 ) (3.18 ( 1.3) × 10
cm s ,
-
11
3
-1
11
k∞ ) (4.34 ( 0.4) × 10 cm s , n ) (0.30 ( 0.6), m )
-
3
-1
(1.17 ( 0.1), and k3a ) 1.91 × 10
cm s . Also plotted in
3
0
Figure 4 are data obtained at low pressure by Yamada et al.
32
and Biggs et al.26
2 5 2
Investigations of the reaction of C H with NO at 300 K with
Yamada et al.30 obtained a value of k3a ) 2.5 × 10
-11
cm3
detection of C2H5NO2 and NO products from the termolecular
and bimolecular pathways, respectively, have resulted in a ratio
of termolecular to bimolecular components of ≈0.25 between
s-1 at 295 K in flow tube experiments at 295 K and at pressures
of 1 Torr with direct detection of CH3 using photoionization
mass spectrometry. Within the quoted error limits, this is in
good agreement with the present extrapolation to 1 Torr (see
Figure 4). Nitromethane (CH3NO2) was not observed in those
experiments, indicating that channel 3b was not important at
that pressure. This observation disagrees with the results of
McCaulley et al., who measured yields of CH3NO2 of 0.04 at
98 K and a total pressure of 0.5 Torr of He, increasing to 0.07
at 1 Torr of He. They found no temperature dependence of the
CH3NO2 yield at 1 Torr between 223 and 298 K.
0
.7 and 2.1 Torr of He, in line with the present results for CH3
+
NO2 when extrapolated to low pressures.
The difference between the present observations of a strong
pressure dependence in the reaction of CH3 and NO2 and the
literature data may reflect the fact that the previous measure-
ments were not carried out over extended pressure ranges, and
that the effect at low pressures is partially disguised by the
presence of the fast, pressure independent bimolecular reaction,
(3a). There are no temperature-dependent rate coefficients for
CH3 + NO2 at low pressures (close to 1 Torr) available. The
global fits to our data were of the form k3 ) k3a + k3b(M,T)
and were thus forced to return a temperature-independent rate
3
1
2
In further flow tube experiments at pressures of 1-10 Torr
of He, with LIF detection of CH3O to follow the CH3 kinetics,
Biggs et al. derived values for k3 and k5. They found k3 to be
independent of pressure between 1 and 7 Torr, with a rate
2
6
-11
3 -1
constant for k3a, which was found to be 1.91 × 10
cm s .
-
11
3
-1
The applicability of this assumption was tested by carrying out
experiments at two temperatures, 297 and 255 K, at constant
pressure (20 Torr). In such experiments, a variation in the rate
constant for k3a with temperature will be manifest as a change
in the height of the CH3O signal as the efficiency of CH3O
formation changes. The experiments were carried out back-to-
back and in swift succession so that experimental parameters
such as the fluence of both the photolysis laser and the excitation
laser remained largely unchanged. The profiles obtained at each
temperature were scaled for changes in CH3C(O)CH3 and thus
CH3 due to density changes as the gas mixture was cooled. The
concentration of NO2 in the absorption cell was monitored, and
its concentration in the reaction vessel was calculated as usual.
