Atmospheric oxidation mechanism of methyl pivalate, (CH3)3CC(O)OCH3

Article abstract of DOI:10.1021/jp010308s

Smog chamber/FTIR techniques were used to study the Cl atom and OH radical initiated oxidation of methyl pivalate, (CH₃)₃CC(O)OCH₃, in the presence of NO_x in 700 torr of N₂/O₂ diluent at 296 K. The Cl atom initiated oxidation of methyl pivalate in the presence of 15-600 torr of O₂ and 10-30 mtorr of NO_x in 700 torr total pressure of N₂ diluent at 296 K yielded HCHO, CO, acetone, CO₂ and CH₃OC(O)O₂NO₂. OH radical initiated oxidation of methyl pivalate in air produced acetone in a 51% yield. Flash photolysis-resonance fluorescence techniques were used to measure the rate constant for the reaction of OH radicals with methyl pivalate at 250-370 K. The rate constant showed a weak temperature dependence, increasing at both low and high temperature from a minimum value of about 1.2 × 10^-12 cc/molecule-sec near room temperature. The data obtained were used to formulate a detailed mechanism describing the atmospheric oxidation of methyl pivalate. This mechanism gave predictions of the effects of methyl pivalate on O₃ formation and other measures of reactivity that were in good agreement with results of environmental chamber experiments. It should be used in future airshed models to predict the effects of methyl pivalate emissions on air quality.

Full text of DOI:10.1021/jp010308s

J. Phys. Chem. A 2001, 105, 7225-7235
7225
Atmospheric Oxidation Mechanism of Methyl Pivalate, (CH3)3CC(O)OCH3
T. J. Wallington*
Ford Motor Company, SRL-3083, Dearborn, Michigan 48121-2053
Y. Ninomiya, M. Mashino, and M. Kawasaki
Department of Molecular Engineering, Kyoto UniVersity, Kyoto 606-8501, Japan
V. L. Orkin, R. E. Huie, and M. J. Kurylo
Physical and Chemical Properties DiVision, National Institute of Standards and Technology,
Gaithersburg, Maryland 20899-8381
W. P. L. Carter, D. Luo, and I. L. Malkina
CE-CERT, UniVersity of California, RiVerside, California 92521
ReceiVed: January 26, 2001; In Final Form: May 7, 2001
Flash photolysis-resonance fluorescence techniques were used to measure the rate constant for the reaction
of OH radicals with methyl pivalate, (CH CC(O)OCH over the temperature range 250-370 K. The rate
constant exhibited a weak temperature dependence, increasing at both low and high temperature from a
3
)
3
3
-
12
3
-1 -1
minimum value of approximately 1.2 × 10 cm molecule
spectrum of methyl pivalate was measured between 160 and 500 nm at room temperature. Smog chamber/
FTIR techniques were used to study the Cl atom and OH radical initiated oxidation of (CH CC(O)OCH in
the presence of NO in 700 Torr of N /O
CC(O)OCH
) ) (4.1 ( 0.5) × 10 , k(Cl+(CH
CC(O)OC(O)OH) ) (1.7 ( 0.2) × 10
s
near room temperature. The UV absorption
3
)
3
3
x
2
2
diluent at 296 K. Relative rate techniques were used to measure
-
11
-11
k(Cl+(CH
k(Cl+(CH
3
3
)
)
3
3
3
3
)
3
CC(O)OCH
2
Cl) ) (1.8 ( 0.3) × 10 , and
-
11
3
-1 -1
cm molecule s . The reaction of Cl atoms with
(CH
3
)
3
CC(O)OCH
3
was found to proceed (11 ( 3) % via H-abstraction at the -OCH
CC(O)OCH in the presence of 15-600 Torr of O and 10-30 mTorr of NO
diluent at 296 K gives HCHO, CO, acetone, CO , and CH OC(O)O NO products.
OH radical initiated oxidation of (CH CC(O)OCH in air produces acetone in a yield of 51 ( 6%.
3
site. The Cl atom
initiated oxidation of (CH
3
)
2
3
3
2
x
in
7
00 Torr total pressure of N
2
3
2
2
3
)
3
3
Environmental chamber experiments were performed to quantify the effect of methyl pivalate on ozone
formation under simulated atmospheric conditions. An expression representing the atmospheric oxidation
mechanism of methyl pivalate in computer models of atmospheric chemistry is recommended.
1. Introduction
reaction 1 react with oxygen to give peroxy radicals
Industry consumes multibillion pound quantities of solvents
(
CH ) CC(O)OCH • + O f (CH ) CC(O)OCH O • (2)
each year in a variety of applications including coatings, inks,
adhesives, consumer products, and chemical processing. Regula-
tions that encourage the reduction of solvent emissions are
leading manufacturers to consider chemicals which have lower
environmental impact. The concept of relative reactivity is being
actively pursued for use in formulated products. Methyl pivalate
3 3 2 2 3 3 2 2
•
CH (CH ) CC(O)OCH + O f
2 3 2 3 2
•
O CH (CH ) CC(O)OCH (3)
2 2 3 2 3
Peroxy radicals react with NO, NO2, HO2, and other peroxy
radicals in the atmosphere. Reaction with NOx dominates in
(
2,2-dimethylpropanoic acid methyl ester, (CH3)3CC(O)OCH3)
1
is under consideration for use in a broad number of applications
as a solvent with low ozone formation potential. Methyl pivalate
is a volatile compound (bp 101 °C) and may be released into
the atmosphere during its use. Prior to large scale production
and use of methyl pivalate its reactivity in the atmosphere needs
to be established.
polluted air masses. Accurate assessment of the environmental
impact of methyl pivalate requires detailed knowledge of its
atmospheric chemistry. Unfortunately, there is little available
information concerning the atmospheric oxidation mechanism
of this compound. To remedy this situation, we report herein
the results of a collaborative investigation of the atmospheric
chemistry of methyl pivalate.
The atmospheric oxidation of methyl pivalate is initiated by
reaction with OH radicals
The investigation comprised of five main elements. First,
measurement of the kinetics of the reaction of OH radicals with
methyl pivalate. Second, measurement of the UV-visible
absorption spectrum of methyl pivalate. Third, determination
of the products of the Cl atom and OH radical initiated oxidation
of methyl pivalate in air. Fourth, quantification of the impact
of methyl pivalate on ozone formation in environmental chamber
experiments. Fifth, construction and validation of a detailed
mechanism of the atmospheric oxidation mechanism of methyl
pivalate.
(
CH ) CC(O)OCH + OH f (CH ) CC(O)OCH • + H O
3 3 3 3 3 2 2
(1a)
(CH ) CC(O)OCH + OH f •CH (CH ) CC(O)OCH +
3
3
3
2
3 2
3
H O (1b)
2
Under atmospheric conditions, the alkyl radicals produced in
*
To whom correspondence should be addressed. Email: twalling@
ford.com
1
0.1021/jp010308s CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/04/2001

7
2
226 J. Phys. Chem. A, Vol. 105, No. 30, 2001
. Experimental Section
Wallington et al.
(4)
Cl + hν f 2 Cl
2
The experimental systems used have been described previ-
ously.
Cl + (CH ) CC(O)OCH f products
(5)
(6)
(7)
(8)
(1)
3
3
3
2
,3,4,5-9
2
.1 Flash Photolysis Resonance Fluorescence and UV
CH ONO + hν f CH O + NO
3 3
Absorption Systems at NIST. A flash photolysis-resonance
fluorescence (FP-RF) apparatus was used to measure k1 over
the temperature range 250-370 K. The reaction cell was a 60
CH O + O f HCHO + HO
3
2
2
3
cm Pyrex sphere thermostated ((1 K) with a fluid circulated
HO
2
+ NO f OH + NO
2
through its outer jacket. Dry argon, argon bubbled through water
thermostated at 276 K, and a reactant/argon mixture were
premixed and flowed through the reaction cell at a total flow
OH + (CH ) CC(O)OCH f products
3 3 3
3
-1
rate of ca. 1.4 cm s , STP. Flow rates of both argon and H2O/
argon were measured to an accuracy of better than 1% using
calibrated mass flow meters. Two different methods were used
to measure the flow of methyl pivalate/argon mixtures into the
reactor: a mass flow meter (calibration accuracy better than
Loss of (CH3)3CC(O)OCH3 and formation of products were
monitored by Fourier transform infrared spectroscopy using an
-
1
infrared path length of 28 m, and a resolution of 0.25 cm .
