Methyl acetate reaction with OH and Cl: Reaction rates and products for a biodiesel analogue

Article abstract of DOI:10.1016/j.cplett.2009.02.066

Relative rate experiments performed in a photochemical reactor at 293 ± 0.5 K and a total air pressure of 980 mbar were used to determine k(CH₃C(O)OCH₃ + Cl) = (1.93 ± 0.27) × 10^-12 cm³ molecule^-1 s

Full text of DOI:10.1016/j.cplett.2009.02.066

Chemical Physics Letters 472 (2009) 23–29
Contents lists available at ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett
Methyl acetate reaction with OH and Cl: Reaction rates and products
for a biodiesel analogue
*
Vibeke F. Andersen, Elna J.K. Nilsson, Solvejg Jørgensen, Ole John Nielsen, Matthew S. Johnson
Copenhagen Center for Atmospheric Research, Department of Chemistry, University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen Ø, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history:
Relative rate experiments performed in a photochemical reactor at 293 0.5 K and a total air pressure of
980 mbar were used to determine k(CH3C(O)OCH3 + Cl) = (1.93 0.27) ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1 and
k(CH3C(O)OCH3 + OH) = (3.18 0.13) ꢀ 10ꢁ13 cm3 moleculeꢁ1 sꢁ1. The product branching ratio for the
Cl-initiated oxidation of CH3C(O)OCH3 was investigated at 293 0.5 K and 980 mbar N2, air, or O2, in
the presence and absence of NO. Products observed were CH3C(O)OCHO, CH3C(O)OH, HC(O)OH, CO,
and CO2. The product branching ratios were dependent on NO, but not on the oxygen partial pressure.
Ó 2009 Elsevier B.V. All rights reserved.
Received 20 August 2008
In final form 23 February 2009
Available online 27 February 2009
1. Introduction
which is present at a concentration of approximately 1 ꢀ 106 radi-
cals cmꢁ3 [9].
The production and use of transportation biofuels has grown
rapidly during the last decade due to the desire to reduce carbon
dioxide emissions and dependence on oil. In addition to the carbon
budget, biofuel production and use can involve emission of green-
house gases such as N2O due to use of fertilizer [1], and the use of
ethanol blends may increase the concentration of atmospheric pol-
lutants including formaldehyde and acetaldehyde [2]. Production
of biodiesel has been commercially important since the beginning
of the 1990’s. From 2001 to 2006 the global production of biodiesel
expanded six fold. In 2006, 73% of global biodiesel was produced in
Europe with rapeseed and sunflower being the major sources [3,4].
Biodiesel consists of methyl esters derived from plant oils,
which are long-chain fatty acids such as methyl oleate [5]. When
used in engines, incomplete combustion leads to the formation of
a variety of chemical compounds in the exhaust. In addition, an
estimated 0.01–0.1% of all transportation fuel is emitted directly
to the atmosphere [6]. Combustion products from biodiesel include
shorter chained methyl esters, aldehydes, and methyl acrylates
[7,8].
In order to assess the environmental impact of biofuels, life cy-
cle analysis of the fuels must be carried out. The last step of the life
cycle involves combustion of the fuel. In the present work, the
atmospheric fate of a short chained methyl ester, used as an ana-
logue to uncombusted biodiesel compounds, has been investi-
gated. An important component of a compound’s environmental
impact is its lifetime in the atmosphere. This can be estimated from
the reaction rate of the compound with key oxidants. In ambient
air the most important reactive species is the hydroxyl radical
Another atmospheric oxidant is the chlorine atom. The most
important source of Cl is photolysis of chlorine containing species
in sea salt aerosols [10]. Reaction mechanisms involving Cl are sim-
ilar to those of the OH-radical, with rate constants 10–100 times
larger than for the corresponding OH-reactions. Therefore, also
due to their ease of production in the laboratory, it is useful to
use Cl-atom initiated reactions as models for OH-radical initiated
reactions. In the atmosphere, oxidation by Cl atoms may be impor-
tant in the marine boundary layer where peak concentrations have
been observed to be 103–106 cmꢁ3 [11].
In this study methyl acetate has been used as a model for bio-
diesel. This short chained compound has a relatively high vapor
pressure for a methyl ester, and is therefore feasible for study in
the laboratory. The reactivity of methyl acetate with OH and Cl
has been studied in the presence of varying concentrations of oxy-
gen and in the presence and absence of NOx. For the Cl reaction the
product distributions given various conditions have been
determined.
There are three previous studies of the rate constant for the
reaction of methyl acetate with Cl published in the literature,
and the details for these are given in Table 2. The product branch-
ing ratio for the reaction has been determined in the presence and
absence of NO with varying oxygen partial pressures [12], and the
temperature dependence of this branching ratio was studied by
Tyndall et al. in the temperature range 253–324 K, in the absence
of NO, and at a total pressure of 910 mbar with varying oxygen par-
tial pressures [13].
There are four previous studies of the rate constant for the reac-
tion of methyl acetate with OH, which are summarized in Table 2.
The product branching ratio of the OH-initiated oxidation of
methyl acetate at 296 K has been determined in the presence of
* Corresponding author. Fax: +45 3532 0322.
E-mail address: msj@kiku.dk (M.S. Johnson).
0009-2614/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2009.02.066