The data were then simulated using the parametrized rate
constants given above for k3 and k5. For both temperatures, not
only the shape but also the height of the profiles were well
constant of (2.3 ( 0.3) × 10 cm s . Although the result at
1
Torr is in good agreement with the present data set, the lack
of a pressure dependence is not commensurate with the present
study, which revealed a pronounced pressure dependence at
room temperature. A partial explanation for this difference may
be related to the use of a low-pressure flow tube to determine
the kinetics of a fast process (CH3 + NO2) and a significantly
slower process (CH3O + NO2) simultaneously. Close examina-
tion of the data presented by these authors shows that the raw
data are not accurately described by the fit to eq i, and that the
resulting scatter in the plot of k′ versus [NO2] encompasses rate
-
11
3
-1
constants between 3.3 and 1.5 × 10
cm s . This may be
related to problems with spatial resolution in the flow tube when
attempting to measure first-order decay rates of close to 2000
-
1
s
and to heterogeneous loss of CH3 and/or CH3O on the flow
29
tube walls. Gl a¨ nzer and Troe determined values for k3a and
k3b in shock-tube studies at 900-1400 K at pressures between
-
11
3
-1
reproduced when k3a was held at 1.9 × 10
cm s , which
-
11
3
-1
1
and 30 atm. They determined k3a ) 2.16 × 10
cm s
rules out a strong temperature dependence in k3a. Within

Kinetics of CH3, NO2, and CH3C(O)CH3 Reactions
J. Phys. Chem. A, Vol. 104, No. 27, 2000 6435
TABLE 3: Experimental Conditions for the Investigation of Reaction 2a
3
expt
T
CH
3
C(O)CH
3
HONO
NO
2
NO
OH
O( P)
1
1
1
1
1
1
5
5
5
5
5
5
14
14
14
14
14
14
14
14
14
14
14
14
13
13
14
13
13
13
12
12
12
11
12
12
12
12
12
12
12
12
11
13
15
6
7
8
233
233
233
297
297
297
9.7 × 10
3.9 × 10
8.2 × 10
7.0 × 10
7.4 × 10
3.4 × 10
2.0 × 10
1.8 × 10
2.5 × 10
1.4 × 10
1.5 × 10
1.5 × 10
1.6 × 10
2.0 × 10
3.3 × 10
2.6 × 10
1.6 × 10
1.6 × 10
6.8 × 10
9.2 × 10
1.2 × 10
9.8 × 10
8.2 × 10
8.8 × 10
1.3 × 10
1.2 × 10
1.7 × 10
9.4 × 10
1.0 × 10
1.0 × 10
2.3 × 10
2.9 × 10
4.8 × 10
3.9 × 10
2.4 × 10
3.9 × 10
a
-3
3
Temperature (T) in K. Concentrations in cm . All experiments at 20 Torr of total pressure. OH and O( P) concentrations are based on the
measured laser fluence at 351 nm and on the absorption cross sections of HONO and NO respectively, at this wavelength (in both cases a quantum
yield of unity is assumed.). The NO concentration at 233 K has been corrected for dimerization to form N
(size of the correction was ≈ 10%).
The NO concentration listed is that observed by optical absorption. This is incremented in the numerical simulation by the same concentration as
2
2
2 4
O
3
2
O( P), as both are formed in the dissociation of NO .
experimental uncertainty, we can place an upper limit of E/RT
-300/T for the temperature dependence of k3a.
.2. OH + CH3C(O)CH3 f CH3 + CH3C(O)OH. This
)
3
reaction was investigated by photolyzing HONO in the presence
of CH3C(O)CH3 and NO2 and by following the formation and
decay of CH3O by LIF. The main aim of these experiments
was to quantify the branching ratio to CH3 formation in reaction
2
, and thus an absolute determination of CH3O was necessary.
For this reason, directly prior to the experiments, the sensitivity
of the LIF detection scheme to CH3O was measured. This was
carried out by 351 nm photolysis of a known concentration of
2
CH3ONO at a known laser fluence (typically ≈20 mJ/cm ) and,
most importantly, in the presence of the same concentration of
acetone to be used in the studies of the OH + acetone reaction.
It was found that the sensitivity to CH3O was severely degraded
15
-3
in the presence of ≈(3-7) × 10 cm of acetone due to
fluorescence quenching. Such high concentrations of CH3C-
(O)CH3 were necessary to ensure that a high proportion of the
OH formed reacted with the CH3C(O)CH3 and not with HONO
or NO2. Optical measurements of both acetone and CH3ONO
revealed no reactivity between these two molecules. Also,
variation of the CH3ONO concentration made no significant
change in the sensitivity to CH3O, implying that at concentra-
tions of ≈1 × 10 cm the CH3ONO did not significantly
quench the CH3O fluorescence. The dye laser and excimer laser
fluences were constantly monitored to enable corrections to the
calibration factor for the CH3O concentration to be made if
necessary.