Infrared spectra were derived from 32 co-added interferograms.
Calibration of (CH3)3CC(O)OCH2Cl, CH3C(O)CH3, HCHO,
CO, and CO2 reference spectra was achieved by expanding
known volumes of these compounds into the chamber. Ultrahigh
purity N2, O2, and air diluent gases were obtained from Michigan
Airgas Corp. All other reagents were obtained from Aldrich
Chemical Co. at purities >99%.
In smog chamber experiments, unwanted loss of reactants
and products via photolysis, dark chemistry, and wall reactions
have to be considered. Control experiments were thus performed
to check for such unwanted losses of (CH3)3CC(O)OCH3,
1
%) and direct measurement of the rate of pressure change in
a calibrated volume. Different methyl pivalate/argon mixtures
0.05% to 1%) were used to verify that the dilution process did
(
not introduce any systematic error into the rate constant
measurement. UV analysis was used to check the composition
and stability of methyl pivalate/argon mixtures prepared mano-
metrically in the glass bulbs. Methyl pivalate concentrations
determined by UV analysis and calculated from reagent pres-
sures agreed to within 1-2%. The total pressure in the reactor
was typically (100.0 ( 0.1) Torr, with a water vapor component
of approximately 0.07%. The pressure in the reactor was
measured using a MKS Baratron manometer with an absolute
accuracy of better than 0.5%. We estimate that the overall
systematic uncertainty associated with measurements of the
reactant concentration does not exceed 4%. Hydroxyl radicals
were produced by the pulsed photolysis (2.5 Hz repetition rate)
of H2O using a xenon flash lamp focused into the reactor. OH
radicals were monitored using their resonance fluorescence near
(
CH3)3CC(O)OCH2Cl, (CH3)3CC(O)OC(O)OH, CH3C(O)CH3,
HCHO, CO, and CO2. No loss (<2%) was observed when these
compounds were left to stand in the chamber in the dark for 10
min and then irradiated using the output from the UV fluorescent
lamps for 10 min. Photolysis and deposition on the chamber-
walls are not important.
2
.3 Ozone Reactivity Environmental Chamber Experi-
ments at CE-CERT. Environmental chamber experiments were
performed to assess the ability of the chemical mechanism to
predict the effects of methyl pivalate on ozone formation and
other measures of reactivity under simulated atmospheric
conditions. Experiments consisted of the simultaneous irradiation
of two mixtures in a dual-reactor environmental chamber at 297
3
08 nm excited by a microwave discharge resonance lamp
focused into the reactor center. The resonance fluorescence
signal was recorded on a computer based multichannel scaler
(
time channel width 100 µs) as a summation of 1000-8000
(
1 K. One reactor contained a reactive organic gas (ROG)-
consecutive flashes. The radical decay signal at each reactant
2
NOx-air mixture “base case” representing a polluted urban
atmosphere. The other reactor contained the same mixture but
with varying amounts of methyl pivalate added. Three different
base case ROG-NOx mixtures were employed.
concentration was analyzed as described by Orkin et al. to
-
1
obtain a first-order decay rate (τ MP).
The absorption spectrum of (CH3)3CC(O)OCH3 was mea-
sured over the wavelength range 160-300 nm at 295 ( 1 K
First, the “Mini-Surrogate” base case consisting of a simple
-component ROG mixture at relatively low ROG/NOx levels
3
using the single beam apparatus described elsewhere. The
3
complete absorption spectrum was constructed from data taken
over several overlapping wavelength ranges for methyl pivalate
pressure between 0.1 and 20 Torr (13 Pa - 2.7 kPa) inside the
where O3 formation is most sensitive to VOC levels. This
mixture provides a sensitive test of the effect of methyl pivalate
5
,6
on overall radical levels. The “Mini-Surrogate” is a mixture
of NO (0.30), NO2 (0.10), n-hexane (0.50), ethene (0.85), and
m-xylene (0.13). Values in parentheses indicate the initial
concentrations in ppm (at 300 K and 740 Torr pressure 1 ppm
1
6.9 ( 0.05 cm absorption cell. The absorption spectrum of
the liquid methyl pivalate was also recorded over the wavelength
range of 240-500 nm using a Varian Cary double beam
spectrophotometer.
1
3
-3
)
2.4 × 10 molecules cm ).
Methyl pivalate, (CH3)3CC(O)OCH3, was obtained from
Aldrich Chemical Co. at a stated purity >99.6% and was
subjected to repeated freeze-pump-thaw cycles before use.
GC analysis confirmed that the sample was free of impurities.
Second, the “High NOx Full Surrogate” base case, which
employed a more realistic 8-component ROG mixture, also at
relatively low ROG/NOx ratios where O3 formation is sensitive
to VOCs was employed. This base case mixture is a closer
3
2.2 FTIR-Smog Chamber System at Ford. All experiments
approximation of atmospheric conditions in terms of O impacts
6
were performed in a 140-liter Pyrex reactor interfaced to a
Mattson Sirus 100 FTIR spectrometer.4 The reactor was
surrounded by 22 fluorescent blacklamps (GE F15T8-BL) used
to photochemically initiate the experiments. The oxidation of
of added methyl pivalate. The initial mixture (average con-
centrations in ppm) was the following: NO (0.27), NO (0.05),
n-butane (0.41), n-octane (0.10), ethene (0.07), propene (0.06),
trans-2-butene (0.06), toluene (0.09), m-xylene (0.09), and
formaldehyde (0.10).
2
(
CH3)3CC(O)OCH3 was initiated by reaction with Cl atoms or
OH radicals generated by photolysis of molecular chlorine, or
methyl nitrite, in 700 Torr of O2/N2 diluent at 296 ( 2 K
Third, the “Low NOx Full Surrogate” base case which
employed the same 8-component ROG mixture as the “High

Atmospheric Oxidation Mechanism of Methyl Pivalate
J. Phys. Chem. A, Vol. 105, No. 30, 2001 7227
NOx Full Surrogate” but with reduced NO and NO2 concentra-
tions of 0.07 and 0.04 ppm, respectively, was employed. This
mixture provides a test of the mechanism under lower NOx
ln([xylene] /[xylene] )
0
t
IntOH )
(II)
t
xylene
kOH
6
conditions. In addition to the experiments with methyl pivalate,
[
xylene]0 and [xylene]t are the initial and time ) t concentrations
a series of control and characterization experiments were carried
out to characterize the conditions of the experiments for
mechanism evaluation.7
xylene
of m-xylene, and kOH
is the rate constant for reaction of
-11
3
-1 -1 17
OH radicals with xylene (2.36 × 10 cm molecule s . ).
The effect of methyl pivalate on the OH radical concentration
is measured by its IntOH incremental reactivity, which is defined
as
,8
All experiments were performed using the CE-CERT “Divid-
able Teflon Chamber” (DTC) consisting of two ∼6000-liter
2
-mil heat-sealed FEP Teflon reaction bags fitted inside an 8′
Test
Base
×
8′ × 8′ framework with two diametrically opposed banks of
IntOHt - IntOHt
IR[IntOH] )
(III)
3
2 Sylvania 40-W BL black lights. The reaction bags were
t
[methyl pivalate]0
interconnected with ports containing fans to exchange the
contents of the bags. Separate fans were employed to mix the
contents within each chamber. The ports were closed and all
fans were turned off during the irradiations. Pure dry air was
produced using an Aadco air purification system and used to
purge the chamber overnight (for 6-9 h) between experiments.
The base case reactants were injected into both reactors and
mixed as described previously,5 with the methyl pivalate added
to one side after the ports connecting the reactors were closed.
The irradiations proceeded for 6 h.