24
V.F. Andersen et al. / Chemical Physics Letters 472 (2009) 23–29
NO and at a total pressure of 910 mbar with varying oxygen partial
pressures by Tyndall et al. [13].
We have undertaken the present study using the new photo-
chemical reactor in Copenhagen to further investigate the reaction
rate and the reaction products. Differences between the results of
the present study and previous studies will be discussed.
were 7–71 ppm and 128–158 ppm, respectively. For the OH exper-
iments reference concentrations were 29–58 ppm. The total pres-
sure of the reaction mixture in each experiment was 980 mbar
and the temperature was 293 0.5 K. For the relative rate experi-
ments, technical air (21% O2 and 79% N2) was used as the back-
ground gas. The product study was done in 980 mbar N2,
980 mbar technical air and 980 mbar O2. The reaction in N2 was
carried out in the absence of NO, the reaction in technical air
was carried out both in the presence and absence of NO. The reac-
tion in O2 was carried out in the presence of NO. This was done in
order to determine the oxygen- and NO-dependence of the reac-
tion yields. In these experiments initial methyl acetate and NO
concentrations were in the range 25–45 ppm.
UV-A light with a maximum emission wavelength of 350 nm
was used to generate Cl atoms from Cl2 gas in the reaction cham-
ber. A concentration of Cl2 of approximately five times the total
concentration of reactants (methyl acetate and reference com-
pounds) was used. One UV-A lamp was used and the photolysis
was performed in the presence of all compounds at time intervals
of 2 s to 5 min. After each photolysis step an IR spectrum was
recorded.
For the generation of OH radicals a mixture of O3 and H2O was
irradiated with four UV-C lamps in the presence of both reactants
in the chamber. O3 was generated from O2 with an ozone genera-
tor, and condensed on a silica gel that was cooled with ethanol and
dry ice to ca. ꢁ70 °C. The initial concentration of ozone and water
vapour was approximately 20 and 10 times the total concentration
of reactants, respectively. The photolysis steps in the OH-experi-
ments were between 5 s and 10 min, and after each photolysis step
an IR spectrum was recorded. Each experiment was performed
twice in order to check reproducibility. The resultant relative rate
plots include data from the two identical experiments. Calibrated
reference spectra of the studied reactant, reference compounds,
and a number of the possible products were recorded in the same
reaction chamber and under the same conditions as the
experiments.
2. Experimental setup
The experimental system used in this work is described else-
where [14] and is summarized here. The system consists of a
100 L quartz reaction chamber sealed in both ends with stainless
steel flanges and surrounded by eight UV-A, sixteen UV-C, and
twelve broadband sun lamps. The system is enclosed in an insulat-
ing box equipped with a temperature control system to ensure sta-
ble (within 0.5 K) temperatures during experiments. Reactants are
introduced into the chamber using a pyrex gas handling manifold
with a calibrated volume. By monitoring the pressure of the cali-
brated volume, the concentrations of the compounds introduced
into the chamber can be determined.
The reactions in the chamber were monitored with a Bruker IFS
66v/s FTIR spectrometer. An absorption path length of 72 m was
obtained using a White type optical system in the reaction cham-
ber. A liquid nitrogen cooled MCT detector with a range of 12000–
850 cmꢁ1 was used, and IR spectra at a resolution of 0.125 cmꢁ1
were obtained by co-adding 64 interferograms.
3. Materials and methods
Methyl acetate was reacted with OH and Cl radicals in the pho-
tochemical reactor, in order to determine the following reaction
rates:
CH3CðOÞOCH3 þ Cl ! products
CH3CðOÞOCH3 þ OH ! products
ð1Þ
ð2Þ
The reaction rates were determined relative to two reference
reactions. In the Cl experiment both references were present at
the same time in each experiment, whereas in the OH experiment
only one reference was present at a time. The reference reactions
were the following:
The reaction rates were determined using the relative rate
method. This method assumes that the investigated reaction is
the only significant loss process for the reactant and references
and that these compounds are not formed in the chamber. The fol-
lowing relation is then valid:
ꢀ
ꢁ
ꢀ
ꢁ
C2H5Cl þ Cl ! products
CH3Br þ Cl ! products
C2H4F2 þ OH ! products
C2H6 þ OH ! products
ð3Þ
ð4Þ
ð5Þ
ð6Þ
½reactantꢂ
½reactantꢂ0t
kreactant
kreference
½referenceꢂ
½referenceꢂ0t
ln
¼
ln
where [reactant]0 and [reference]0 are the initial concentrations of
reactant and reference, respectively, and [reactant]t and [reference]t
are the concentrations of reactant and reference at time t. Plotting
Rate constants for the reference reactions are given in Table 1.
k1 was determined relative to k3 and k4, and k2 was determined
relative to k5 and k6. In order to establish the reliability of the
experimental system, k3 was determined relative to k4 and k5
was determined relative to k6 in separate experiments.
The following chemicals were used: CH3C(O)OCH3(P99%, Rie-
del-de Häen), C2H5Cl (P99.7%, Gerling Holz + CO), CH3Br (99%,
Gerling Holz + CO), C2H6 (P99.95%, Fluka), CH2FCH2F (>99%, [15]).
Initial concentrations of reactants were in the range 9–43 ppm.
For the Cl experiments initial C2H5Cl and CH3Br concentrations
ꢂ
ꢃ
ꢂ
ꢃ
½reactantꢂ
½referenceꢂ
½reactantꢂ0t
½referenceꢂ0t
ln
versus ln
gives a straight line with slope:
kreactant
kreference
[16]. The errors for the reported rate constants are computed
from the standard error on the slope of the relative rate plot. Errors
on the relative rate plots are in the order of 0.5–5%.
4. Relative rate results
To test for possible photolytic breakdown, methyl acetate and
the reference compound in air were irradiated in separate experi-
ments under the same conditions (number of lamps and time span)
as a typical relative rate experiment. The tests were performed
both with one UV-A lamp with a spectral range of 320–380 nm
and with 4 UV-C lamps with emission at 254 nm. The time span
was typically 20 minutes. The tests showed no photolysis of the
compounds. To examine any dark reactions that might occur, mix-
tures of methyl acetate, the two reference compounds and chlorine
in air were left for 10–30 min in the chamber with no lights on. A
Table 1
Rate constants for the reference compounds used in this work.
Reaction
Rate constant
Ref.
(3) C2H5Cl + Cl ? products
(4) CH3Br + Cl ? products
(5) C2H4F2 + OH ? products
(6) C2H6 + OH ? products
(8.7 1.0) ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1
(4.29 0.35) ꢀ 10ꢁ13 cm3 moleculeꢁ1 sꢁ1
(8.78 0.88) ꢀ 10ꢁ14 cmꢁ3 moleculeꢁ1 sꢁ1
(2.27 0.93) ꢀ 10ꢁ13 cm3 moleculeꢁ1 sꢁ1
[28]
[29]
[30]
[31]