Figure 6. Formation and reaction of CH
3
O in the 351 nm photolysis
of HONO/CH C(O)CH /NO mixtures at 20 Torr (Ar) and 297 K. The
3
3
2
experimental conditions for each experiment can be taken from Table
3. Solid points are experiment 6; open symbols are experiment 8. The
solid curves are the result of the numerical simulation using the reactions
listed in Table 4 (apart from those labeled “other”).
1
4
-3
reaction scheme, which is listed in Table 4. The reaction scheme
uses the data for the reactions of CH3 and CH3O with NO2
measured as part of the present study. The simulations are able
to reproduce the general features of the CH3O formation and
decay and the change in shape and concentration as experimental
2
Altogether six experiments were carried out at 233 and 298
K, using various reactant concentrations. The experimental
conditions are listed in Table 3. Note that, with the exception
parameters such as the initial NO and acetone concentration
are changed. Although weak, the CH O signals and their time
dependence are clearly good indicators of the formation and
3
3
of OH and O( P), all concentrations were determined by optical
absorption measurements. CH3O profiles from such experiments
are shown (for 297 K) in Figure 6. Due to the low signal, 200
scans over ≈20 min were made to improve the signal-to-noise
ratio, with each data point normalized to the excitation laser
fluence. A correction to each profile was also made to account
for HONO fluorescence following excitation at 292.8 nm. In
experiments with no CH3C(O)CH3 present, we observed a rapid
drop in background signal directly after the excimer laser pulse,
followed by a further slow decrease over several milliseconds.
Analysis of the time dependence of this signal showed that it
was due to the immediate removal of HONO by the 351 nm
laser pulse and then by reaction of HONO with OH. We
conclude therefore that HONO fluoresces weakly at ≈330 nm
following excitation at 292.8 nm. Normally this weak signal
would not be significant, but due to the low CH3O signals in
these experiments, a correction to the CH3O profile had to be
applied to account for it.
reaction of CH in this system. The slight discrepancy between
3
the measured rate of formation of the CH O and the modeled
3
value is presumably due largely to errors associated with the
calculated total loss rate of OH due to reaction with CH C(O)-
3
CH , NO , and HONO.
3
2
If we initially make the assumption that CH is formed only
3
in the title reaction, we can estimate branching ratios (R ) k /
2a
k ) to CH formation by adjusting R (see Table 4) until the
2
3
curves best fit the data points. The best fits were obtained with
a branching ratio R(297 K) ) (0.5 ( 0.15), where the errors
simply reflect the variability of the fitted branching ratio between
the three experiments. At 233 K the branching ratio obtained is
R ) (0.3 ( 0.1). These values are also subject to large potential
systematic errors related to the estimation of radical concentra-
tions (both OH and CH O) by fluence measurements, although
3
this error is somewhat reduced by the fact that the same
Joulemeter at the same laser wavelength was used to estimate
both OH and CH3O concentrations. Large errors are also
associated with the indirect method of determining the CH3
The solid lines through the experimental data are simulated
CH3O profiles, generated by numerical simulation of an assumed