IntOHTest and IntOHBase are the IntOH values measured at time
t in the added methyl pivalate and base case experiment,
respectively. The uncertainties in IntOH and IR[IntOH] were
estimated by assuming an ∼2% imprecision in the measurements
of the m-xylene concentrations.
2.4 Environmental Chamber Modeling Methods. The
ozone reactivity environmental chamber experiments were
modeled using the SAPRC-99 chemical mechanism.18 The
atmospheric oxidation mechanism for methyl pivalate developed
herein reflects improvements in our understanding of the
chemistry of this species and supersedes that given recently by
,9
9,10
The light intensity was measured by NO2 actinometry. The
-
1
average NO2 photolysis rate was 0.162 min . The flexible
design of the reaction bags ensures that dilution due to sampling
was negligible. Ozone and nitrogen oxides were monitored
continuously using commercially available analyzers. Organic
reactants were monitored using gas chromatography with flame
ionization detection. Formaldehyde was monitored using a
1
8
Carter. The chemistry of methyl pivalate and all organic
compounds present initially in the base case experiments was
represented explicitly in the mechanism. Higher molecular
weight organic products were represented using lumped model
species.
diffusion scrubber method.9
,11,12
Model simulations of environmental chamber experiments
require appropriate representations of chamber effects such as
Two measures of the reactivity of methyl pivalate were
considered when evaluating the ability of the chemical mech-
anism to simulate the environmental chamber experiments with
this compound. The first measure of reactivity used was the
change in the quantity ([O3]t - [NO]t) - ([O3]0 - [NO]0), or
1
9
the chamber radical source. These effects were represented
8
,9
as discussed previously. Initial reactant concentrations used
in the simulations were those measured experimentally. Pho-
tolysis rates were calculated using the measured NO2 photolysis
9
rate, the blacklight spectrum, and absorption cross sections and
∆
([O3] - [NO]). This provides a direct measure of the chemical
1
8
quantum yields given by Carter.
processes responsible for ozone formation that is useful even
when O3 is suppressed by high NO levels.1
3,14,15
The incremental
3
. Results
reactivity of methyl pivalate relative to this quantity is given
by
3
.1 Absolute Rate Study of k(OH+Methyl Pivalate). OH
resonance fluorescence decays were analyzed to obtain pseudo
IR[∆([O ] - [NO]) ] )
first-order decay rates for each methyl pivalate concentration.2
3
t
Test
Base
∆
([O ] - [NO])t - ∆([O ] - [NO])t
Figure 1 shows a plot of the pseudo first-order decay rates vs
methyl pivalate concentration at 272 K. Linear least-squares
analysis of the data in Figure 1 gives k1 ) (1.24 ( 0.02) ×
3
3
(I)
[methyl pivalate]0
-12
3
-1 -1
10
cm molecule s . Results obtained at other temperatures
where ∆([O3] - [NO])tTest is the change in ([O3] - [NO])
are listed in Table 1, quoted uncertainties represent the 95%
confidence intervals associated with the statistical analysis. As
discussed in section 2.1, we estimate that possible systematic
errors associated with measurement of the methyl pivalate
concentration contribute an additional 4% uncertainty. The
measured rate constants are plotted in Arrhenius form in Figure
measured at time t in an experiment where methyl pivalate was
Base
added to the base case, ∆([O3] - [NO])t
is the change in
the corresponding base case run, and [methyl pivalate]0 is the
amount of methyl pivalate added. An estimated uncertainty for
IR[∆([O3] - [NO])] was derived by assuming an ∼3%
uncertainty or imprecision in the measurements of O3 and NO
concentrations.
2
.
The methyl pivalate sample had a stated purity >99.6%. Even
The second measure of reactivity used for mechanism
evaluation was the effect of added methyl pivalate on the
integrated hydroxyl radical concentration (IntOH). It is important
that the mechanism be able to predict this quantity because the
effect of methyl pivalate on OH radical levels will affect O3
if present at a level of 0.4%, the impurities that are most likely
to be present in the methyl pivalate sample could not result in
any appreciable overestimation of the measured rate constant.
For example, the rate constant for the reaction between OH and
-
13
3
-1 -1
20
methanol is 8.9 × 10
cm molecule s at T ) 298 K.
6,7,8,16,18
formation rates from all other reactive compounds present.
This would result in a 0.3% error in the rate constant if the
total 0.4% impurity level is assigned to methanol. There is no
information available on reactivity of pivalic acid toward OH.
Nevertheless, we can consider the reactivity of other organic
It was determined from the rate of consumption of m-xylene in
the experiments (m-xylene is consumed only by reaction OH
radicals) using

7228 J. Phys. Chem. A, Vol. 105, No. 30, 2001
Wallington et al.
0- cm molecule s at T ) 298 K.21 At 0.4% impurity level,
12
3
-1 -1
1
this would result in less than 0.7% error.
OH radicals can react with the radical products of reaction 1
or with the relatively stable products that can accumulate in
the reactor because of the multiflash experimental procedure.
As a check on these possibilities, experiments were performed
using different flash energies to increase the initial OH
concentration and, hence, the concentrations of radicals and
stable products. Varying the flash energy by a factor of 9
resulted in variations in the measured rate constant of less than
6
1
%. Nevertheless, the results presented in Table 1 and Figures
and 2 were obtained only from experiments performed at the
lowest flash energy of 0.24 J to minimize any possible error
due to “secondary chemistry”.
There are no other studies of reaction 1 to compare with the
present results. Nevertheless several observations can be made.
First, the rate constant obtained at room temperature for methyl
pivalate is the same as measured for 3,3-dimethyl-2-butanone,
-
12
3
-1 -1 22
(
CH3)3CC(O)CH3, 1.2 × 10
cm molecule s . Second, it
is noticeably higher than the rate constant measured for dimethyl
-
12 23
carbonate, CH3OC(O)OCH3, 0.31 × 10
,
methyl acetate,
-
12 24,25
CH3OC(O)CH3, 0.33 × 10
,
and acetone, CH3C(O)CH3,
These comparisons suggest
Figure 1. Pseudo first-order OH decay rate vs methyl pivalate
concentration from experiments at 272 K (40 data points).
-12
3
-1 -1 26,27
0
.18 × 10 cm molecule s .
that the majority of the reaction occurs at the -C(O)C(CH3)3
site of methyl pivalate (reaction 1b). This conclusion is
consistent with the results of the product studies presented in
section 3.6.
TABLE 1: Kinetic Data for the Reaction of OH with
a
(CH
3
3
) C-C(O)-O-CH
3
1
2
[
(CH
3
)
3
CC(O)OCH
3
]
k
1
× 10
T, K
range, (10 molecule cm-3)
1
4
(cm molecule-1s-1
3
)
Figure 2 shows the unusual temperature dependence for
reaction 1, with a slight increase in the rate constant as the
temperature is either raised or lowered about room temperature.
These changes lie outside of the statistical uncertainty of the
individual points. This behavior is identical to that observed
250
272
298
330
370
0.099-3.06
0.364-2.09
0.167-3.79
0.301-2.88
0.135-2.56
1.37 ( 0.03
1.24 ( 0.02
1.20 ( 0.03
1.22 ( 0.02
1.38 ( 0.03
23
for dimethyl carbonate, CH3OC(O)OCH3, and methyl acetate,
2
4,25
a)Error bars are 95% confidence intervals from the statistical analysis
and do not include potential systematic errors, see text for details.
CH3OC(O)CH3.
Such a temperature dependence may indi-
cate the existence of two reaction pathways, direct hydrogen
abstraction at higher temperature and a reaction proceeding via
an addition complex formation at lower temperature. For the
three above cases, the reactants contain the CH3OC(O) group.