6436 J. Phys. Chem. A, Vol. 104, No. 27, 2000
Wollenhaupt and Crowley
3
TABLE 4: Reaction Scheme Used To Model the CH O Profiles in the OH + CH
3
C(O)CH
3
+ NO
2
Experiments
k(233 K)
k(297 K)
note
Reactions of OH
-
13
-13
OH + CH
OH + CH
3
C(O)CH
C(O)CH
3
3
f CH
f CH
3
C(O)CH
2
+ H
2
O
(1.3 × 10 )R
(1.7 × 10 )R
a
a
b
b
b
-
13
-13
3
3
3
+ CH C(O)OH
(1.3 × 10 )(1 - R)
(1.7 × 10 )(1 - R)
-
-
-
12
12
13
-12
OH + HONO f H
2
O + NO
+ M f HNO
OH + NO + M f HONO + M
2
3.4 × 10
3.4 × 10
9.3 × 10
4.8 × 10
-12
OH + NO
2
3
+ M
1.2 × 10
-13
4.1 × 10
Reactions of Alkyl Radicals
-
-
11
11
-11
CH
CH
CH
CH
CH
3
3
3
3
3
C(O)CH
C(O)CH
+ NO
+ NO
+ NO + M f CH
2
+ NO
+ NO f products
O + NO
+ M f CH NO
NO + M
2
f products
1.6 × 10
1.6 × 10
c
c
-11
2
2.6 × 10
2.6 × 10
-
-
11
-11
2
f CH
3
1.91 × 10
3.43 × 10
1.91 × 10
this work
this work
d
11
-11
2
3
2
+ M
2.54 × 10
-
12
-12
3
2.8 × 10
1.4 × 10
Reactions of CH
3
O
-
11
-12
CH
CH
CH
CH
3
3
3
3
O + NO
O + NO
2
2
+ M f CH
f products
3
ONO
2
+ M
1.69 × 10
7.62 × 10
this work
-
14
11
12
-13
6.9 × 10
2.0 × 10
e
b
f
-
-12
O + NO + M f CH
3
ONO + M
1.1 × 10
5.1 × 10
-
-12
O + NO f HCHO + HNO
3.8 × 10
3.0 × 10
Other
-
11
11
-12
O + NO
2
f NO + O
2
1.1 × 10
9.7 × 10
b
g
h
i
5
5
CH
CH
CH
3
3
3
C(O)CH
C(O)CH
2
O f CH
O + NO
f CH
3
CO + HCHO
+ M f CH C(O)CH
+ CO + NO
2 × 10
2 × 10
2
2
3
2 2
ONO + M
-
-11
CO + NO
2
3
2
2.5 × 10
2.5 × 10
a
8
b
Temperature-dependent rate coefficient from Wollenhaupt et al.
Temperature/pressure-dependent rate coefficients calculated for 20 Torr
22
c
34
using JPL parametrization. Data at room temperature only from Sehested et al. There are no measurements of this parameter at low temperatures.
d
36
e
Data for the low-pressure third-order rate coefficient provided by Davies et al. was used to calculate k at 20 Torr of Ar. McCaulley et al.,
2
5
f
31
g
35
6
-1
1
985.
McCaulley et al., 1990.
Jenkin et al. measured a decomposition frequency of >1 × 10 s at 298 K and 700 Torr of N
2
, which
3
7
6
-1
broadly agrees with the estimation of Baldwin of 8 × 10 s . The value used is corrected for the pressure of 20 Torr in the present experiments.
h
i
38
Assumed to be the same as CH
decomposition of the initially formed CH
3
O + NO
2
+ M f CH
3
ONO
2
+ M. Value obtained at room-temperature only, and assumes immediate
and CO
3
CO
2
product to form CH
3
2
.
concentration and the need to calculate the partitioning of OH
to a number of reactants such as CH3C(O)CH3, NO2, HONO,
and NO, although this error is reduced somewhat by making
optical measurements of all of the species.
The trend in branching ratio with temperature is contrary to
that expected if CH3 is formed in reaction 2a via an association
complex, the stabilization of which should be more favorable
at low temperature.8 This result can be understood if the
decomposition of the vibrationally excited association complex,
are no measurements of this constant. The fate of alkoxy radicals
in the atmosphere is either reaction with oxygen, decomposition
3
3
or rearrangement. For (CH3)2C(OH)O the rearrangement
pathway is precluded as this is accessible only to longer chain
alkoxy radicals. There are also no H-atoms on the carbonyl
C-atom that can be abstracted by reaction with O2, and the fate
in the atmosphere (where NO2 is not available at the concentra-
tions employed in the laboratory work) will probably be
decomposition.
[
(CH3)2C(OH)O], proceeds either partially or completely via
One further explanation for the present observation of a
decreasing CH3 yield with decreasing temperature, which we
have already touched upon, is that the bimolecular rate coef-
ficient for the reaction of CH3 + NO2 is significantly lower at
formation of an alkoxy radical:
OH + CH C(O)CH f [(CH ) C(OH)O] f
3
3
3 2
2
33 K than at 298 K, meaning that more CH3 is converted to
(
CH ) C(OH)O (9)
3 2
CH3NO2 and less is transformed to detectable CH3O (see above).
Neither the present data on the CH3 + NO2 reaction at pressures
of 10 Torr or more nor the literature data have accurately
established the temperature dependence of k3a. The reaction of
NO2 with H-atoms has some similarity with the bimolecular
component of the reaction of CH3 with NO2 as both reactions
may be considered O-atom transfer to form NO and a further
oxygen-containing radical:
where (CH3)2C(OH)O is the R-hydroxyisopropoxy radical,
which can either decompose via C-C bond fission to form
CH3C(O)OH + CH3 or react with NO2 to form a nitrate:
(
CH ) C(OH)O + M f CH + CH C(O)OH + M
(10)
3
2
3
3
(
CH ) C(OH)O + NO + M f
3
2
2
(
CH ) C(OH)ONO + M (11)
3 2 2
H + NO f OH + NO
(12)
2
As the decomposition rate of the alkoxy radical will be
strongly dependent on temperature (decreases with decreasing
temperature), it is reasonable to assume that the competition
between reactions 10 and 11 will favor nitrate formation at low
temperatures, and the apparent yield of CH3 will decrease. On
This reaction has a rate coefficient described by k12 ) 4.0 ×
10 and thus displays a weak
temperature dependence. We therefore examined the effect of
-
10
3
-1 22
exp(-340/T) cm
s
fitting all the data in Figure 4 to k3 ) k3a(T) + k3b(M,T), where
14
-3
-11
3
-1
the basis of the known concentration of NO2 of ≈2 × 10 cm
k3a(T) ) 4.5 × 10
exp(-250/T) cm s . These parameters
-
11
and an estimated rate coefficient for reaction 14 of ≈4 × 10
were chosen to reproduce the room-temperature rate coefficient
3
-1
-11
3 -1
cm s , we estimate a loss rate of (CH3)2C(OH)O of ≈8000
of 1.9 × 10
cm s for k3a but remain within the limits we
-1
s
due to reaction with NO2. The decomposition frequency of
CH3)2C(OH)O at 20 Torr would have to be of a similar
magnitude to compete, though, as far as we are aware, there
set for the variation with temperature given above (i.e., less
than -300/T). The global fits thus obtained are not significantly
different from those obtained with a temperature-independent
(

Kinetics of CH3, NO2, and CH3C(O)CH3 Reactions
J. Phys. Chem. A, Vol. 104, No. 27, 2000 6437
Figure 7. Schematic diagram illustrating the reactions studied and considered in the present work. The species in brackets is the association
complex, which may have some characteristics of a vibrationally excited alkoxy radical. The numbers adjacent to the arrows represent the reaction
numbers as given in the text.
rate coefficient of k3a ) 1.9 × 10-11 cm s . However, at low
3
-1
Any CH3 radicals formed in this sequence will react with
NO to form CH NO and CH O which would result in an
temperatures and 20 Torr, the ratio of k3b/k3 increases to almost
2
3
2
3
0
.7 and the conversion of CH3 to CH3O is less efficient. This
erroneous estimation of the yield of CH from reaction 2.
3
results in an underestimation of the branching to CH3 at low
temperatures.
Finally, we also consider the possibility that CH3 can be
formed in reactions of the acetonyl radical (CH3C(O)CH2)
formed in the abstraction reaction of OH with CH3C(O)CH3
The results of Sehested et al.34 indicate that the reaction
between CH3C(O)CH2 and NO2 does not have a significant
reaction pathway to form the R-carbonylalkoxy radical under
their experimental conditions (1000 mbar of total pressure,
mainly SF6), implying that the reaction proceeds predominantly
via a termolecular pathway to form CH3C(O)CH2NO2 (13b).