26,27
However, recent studies
demonstrated similar non-Arrhenius
behavior for acetone, CH3C(O)CH3 as well. These observations
support the hypothesis that the CdO group in the reactant is
associated with the reaction pathway involving complex forma-
tion. On the basis of the observed temperature dependencies
one can conclude that both reaction pathways are comparable
near room temperature for all three molecules, methyl pivalate,
dimethyl carbonate, methyl acetate. However, the fact that
methyl pivalate is four times more reactive than dimethyl
carbonate suggests that OH forms a transition complex with
the tert-butyl rather than with methyl side of methyl pivalate,
as discussed above:
Figure 2. Arrhenius plot showing the average values obtained for k
1
3
.2 UV Absorption Spectrum of Methyl Pivalate. Figure
at each temperature. The thick error bars at each temperature represent
the 95% confidence intervals from statistical analysis only. The fine
lines represent total errors that include the estimated systematic error
of 4%.
3
shows the gas-phase UV absorption spectrum obtained over
the spectral range 160-300 nm. The position of the long
wavelength absorption band corresponds reasonably to those
of methyl and ethyl acetate. Analysis of the absorption spectra
of liquid samples resulted in the absorption cross sections of
methyl pivalate smaller than 10 cm /molecule at wavelengths
28
acids. The rate constant for reaction of OH with the structurally
similar acid, 2-methylpropanoic acid, (CH3)2CHCOOH, is 2 ×
-22
2

Atmospheric Oxidation Mechanism of Methyl Pivalate
J. Phys. Chem. A, Vol. 105, No. 30, 2001 7229
Figure 3. Ultraviolet absorption spectrum (base e) of gaseous methyl
pivalate at 295 K.
longer than 280 nm. In contrast to ketones and aldehydes,28
methyl pivalate does not absorb radiation at wavelengths
exceeding 290 nm and will not undergo photolysis in the
troposphere.
3
.3 Relative Rate Studies of the Reactions of Cl Atoms
with (CH3)3CC(O)OCH3, (CH3)3CC(O)OCH2Cl, and (CH3)3-
CC(O)OC(O)OH. Prior to investigating the atmospheric oxida-
tion products of methyl pivalate, relative rate experiments were
performed in which the kinetics of reactions (9), (10), and (11)
were measured relative to (12) and (13). Initial concentrations
used were 7-30 mTorr (CH3)3CC(O)OCH3, 5-6 mTorr (CH3)3-
CC(O)OCH2Cl, 4-5 mTorr (CH3)3CC(O)OC(O)OH, 4-5 mTorr
of C2H4, 30-53 mTorr of C2H5Cl, and 75-240 mTorr of Cl2,
in 700 Torr of air, or N2, diluent.
Cl + (CH ) CC(O)OCH f products
(9)
(10)
(11)
(12)
(13)
3
3
3
Cl + (CH ) CC(O)OCH Cl f products
3
3
2
Cl + (CH ) CC(O)OC(O)OH f products
3
3
Cl + C H f products
2
4
Cl + C H Cl f products
2
5
Figure 4. Loss of methyl pivalate (MP), chloromethyl pivalate (MP-
The observed losses of (CH3)3CC(O)OCH3 (MP), (CH3)3CC-
O)OCH2Cl (MP-Cl), and (CH3)3CC(O)OC(O)OH (PA) vs
those of reference compounds in the presence of Cl atoms are
shown in Figure 4. Indistinguishable results were obtained in
N2 and air diluent. Linear least-squares analysis of the data in
Figure 4, give the rate constant ratios listed in Table 2. Values
Cl), and pivalic acid (PA) vs C
of Cl atoms in 700 Torr of N
2
H
5
2 4
Cl (O) and C H (b) in the presence
(
2
at 296 K.
CC(O)OCH3/Cl2/N2 mixtures. Initial concentrations were 0.5
Torr of Cl2 and 15 mTorr (CH3)3CC(O)OCH3 in 700 Torr of
N2. Reaction mixtures were subjected to 3-5 successive
irradiations each having a duration of 0.5-1 s. In such
experiments, CH2Cl(CH3)2CC(O)OCH3 and (CH3)3CC(O)OCH2-
Cl are produced by the following sequence of chain reactions
-
11 29
of k9, k10, and k11 were derived using k12 ) 9.29 × 10
and
-
12 30
k13 ) 8.04 × 10
.
We choose to cite final values of k9, k10,
and k11 which are the averages of the individual determinations
given in Table 2 with uncertainties which encompass the
extremes of the individual determinations. Hence, k9 ) (4.1 (
.5) × 10 , k10 ) (1.8 ( 0.3) × 10 , and k11 ) (1.7 ( 0.2)
-
11
-11
0
×
Cl + (CH
3
)
3
CC(O)OCH
3
f (CH
3
)
3
CC(O)OCH
2
• + HCl
-
11
3
-1 -1
10
cm molecule s .
(9a)
3
.4 Mechanism of the Reaction of Cl Atoms with
Cl + (CH ) CC(O)OCH f •CH (CH ) CC(O)OCH +
(CH3)3CC(O)OCH3. To investigate the mechanism of reaction
3 3
3
2
3 2
3
9, experiments were performed using UV irradiation of (CH3)3-
HCl (9b)

7
230 J. Phys. Chem. A, Vol. 105, No. 30, 2001
TABLE 2: Kinetic Data for Reactions Involving Cl Atoms Measured at 296 ( 2 K
CH CC(O)OCH (CH CC(O)OCH Cl
Wallington et al.
(
3
)
3
3
3
)
3
2
3 3
(CH ) CC(O)OC(O)OH
a
a
a
reference
k
9
/kreference
k
9
k
10/kreference
k
10
k
11/kreference
k
11
C
C
2
H
H
4
0.44 ( 0.03
5.06 ( 0.60
4.09 ( 0.28
4.07 ( 0.48
0.18 ( 0.02
2.33 ( 0.10
1.67 ( 0.19
1.87 ( 0.08
0.18 ( 0.02
2.01 ( 0.10
1.67 ( 0.19
1.62 ( 0.08
2
5
Cl
a
Units of 10-11 cm molecule s-1
3
-1
.
(
CH ) CC(O)OCH • + Cl f (CH ) CC(O)OCH Cl + Cl
3 3 2 2 3 3 2
(14)
•
CH (CH ) CC(O)OCH + Cl f
2 3 2 3 2
CH Cl(CH ) CC(O)OCH + Cl (15)
2
3 2
3
Unlike CH2Cl(CH3)2CC(O)OCH3, (CH3)3CC(O)OCH2Cl is com-
mercially available and the FTIR system can be calibrated for
this species. The yield of (CH3)3CC(O)OCH2Cl provides a
measure of the importance of channel (9a). Figure 5 shows a
plot of the formation of (CH3)3CC(O)OCH2Cl vs loss of (CH3)3-
CC(O)OCH3 following UV irradiation of (CH3)3CC(O)OCH3/
Cl2/N2 mixtures. Filled symbols are the observed data while
the open symbols have been corrected (as described by Meagher
3
1
et al. ) for loss of (CH3)3CC(O)OCH2Cl via reaction with Cl
atoms using k10/k9 ) 0.44 (see previous section). Linear least-
squares analysis of the corrected data in Figure 5 gives a molar
(
2
CH3)3CC(O)OCH2Cl yield of 0.11 ( 0.01 (quoted errors are
standard deviations). We estimate that systematic uncertainties
associated with the calibration of the reference spectra contribute
an additional 20% uncertainty. Propagating this additional
uncertainty we conclude that k9a/(k9a + k9b) ) 0.11 ( 0.03 and,
by inference, k9b/(k9a + k9b) ) 0.89 ( 0.03.