(2b):
However, this might not reflect the situation at 20 Torr, and
to estimate the potential formation of CH3 via this mechanism,
we carried out numerical simulations of the reaction scheme in
Table 4, in this case including the reactions designated “other”
and assuming a worse-case scenario in which the reaction of
CH3C(O)CH2 with NO2 always generates CH3C(O)CH2O, i.e.,
k13a/k13 ) 1. A detailed reaction scheme of this chemistry and
other reactions studied as part of the present investigations is
given in Figure 7. We found that the variation in the CH3O
signal with experimental parameters could not be simulated by
assuming CH3 formation in reactions of CH3C(O)CH2 only.
Generally the formation of CH3 from reactions 2b, 13a, and
OH + CH C(O)CH f CH C(O)CH + H O (2b)
3
3
3
2
2
The acetonyl radical is known34 to react with NO2 (13) and
NO (14):
CH C(O)CH + NO f CH C(O)CH O + NO
(13a)
3
2
2
3
2
CH C(O)CH + NO + Mf CH C(O)CH NO + M
3
2
2
3
2
2
(13b)
CH C(O)CH + NO + M f CH C(O)CH NO + M (14)
3
2
3
2
1
4
5-17 and the conversion of CH3 to CH3O was delayed by
00-500 µs relative to the observed signal, even if the
The R-carbonylalkoxy radical CH3C(O)CH2O formed in
reaction 13a is expected to decompose readily35 to form the
acetyl radical and HCHO (15). The fate of the acetyl radical in
these experiments is reaction with NO2 to form CH3CO2 which
will unimolecularly decompose to form CH3 radicals:
6
decomposition of CH3C(O)CH2O was assumed to be 1 × 10
-
1
s
(as measured at atmospheric pressure). This is due to the
need for two reactions with NO2 (one by CH3C(O)CH2 and one
by CH3CO) to form the CH3 radical (see reaction scheme in
Table 4). The results of this simulation (Figure 8) clearly show
that CH3 cannot arise solely from secondary reactions of the
CH3C(O)CH2 radical. We therefore conclude that CH3 is formed
directly in the reaction of OH radicals with CH3C(O)CH3 and
with a yield of ≈50% at room temperature, though the indirect
nature of the present experiments does not allow us to place a
CH C(O)CH O + M f CH CO + HCHO + M (15)
3
2
3
CH CO + NO f CH CO + NO
(16)
(17)
3
2
3
2
CH CO + M f CH + CO + M
3
2
3
2

6438 J. Phys. Chem. A, Vol. 104, No. 27, 2000
Wollenhaupt and Crowley
(3) Arnold, F.; B u¨ rger, V.; Droste-Fanke, B.; Grimm, F.; Krieger, A.;
Schneider, J.; Stilp, T. Geophys. Res. Lett. 1997, 24, 3017.
4) Jaegl e´ , L.; Jacob, D. J.; Wennberg, P. O.; Spivakovsky, C. M.;
(
Hanisco, T. F.; Lanzendorf, E. J.; Hintsa, E. J.; Fahey, D. W.; Keim, E. R.;
Proffitt, M. H.; Atlas, E. L.; Flocke, F.; Schauffler, S.; McElroy, C. T.;
Midwinter, C.; Pfister, L.; Wilson, J. C. Geophys. Res. Lett. 1997, 24, 3181.
(5) Jaegl e´ , L.; Jacob, D. J.; Brune, W. H.; Tan, D.; Faloona, I. C.;
Weinheimer, A. J.; Ridley, B. A.; Campos, T. L.; Sachse, G. W. Geophys.
Res. Lett. 1998, 25, 1709.
(
6) Folkins, I.; Chatfield, R.; Singh, H.; Chen, Y.; Heikes, B. Geophys.