Figure 5. Formation of chloromethyl pivalate vs loss of methyl pivalate
following UV irradiation of methyl pivalate/Cl /N mixtures (filled
symbols are observed data, open symbols have been corrected for
2
2
3
.5 Products of the Cl Initiated Oxidation of Methyl
Pivalate in the Presence of NOx. The oxidation of methyl
pivalate was studied in the presence of NO. Initial concentrations
were 15 mTorr methyl pivalate, 100 mTorr of Cl2, 9-38 mTorr
of NO, 15-600 Torr of O2 in 700 Torr total pressure of N2
diluent at 296 K. As discussed above, the reaction of Cl atoms
with methyl pivalate in the presence of O2 leads to the formation
of the peroxy radicals •O2CH2(CH3)2CC(O)OCH3 and (CH3)3-
CC(O)OCH2O2• in yields of 89 and 11%, respectively. In the
presence of NO, these peroxy radicals react to give the
corresponding alkoxy radicals. Information concerning the
atmospheric fate of these alkoxy radicals can be derived by
studying the products following UV irradiation of methyl
pivalate/Cl2/NO/O2/N2 mixtures. Figure 6 shows IR spectra
acquired before (A) and after (B) a 45 s UV irradiation of a
mixture containing 15.1 mTorr of methyl pivalate, 96 mTorr
of Cl2, and 18 mTorr of NO in 700 Torr of air. Panel C shows
the product spectrum obtained after subtraction of features
attributable to methyl pivalate from panel B. Comparison of
panel C with the reference spectrum of acetone in panel D
indicates the formation of acetone. In a similar fashion HCHO,
CO, CO2, and CH3OC(O)O2NO2 were also observed as carbon
containing products. Figure 7 shows a plot of the observed
formation of acetone vs the loss of methyl pivalate following
the successive UV irradiation of methyl pivalate/Cl2/NO/O2/
N2 mixtures. As seen from Figure 7, independent variation of
secondary reaction with Cl atoms - see text for details).
peroxy radical will react with NO, such reactions are well-known
to proceed via two channels giving an alkoxy radical as the
major product and an organic nitrate as the minor product.32
The resulting alkoxy radical (•OCH2(CH3)2CC(O)OCH3) can
either react with O2 to give an aldehyde or undergo unimolecular
decomposition via elimination of HCHO. The observation of
an acetone yield of 54 ( 4%, independent of [O2], shows that
the later pathway dominates. Elimination of HCHO produces
an alkyl radical which adds O2 to give a peroxy radical which
in turn will react with NO via alkoxy radical and organic nitrate
pathways. The alkoxy radical that is formed (•O(CH3)2CC(O)-
OCH3) can isomerize to give a HO(CH3)2CC(O)OCH2• radical,
or decompose by elimination of either a CH3 group to give
CH3C(O)C(O)OCH3, or a C(O)OCH3 fragment to give acetone.
The observation of acetone product shows that the later pathway
dominates. Using a reference spectrum of CH3C(O)C(O)OCH3
(methyl pyruvate) an upper limit of 7% for the yield of this
species was established. The yield of acetone (54 ( 4%) is
significantly less than the fraction of Cl atom attack that
proceeds at the (CH3)3C- end of the methyl pivalate molecule
(89 ( 3%). We attribute this difference to the formation of
1,33,34
nitrates in peroxy radical + NO reactions
of the •O(CH3)2CC(O)OCH3 radical
and isomerization
[
6
NO] over the range 9-38 mTorr and [O2] over the range 15-
00 Torr had no discernible effect on the observed yield of
RO • + NOf RO• + NO
(16a)
(16b)
2
2
acetone. Linear least-squares analysis of the acetone data in
Figure 7 gives a molar yield of 54 ( 4%.
RO • + NOf RONO
2
2
The observation of acetone as product provides important
information regarding the atmospheric degradation mechanism
of methyl pivalate. As discussed above, 89 ( 3% of the Cl
atom attack proceeds at the (CH3)3C- group. The resulting
Spectroscopic evidence for the formation of organic nitrates was
observed in the form of a residual absorption feature at 1662
cm which is characteristic of the asymmetric NO2 stretch in
-
1

Atmospheric Oxidation Mechanism of Methyl Pivalate
J. Phys. Chem. A, Vol. 105, No. 30, 2001 7231
Figure 7. Formation of acetone vs loss of methyl pivalate following
UV irradiation of methyl pivalate/Cl
show the results using “standard” initial conditions of [methyl pivalate]
15 mTorr, [Cl ] ) 100 mTorr, [NO] ) 18 mTorr, and [O ] ) 147
Torr in 700 Torr total pressure of N diluent at 296 K. Open circles
2 2 2
/NO/O /N mixtures. Filled circles
)
2
2
2
and filled triangles are “standard conditions” but with [NO] decreased
to 9, or increased to 38 mTorr, respectively. The open triangles and
diamonds are the results obtained using “standard conditions” but with
2
[O ] decreased to 15, or increased to 600 Torr, respectively.
Figure 6. IR spectra before (A) and after (B) a 45 s irradiation of a
mixture of 15.1 mTorr of methyl pivalate, 96 mTorr of Cl , and 18
2
mTorr of NO in 700 Torr of air. The consumption of methyl pivalate
was 36%. Panel C shows the product spectrum obtained after subtraction
of features attributable to methyl pivalate from panel B. Panel D shows
the reference spectrum of acetone.
organic nitrates.35 The yield of RONO2 was unaffected by
variation of the [NO] over the range 9-38 mTorr and [O2] over
the range 15-600 Torr showing that under the present experi-
mental conditions the RONO2 product is not formed to any
significant degree by addition of NO2 to the alkoxy radicals.
Using an integrated absorption cross section of (2.5 ( 0.2) ×
35
-17
-1
10
cm molecule (base 10) for the asymmetric NO2 stretch
gives an estimate of 20 ( 5% for the RONO2 yield. The IR
technique is unable to distinguish between the possible organic
nitrates formed in the system, and hence, the yield quoted above
is an overall yield of organic nitrates in the system.
Thus far, we have not addressed the fate of the peroxy radical
generated following Cl atom attack on the -OCH3 end of
methyl pivalate. Only 11% of the reaction of Cl atoms with
methyl pivalate proceeds at the -OCH3 group. By analogy to
the behavior of methyl acetate, we expect 65% of the alkoxy
radicals that are formed to undergo R-ester rearrangement and
Figure 8. Formation of acetone vs loss of methyl pivalate following
3 2 2
UV irradiation of methyl pivalate/CH ONO/NO/O /N mixtures in 700
3
5% to react with O2.36 Hence, we would predict a pivalic acid
Torr of air at 296 K.
yield of 11 × 0.65 ) 7% with the remaining 4% giving (CH3)3-
CC(O)OCHO. Using a calibrated reference spectrum of pivalic
acid an upper limit of 8% was established for this species. A
reference spectrum of (CH3)3CC(O)OCHO is not available.
of air diluent at 296 K. The spectral congestion caused by the
presence of large concentrations of CH3ONO in the chamber
precluded the observation of any carbon containing products
except for acetone. Figure 8 shows a plot of the observed
formation of acetone following UV irradiation of methyl
pivalate/ CH3ONO/NO/air mixtures. Linear least-squares analy-
sis gives an acetone yield of 51 ( 6%. It is interesting that the
acetone yield measured in the OH radical and Cl atom initiated
3
.6 Products of the OH Initiated Oxidation of Methyl
Pivalate in the Presence of NOx. The products of the OH
radical initiated oxidation of methyl pivalate were investigated
by the UV irradiation of mixtures of 15 mTorr methyl pivalate,
7
5-110 mTorr of CH3ONO, and 7 mTorr of NO in 700 Torr

7
232 J. Phys. Chem. A, Vol. 105, No. 30, 2001
Wallington et al.