Res. Lett. 1998, 25, 1305.
7) Wennberg, P. O.; Hanisco, T. F.; Jaegl e´ , L.; Jacob, D. J.; Hintsa,
(
E. J.; Lanzendorf, E. J.; Anderson, J. G.; Gao, R.-S.; Keim, E. R.; Donnelly,
S. G.; Del Negro, L. A.; Fahey, D. W.; McKeen, S. A.; Salawitch, R. J.;
Webster, C. R.; May, R. D.; Herman, R. L.; Proffitt, M. H.; Margitan, J. J.;
Atlas, E. L.; Schauffler, S. M.; Flocke, F.; McElroy, C. T.; Bui, T. P. Science
1
997, 279, 49.
8) Wollenhaupt, M.; Carl, S. A.; Horowitz, A.; Crowley, J. N. J. Phys.
Chem. A 2000, 104, 2695.
9) Jacob, D. J.; Heikes, B. G.; Fan, S.-M.; Logan, J. A.; Mauzerall,
(
(
D. L.; Bradshaw, J. D.; Singh, H. B.; Gregory, G. L.; Talbot, R. W.; Blake,
D. R.; Sachse, G. W. J. Geophys. Res. 1996, 101, 24235.
(10) Flocke, F.; Atlas, E.; Madronich, S.; Schauffler, S. M.; Aikin, K.;
Margitan, J. J.; Bui, T. P. Geophys. Res. Lett. 1998, 25, 1891.
(
(
(
11) Smith, I. W. M. J. Chem. Soc., Faraday Trans. 1991, 87, 2271.
12) Inoue, G.; Akimoto, H.; Okuda, M. J. Chem. Phys. 1980, 72, 1769.
13) Wantuck, P. J.; Oldenborg, R. C.; Baughcum, S. L.; Winn, K. R.
Figure 8. Numerical simulation of the CH
b as CH source (solid lines) and using the formation and reactions
of CH C(O)CH and CH C(O)CH O in the presence of NO as CH
source (broken lines, see text). Solid circles, data gathered at 297 K
experiment 6); open circles, data gathered at 233 K (experiment 11).
3
O profiles using reaction
2
3
3
2
3
2
2
3
O
J. Phys. Chem. 1987, 91, 3253.
14) Bongartz, A.; Kames, J.; Welter, F.; Schurath, U. J. Phys. Chem.
991, 95, 1076.
15) Bongartz, A.; Kames, J.; Schurath, U.; George, C.; Mirabel, P.;
Ponche, J. L. J. Atmos. Chem. 1994, 18, 149.
16) Lightfoot, P. D.; Kirwan, S. P.; Pilling, M. J. J. Phys. Chem. 1988,
2, 4938.
(
(
1
(
firm number on this branching ratio, and especially the
temperature dependence of this process.
(
9
(
(
17) Donaldson, D. J.; Leone, S. R. J. Phys. Chem. 1987, 91, 3128.
18) Taylor, W. D.; Allston, T. D.; Moscato, M. J.; Fazekas, G. B.;
4. Conclusions
Kozlowski, R.; Takacs, G. A. Int. J. Chem. Kinet. 1980, 12, 231.
19) Cox, R. A.; Derwent, R. G.; Kearsey, S. V.; Batt, L.; Patrick, K.
G. J. Photochem. 1980, 13, 149.
We have investigated the reactions of CH3 and CH3O with
(
NO2 under a variety of conditions of temperature and pressure
and derived a parametrization of the rate coefficients of these
reactions, which proceed by both pressure-dependent and
pressure-independent pathways. The reaction of CH3 with NO2
was found to have a large contribution from a pressure-
dependent term, even at pressures of less than 10 Torr. The
termolecular rate coefficient in the limit of low pressure is given
(20) Lake, J. S.; Harrison, A. J. J. Chem. Phys. 1959, 30, 361.
(
21) Schneider, W.; Moortgat, G. K.; Tyndall, G. S.; Burrows, J. P. J.
Photochem. Photobiol. 1987, 40, 195.
22) 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
Modelling, No. 11; Jet Propulsion Laboratory, 1997.
(23) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F. J.; Kerr,
0
3
-28
-0.3
6
-1
by k ) (3.2 ( 1.3) × 10 (T/297)
cm s . The rate
J. A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1997, 26, 1329.
(24) Troe, J. J. Phys. Chem. 1979, 83, 114.
(25) McCaulley, J. A.; Anderson, S. M.; Jeffries, J. B.; Kaufman, F.
∞
3
coefficient at the high-pressure limit is given by k ) (4.3 (
-11
-1.2
3
-1
0
.4) × 10 (T/297)
cm s . An extrapolation of our results
Chem. Phys. Lett. 1985, 115, 180.
-
11
to 1 Torr yields a total rate coefficient of close to 2.5 × 10
(26) Biggs, P.; Canosa-Mas, C. E.; Fracheboud, J.-M.; Parr, A. D.;
3
-1
cm s , in good agreement with previous low-pressure studies.
Shallcross, D. E.; Wayne, R. P. J. Chem. Soc., Faraday Trans. 1993, 89,
4163.
The bimolecular reaction channel to give CH3O + NO has a
-
11
3
-1
(27) Frost, M. J.; Smith, I. W. M. J. Chem. Soc., Faraday Trans. 1990,
6, 1751.
rate coefficient of close to 1.9 × 10
cm s . For CH3O +
8
0
NO2 the data are well simulated by k ) (5.3 ( 0.3) ×
5
(28) Frost, M. J.; Smith, I. W. M. J. Chem. Soc., Faraday Trans. 1990,
-
29
-4.4
6
-1
∞
5
-11
86, 1757.
(29) Gl a¨ nzer, K.; Troe, J. Ber. Bunsen-Ges. Phys. Chem. 1974, 78, 182.
30) Yamada, F.; Slagle, I. R.; Gutman, D. Chem. Phys. Lett. 1981, 83,
09.
(31) McCaulley, J. A.; Moyle, A. M.; Golde, M. F.; Anderson, S. M.;
Kaufman, F. J. Chem. Soc., Faraday Trans. 1990, 86, 4001.
1
2
0
(T/297)
cm s and k ) (1.9 ( 0.05) × 10 (T/
-
1.9
3
-1
97)
cm s . The hypothesized formation of significant
(
amounts of CH3 in the reaction of OH with CH3C(O)CH3 could
be confirmed in the present work. A room-temperature branch-
ing ratio of 0.5 was established, though this result is associated
with large uncertainty. Direct measurements of CH3 formation
in the reaction of OH with CH3C(O)CH3 would be most useful
to resolve this issue.
4
(
(
(
32) Park, J.-Y.; Gutman, D. J. Phys. Chem. 1983, 87, 1844.
33) Atkinson, R. Int. J. Chem. Kinet. 1997, 29, 99.
34) 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.
(
35) Jenkin, M. E.; Cox, R. A.; Emrich, M.; Moortgat, G. K. J. Chem.
References and Notes
Soc., Faraday Trans. 1993, 89, 2983.
(
1) Singh, H. B.; O’Hara, D.; Herlth, D.; Sachse, W.; Blake, D. R.;
(36) Davies, K. W.; Green, N. J. B.; Pilling, M. J. J. Chem. Soc., Faraday
Trans. 1991, 87, 2317.
Bradshaw, J. D.; Kanakidou, M.; Crutzen, P. J. J. Geophys. Res. 1994, 99,
1
805.
(37) Baldwin, A. C.; Barker, J. R.; Golden, D. M.; Hendry, D. G. J.
Phys. Chem. 1977, 81, 2483.
(2) Singh, H. B.; Kanakidou, M.; Crutzen, P. J.; Jacob, D. J. Nature
1
995, 378, 50.
(38) Slagle, I. R.; Gutman, D. J. Am. Chem. Soc. 1981, 104, 4741.