branchingc
TABLE 3: Atmospheric Oxidation Mechanism for Methyl Pivalate in the Presence of NO
x
reactiona
notesb
1
rxn (%)
total (%)
CH
+ OH f H
+ OH f H
3
C(CH
3
)(CH
3
)C(O)OCH
3
1
1
b
a
2
O + CH
3
C(CH
C(CH
3
)(CH
)(CH
2
•)C(O)OCH
)C(O)OCH •
3
75
25
75
25
2
O + CH
3
3
3
2
radicals formed from reaction at tert-butyl end
3
CH
CH
3
C(CH
C(CH
3
)(CH
)(CH
2
•)C(O)OCH
OO•)C(O)OCH
OC(O)C(CH )(CH
+ CH C(CH )(CH
3
+ O
3
2
f CH
3
C(CH
3
)(CH
2
OO•)C(O)OCH
3
75
3
3
2
1
1
7
8
+ NO f CH
+ NO f NO
3
3
3
)CH
2
ONO
2
2
13
87
10
65
2
3
3
2
O•)C(O)OCH
3
CH
3
C(CH
3
)(CH
2
O•)C(O)OCH
3
1
2
2
9
0
1
+ O
f HCHO + CH
2
f CH
3
C(CH
3
)(CHO)C(O)OCH
)C(O)OCH
+ O f CH C(OO•)(CH
)C(O)OCH
C(CH )(ONO )C(O)OCH
+ CH C(O•)(CH )C(O)OCH
)C(O)OCH
C(O)CH + CH
C(CH )(OH)C(O)OCH
-OC(O) • + O
3
+ HO
2
•
3
4
5
95
3
62
62
3
C(•)(CH
)C(O)OCH
C(OO•)(CH
3
3
CH
CH
3
C(•)(CH
3
3
2
3
3 3
)C(O)OCH
3
3
3
2
2
2
3
+ NO f CH
+ NO f NO
CH C(O•)(CH
f CH
f CH
3
3
2
3
5
6
94
4
58
2
3
3
3
3
3
3
2
2
2
2
2
4
5
6
7
8
3
3
3
OC(O) •
6
7
86
14
50
8
50
8
3
3
2
•
CH
CH
CH
CH
3
3
3
3
2
f CH
3
2
2
2
OC(O)OO•
• + O f CH
C(CH
C(CH
C(CH
3
3
3
)(OH)C(O)OCH
)(OH)C(O)OCH
)(OH)C(O)OCH
2
3
C(CH
3
)(OH)C(O)OCH
2
OO•
OO• + NO f NO
2
+ CH C(CH )(OH)C(O)OCH
3
3
2
O•
8
O•
2
3
3
9
0
1
+ O
f CH
HCO• + O
radicals formed from reaction at methoxy end
2
f CH
C(CH
f CO + HO
3
C(CH
3
)(OH)C(O)OCHO + HO
2
•
8
9
34
66
3
5
19
3
3
)(OH)C(O)OH + HCO•
2
2
•
2
CH
CH
3
C(CH
C(CH
3
)(CH
)(CH
3
)C(O)OCH
)C(O)OCH
2
• + O
OO•
2
f CH
3
C(CH
3
)(CH
3
)C(O)OCH
2
OO•
25
3
3
3
2
3
3
2
3
+ NO f CH
+ NO f NO
3
C(CH
+ CH
)C(O)OCH
C(CH )(CH
)(CH )C(O)OH + HCO•
3
)(CH
C(CH
O•
)C(O)OCHO + HO
3
)C(O)OCH
2
ONO
2
2
13
87
3
22
2
3
3
)(CH )C(O)OCH
3
2
O•
CH
3
C(CH
3
)(CH
3
2
3
3
4
5
+ O
f CH
2
f CH
3
3
3
2
•
8
9
34
66
7
14
3
C(CH
3
3
a
Predicted stable organic products are shown in bold font. b Documentation notes for branching ratios are as follows. See Carter18 for details
concerning the estimates of the alkoxy radical reactions. 1 This work (see section 3.7). 2 This work (see section 3.8). 3 Estimated rate constant for
-
14
-1
-14
3
-1 -1
reaction with O
2
is 6.0 × 10 exp(-0.63 kcal mol /RT) ) 2.08 × 10 cm molecule
s
at 300 K. 4 Estimated decomposition rate constant
1
4
-1
6
-1
is 2.0 × 10 exp(-10.99 kcal mol /RT) ) 1.96 × 10 s at 300 K. 5 Nitrate yield adjusted to fit environmental chamber reactivity data for
1
8
14
-1
3
-1
methyl isobutyrate. 6 Estimated decomposition rate constant is 2.0 × 10 exp(-14.50 kcal mol /RT) ) 5.44 × 10 s at 300 K. 7 Rate
11
-1
2 -1
constant for 1,4-H-shift isomerization is estimated to be 2.4 × 10 exp(-8.49 kcal mol /RT) ) 8.97 × 10 s at 300 K. 8 Estimated rate constant
-
14
-1
-15
3
-1 -1
for reaction with O
2
is 6.0 × 10
exp(-1.45 kcal mol /RT) ) 5.27 × 10
cm molecule
s
at 300 K. 9 The rate constant for the ester
1
0
-1
4
-1
c
rearrangement reaction is estimated to be 8.0 × 10 exp(-8.49 kcal mol /RT) ) 5.24 × 10 s at 300 K. Predicted branching ratios. The “Rxn”
column shows the importance of the reaction relative to the competing reactions of the species or radical. The “Total” column show the importance
of the reaction relative to the overall process.
-
13
experiments is indistinguishable. This suggests that as was the
case for Cl atom attack, the reaction of OH radicals with methyl
pivalate proceeds mainly at the (CH3)3C- group. Taking the
ratio of observed acetone yields, the measured value of k9b/(k9a
+ k1b) ) 0.7 and an overall rate constant k1 ) 7.34 × 10
3 -1 -1
cm molecule s . The value of k1 predicted by the group
additivity method is a factor of 1.6 lower than that which we
measure experimentally (see section 3.1). We conclude that the
group additivity method under-estimates either k1a, or k1b, or
both. Interestingly, use of the branching ratio k1b/(k1a + k1b) )
0.7 estimated by the group additivity method gives a predicted
acetone yield of 46%, which is only slightly lower than the
measured value of 51%. We choose to adopt a branching ratio
k1b/(k1a + k1b) ) 0.75, so that the 51% acetone yield is predicted
more closely. This is well within the uncertainty of the branching
ratio estimate.
Under conditions where NOx is present and O3 formation
occurs, the fate of peroxy radicals is reaction with NO giving
NO2 and alkoxy radicals as major products and alkyl nitrates
as minor product (see reaction 16). Consistent with the 20 (
5% overall nitrate yield estimated from the IR data (see section
3.5) it was found that assuming an overall nitrate yields of
∼18% gives the best model predictions of the ozone reactivity
in environmental chamber data (see section 3.8). Allowing for
the sequential formation of peroxy radicals in the multistep
mechanism, this gives about a 9% nitrate formation branching
ratio in the reactions of NO with the initially formed peroxy
+
k9b) ) 0.89 ( 0.03, and assuming that attack on the -OCH3
end of methyl pivalate does not lead to acetone formation we
can estimate that k1b/(k1a + k1b) ) (0.89 ( 0.03) × (0.51 (
0
.06)/(0.54 ( 0.04) ) 0.84 ( 0.12, and, by inference, k1a/(k1a
+
k1b) ) 0.16 ( 0.12. This estimate is consistent with the
conclusions presented in section 3.1.
.7 Mechanism of the OH Initiated Atmospheric Oxida-
3
tion of Methyl Pivalate. Table 3 shows the detailed mechanism
describing the OH initiated oxidation of methyl pivalate under
atmospheric conditions in the presence of NOx derived in the
present work. This mechanism is based upon the results
presented above with estimation methods used to assess the
relative importance of competing reactions of alkoxy and peroxy
radicals. Table 3 lists the estimated branching ratios and overall
importance of competing processes. The footnotes to Table 3
indicate how the branching ratios were derived.
The branching ratio for the reaction of OH radicals with
methyl pivalate can be estimated using the group additivity
37
method of Kwok and Atkinson. This method predicts k1b/(k1a

Atmospheric Oxidation Mechanism of Methyl Pivalate
J. Phys. Chem. A, Vol. 105, No. 30, 2001 7233
Figure 9. Experimental and calculated concentration time plots for ∆([O
IntOH for the environmental chamber reactivity experiments.
3
] - [NO]) and incremental reactivities relative to ∆([O ] - [NO]) and
3
-
12
cm molecule s-1, based on the 298K measurement
3
-1
radicals in the methyl pivalate reaction. This is very close to
× 10
18
the estimate of Carter of ∼10% nitrate yields in these reactions,
based on the nitrate yields needed to fit model simulations to
environmental chamber data for other compounds.
made in this work.
3.8 Results of Ozone Reactivity Environmental Chamber
Experiments. Six dual-chamber incremental reactivity experi-
ments were carried out for this program, two with each of the
three kinds of base case ROG-NOx systems. Concentration time
plots for ∆([O3] - [NO]) data and the incremental reactivities
relative to ∆([O3] - [NO]) and IntOH are shown in Figure 9.
Figure 10 shows concentration time plots for the two organic
compounds monitored, formaldehyde and acetone. Note that
formaldehyde is formed in the base case experiments but acetone
is not. Because of its relatively low rate of reaction, extremely
large amounts of methyl pivalate had to be added in these
experiments to have a measurable effect on the results (10-30
ppm of methyl pivalate vs. 4-6 ppmC of base ROG mixture).
Figure 9 shows that methyl pivalate has a negative effect on
rates of NO oxidation and O3 formation in the mini-surrogate
experiments and has negative or small effects in the beginning
of the full surrogate experiments, but has positive effects on
O3 at the end of the full surrogate runs. It also has negative
effects on OH radical levels in all experiments. The negative
effect of methyl pivalate on radical levels is ascribed to radical
termination due to nitrate formation in the RO2 + NO reactions,
and CH3OC(O)OONO2 formation. This slows down the reac-
tions of all reactive organics present and thereby inhibits NO
oxidation and O3 formation. However, methyl pivalate’s reac-
tions also convert NO to NO2, which has a positive effect on
NO oxidation and O3 formation rates. The mini-surrogate
experiments are relatively more sensitive to radical inhibition
effects compared to direct NO to NO2 conversions so the overall
effect on ∆([O3] - [NO]) is negative. The full surrogate
experiments are less sensitive to radical inhibition effects than
Formaldehyde, acetone, organic nitrate, and CH3OC(O)•
radicals are predicted to be the major products following OH
attack at the tert-butyl group. Formation of ∼3% CH3C(CH3)-
(CHO)C(O)OCH3 is predicted from the minor reaction of O2
with the initially formed peroxy radical, and formation of ∼3%
CH3C(CH3)(OH)C(O)OCHO and ∼5% CH3C(CH3)(OH)C(O)-
OH is predicted following the isomerization of the CH3C(O•)-
(CH3)C(O)OCH3 radical, which competes with the decompo-
sition forming acetone. Although these products were not
detected, the data obtained in this study does not rule out their
formation in small yields.
In terms of model species used in the SAPRC-99 mecha-
1
8
nism, the overall process for the reactions shown in Table 3
can be represented as follows:
OH + Methyl Pivalate f 0.317 RO2-R• + 0.176 RO2-N•
1.221 R2O2• + 0.507 RCO-O2• + 0.187 CO + 0.631
+
HCHO + 0.034 RCHO + 0.507 ACET + 0.096 MEK + 0.187
RCO-OH
RO2-R• represents peroxy radicals that react to convert NO
to NO2 and form HO2. RO2-N• represents peroxy radicals that
react with NO to form organic nitrates. R2O2• represents extra
NO to NO2 conversions caused by peroxy radicals formed in
multistep mechanisms. RCO-O2• represents the lumped higher
acyl peroxy radical (used for CH3OC(O)OO• in this case).
HCHO represents formaldehyde. RCHO represents the aldehyde
CH3C(CH3)(CHO)C(O)OCH3. ACET represents acetone. MEK
represents the lumped lower reactivity oxygenated products
CH3C(CH3)(CH3)C(O)OCHO and CH3C(CH3)(OH)C(O)OCHO.
Finally, RCO-OH represents the lumped higher organic acid
CH3C(CH3)(CH3)C(O)OH and CH3C(CH3)(OH)C(O)OH. This
mechanism was used in the model simulations discussed in the
following section. The overall rate constant used was k1 ) 1.18
6
to direct NO to NO2 conversions, and in the case of methyl
pivalate the net effect on ∆([O3] - [NO]) is positive.
As shown in Figure 10, the addition of methyl pivalate causes
measurable acetone formation in all experiments. The mecha-

7234 J. Phys. Chem. A, Vol. 105, No. 30, 2001
Wallington et al.
Figure 10. Experimental and calculated concentration time plots for formaldehyde and acetone for the environmental chamber reactivity experiments.
nism predicts that formaldehyde is a coproduct with methyl
pivalate, and consistent with this the methyl pivalate causes an
increase in the formaldehyde yields in the full surrogate runs.
However, the methyl pivalate causes formaldehyde to de-
crease in the mini-surrogate runs. This is because the radical
inhibition caused by methyl pivalate, which is relatively more
important in the mini-surrogate experiments, decreases the rate
of formaldehyde formation from the ethene in the base case
mixture.
4. Implications for Atmospheric Chemistry
We report here a large body of experimental and modeling
data which improve our understanding of the atmospheric
chemistry of methyl pivalate. Photolysis is a negligible atmo-
spheric loss mechanism for methyl pivalate. At a temperature
of 270 K (appropriate for evaluation of atmospheric lifetimes
3
9
with respect to reaction with OH ) hydroxyl radicals react 214
-
12
-15
(
1.24 × 10 /5.8 × 10 ) times faster with methyl pivalate
than with methyl chloroform. Methyl chloroform has an
atmospheric lifetime of 5.9 years with respect to reaction with
OH radicals. Hence, the atmospheric lifetime of methyl pivalate
can be estimated to be approximately 10 days.
3
.9 Results of Mechanism Evaluation Using Ozone Re-
activity Chamber Experiments. Figure 9 shows the results of
the model simulations of the methyl pivalate reactivity results
using the SAPRC-99 mechanism with the methyl pivalate
mechanism given in Table 3. The recommended mechanism
fits all the chamber data well except for run DTC701, where
the reactivity predictions are considered to be less reliable
because the mechanism does not simulate the base case
experiment particularly well (the mini-surrogate experiments
tend to be sensitive to uncertainties and variations in the chamber
The data obtained in this work were used to formulate a
detailed mechanism (see Table 3) describing the atmospheric
oxidation of methyl pivalate. This mechanism was found to give
predictions of the effects of methyl pivalate on O3 formation
and other measures of reactivity that are in good agreement with
results of environmental chamber experiments. We recommend
that this mechanism be used in future airshed models to predict
the effects of methyl pivalate emissions on air quality.
Finally, it is of interest to consider the potential impact of
the atmospheric oxidation of methyl pivalate on the global
acetone budget. The global source strength of acetone is
5
radical source ). Fits to the low NOx full surrogate runs are not
quite as good as those to the higher NOx runs, but no consistent
biases are seen except for a tendency to under predict the
inhibition of IntOH reactivities. However, as discussed else-
where,6 inhibitions of IntOH reactivities in the low NOx
experiments tend to be under predicted for almost all VOCs,
suggesting a problem with the overall base mechanism under
low NOx conditions. Figure 10 shows that the model gives
reasonably good predictions of the qualitative effects of methyl
pivalate on formaldehyde and acetone formation, though it has
a slight bias toward under predicting their yields in some
experiments. Considering the overall performance of the model
,8
40
12
approximately 50 Tg per year (1 Tg ) 10 g). If methyl
pivalate finds large scale industrial use it is likely to be produced
at a level of 0.01-0.10 Tg per year. Methyl pivalate oxidation
gives acetone in a 20% mass yield. Assuming an industrial
production of 0.10 Tg and 100% release, the atmospheric
oxidation of methyl pivalate would perturb the global atmo-
spheric acetone budget by < 0.04%.
in simulating reactivity data for compounds with known
Acknowledgment. We thank Peter Ellis (ExxonMobil
Chemical Company) for his help and advice during the course
of this work.
mechanisms5
,14,18,38
the fits shown in Figures 9 and 10 are good.
The fits of the model simulations of the reactivity data shown
in Figure 9 were obtained after adjusting the nitrate yields in
the reactions of RO2 with NO to minimize the biases in the
model simulations to the data. The model predictions are
sensitive to this parameter. To illustrate this sensitivity, the
dotted lines in Figure 9 show results obtained using a 25%
overall nitrate yield (upper extreme of experimental range given
in section 3.5) instead of the best fit ∼18% yield. Assuming
higher nitrate yields causes significantly lower predicted ∆([O3]
References and Notes
(1) 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, in press.
(
2) Orkin, V. L.; Huie, R. E.; Kurylo, M. J. J. Phys. Chem. 1996, 100,
907.
3) Orkin, V. L.; Huie, R. E.; Kurylo, M. J. J. Phys. Chem. 1997, 101,
9118.
8
(
(
(
4) Wallington, T. J.; Japar, S. M. J. Atmos. Chem. 1989, 9, 399.
5) Carter, W. P. L.; Pierce, J. A.; Luo, D.; Malkina, I. L. Atmos.
-
[NO]) reactivities in the higher NOx full surrogate experiments
and causes a bias toward under predicting it in the lower NOx
full surrogate runs and in the mini-surrogate experiment where
the model gives the better fit to the base case experiment.
EnViron. 1995, 29, 2499.
6) Carter, W. P. L.; Luo, D.; Malkina, I. L.; Pierce, J. A., “Environ-
mental Chamber Studies of Atmospheric Reactivities of Volatile Organic
(

Atmospheric Oxidation Mechanism of Methyl Pivalate
J. Phys. Chem. A, Vol. 105, No. 30, 2001 7235
Compounds. Effects of Varying ROG Surrogate and NOx,” Final report to
Coordinating Research Council, Inc., Project ME-9, California Air
Resources Board, Contract A032-0692, and South Coast Air Quality
Management District, Contract C91323. March 24, 1995. Available at http://
www.cert.ucr.edu/∼carter/absts.htm#rct2rept.
M. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling, EValuation No. 12; JPL Publication 97-4. Jet Propulsion
Laboratory, California Institute of Technology: Pasadena, CA, 1997.
(21) Dagaut, P.; Wallington, T. J.; Liu, R.; Kurylo, M. J. Int. J. Chem.
Kinet. 1988, 20, 331.
(22) Wallington, T. J.; Kurylo, M. J. J. Phys. Chem. 1987, 91, 5050.
(23) Bilde, M.; Møgelberg, T. E.; Sehested, J.; Nielsen, O. J.; Wallington,
T. J.; Hurley, M. D.; Japar, S. M.; Dill, M.; Orkin, V. L.; Buckley, T. J.;
Huie, R. E.; Kurylo, M. J. J. Phys. Chem. 1997, 101, 3514.
(24) Wallington, T. J.; Dagaut, P.; Liu, R.; Kurylo, M. J. Int. J. Chem.
Kinet. 1988, 20, 177.
(25) El Boudali, A.; Le Calve, S.; Le Bras, G.; Mellouki, A. J. Phys.
Chem. 1996, 100, 12 364.
(26) Wollenhaupt, M.; Carl, S. A.; Horowitz, A.; Crowley, J. N. J. Phys.
Chem. A 2000, 104, 2695.
(27) Le Calv e´ , S.; Hitier, D.; Le Bras, G.; Mellouki, A. J. Phys. Chem.
A 1998, 102, 4579.
(28) Calvert, J. G.; Pitts, N. J., Jr. Photochemistry; John Wiley and Sons
Inc.: New York, 1967.
(7) Carter, W. P. L.; Luo, D.; Malkina, I. L., “Environmental Chamber
Studies for Development of an Updated Photochemical Mechanism for VOC
Reactivity Assessment,” Final report to the California Air Resources Board,
the Coordinating Research Council, and the National Renewable Energy
Laboratory, November 26, 1997. Available at http://www.cert.ucr.edu/
∼
carter/absts.htm#rct3rept.
8) Carter, W. P. L.; Luo, D.; Malkina, I. L., “Investigation of
(
Atmospheric Reactivities of Selected Consumer Product VOCs,” Report
to California Air Resources Board, May 30, 2000. Available at http://
www.cert.ucr.edu/∼carter/absts. htm#cpreport.
(9) Carter, W. P. L.; Luo, D.; Malkina, I. L.; Fitz, D., “The University
of California, Riverside Environmental Chamber Data Base for Evaluating
Oxidant Mechanism. Indoor Chamber Experiments through 1993,” Report
submitted to the, U. S. Environmental Protection Agency, EPA/AREAL,
Research Triangle Park, NC.; March 20, 1995. Available at http://
www.cert.ucr.edu/∼carter/absts.htm#databas.
(29) Wallington, T. J.; Andino, J. M.; Lorkovic, I. M.; Kaiser, E. W.;
Marston, G. J. Phys. Chem. 1990, 94, 3644.
(30) Wine, P. H.; Semmes, D. H. J. Phys. Chem. 1983, 87, 3572.
(31) Meagher, R. J.; McIntosh, M. E.; Hurley, M. D.; Wallington, T. J.
Int. J. Chem. Kinet. 1997, 29, 619.
(32) Wallington, T. J.; Dagaut, P.; Kurylo, M. J. Chem. ReV. 1992, 92,
667.
(
10) Zafonte, L.; Rieger, P. L.; Holmes, J. R. EnViron. Sci. Technol.
1
9
1
977, 11, 483-487.
11) Dasgupta, P. K.; Dong, S.; Hwang, H. Atmos. EnViron. 1988, 22,
49.
(
(12) Dasgupta, P. K.; Dong, S.; Hwang, H. Aerosol Sci. Tech. 1990,
2, 98.
(33) Becker, K. H.; Geiger, H.; Wiesen, P. Chem. Phys. Lett. 1991, 184,
256.
(13) Johnson, G. M. “Factors Affecting Oxidant Formation in Sydney
Air,” In The Urban AtmospheresSydney, A Case Study; Carras, J. N.,
Johnson, G. M., Eds.; CSIRO: Melbourne, 1983, pp 393-408.
(34) Atkinson, R.; Aschmann, S. M.; Carter, W. P. L.; Winer, A. M.;
Pitts, J. N., Jr., J. Phys. Chem. 1982, 86, 4563.
(35) Tuazon, E. C.; Aschmann, S.; M.; Atkinson, R. EnViron. Sci. Tech.
1999, 33, 2885.
(
14) Carter, W. P. L.; Atkinson, R. EnViron. Sci. Technol. 1987, 21,
6
70.
(
(
(
(
15) Carter, W. P. L.; Lurmann, F. W. Atmos. EnViron. 1991, 25A, 2771.
(36) Christensen, L.; Ball, J. C.; Wallington, T. J. J. Phys. Chem. A.
2000, 104, 345.
16) Carter, W. P. L. Atmos. EnViron 29, 2513, 1995.
17) Atkinson, R. J. Phys. Chem. Ref. Data 1989, Monograph no 1.
18) Carter, W. P. L., “Documentation of the SAPRC-99 Chemical
(37) Kwok, E. S. C.; Atkinson, R. Atmos. EnViron. 1995, 29, 1685.
(38) Carter, W. P. L.; Lurmann, F. W. Atmos. EnViron. 1991, 25A, 2771.
(39) Spivakovsky, C. M.; Logan, J. A.; Montzka, S. A.; Balkanski, Y.
J.; Foreman-Fowler, M.; Jones, D. B. A.; Horowitz, L. W.; Fusco, A. C.;
Brenninkmeijer, C. A. M.; Prather, M. J.; Wofsy, S. C.; McElroy, M. B. J.
Geophys. Res. 2000, 105, 8931.
Mechanism for VOC Reactivity Assessment,” Report to the California Air
Resources Board, Contracts 92-329 and 95-308, May 8, 2000. Available
at http://www.cert.ucr.edu/∼carter/absts.htm#saprc99.
(19) Carter, W. P. L.; Atkinson, R.; Winer, A. M.; Pitts, J. N., Jr. Int.
J. Chem. Kinet. 1982, 14, 1071.
20) 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,
(40) Singh, H. B.; O’Hara, D.; Herlth, D.; Sachse, W.; Blake, D. R.;
Bradshaw, J. D.; Kanakidou, M.; Crutzen, P. J. J. Geophys. Res. 1994, 99,
1805.
(