V.F. Andersen et al. / Chemical Physics Letters 472 (2009) 23–29
25
Fig. 1b. Relative rate plots for the OH-reaction of methyl acetate at 293( 0.5) K and
a total air pressure of 980 mbar.
Fig. 1a. Relative rate plots for the Cl-reaction of methyl acetate at 293 K and a total
air pressure of 980 mbar.
similar test was performed in the presence of ozone and water va-
pour. No significant evidence of dark reactions was seen.
The quoted error corresponds to two standard deviations and is
computed as the root mean square of the errors on the two rate
constants obtained relative to k3 and k4. The results are summa-
rized in Table 2. The rate constant is in good agreement with the
rate obtained by Christensen et al. [12]. However, it does not agree
with the rate constants determined by Notario et al. and Cuevas
et al. [17,18] (see Table 2).
We note that the rate constant determined using C2H5Cl is in
good agreement with previous work. The rate constant determined
using CH3Br as a reference is lower than the other rates (see Table
2). However, the ratio of k4/k3 is in excellent agreement with the
literature. We therefore suggest that the rate of reaction of methyl
acetate with atomic chlorine may be lower than presented by
Notario et al. and Cuevas et al. [17,18].
4.1. Cl-experiments
For the reaction of methyl acetate with chlorine, C2H5Cl and
CH3Br were used as reference compounds. C2H5Cl concentrations
were determined using the strong absorption peak at 1289 cmꢁ1
and CH3Br concentrations were determined using the absorption
band at 1306 cmꢁ1. The IR spectrum of methyl acetate shows a
strong absorption from the C@O stretching vibration over the
interval 1740–1800 cmꢁ1 which was used to determine the con-
centrations.
The expected ratio of k4 and k3 (Using the rate constants given in
Table 1) is k4=k3 ¼ 0:049ðꢃ0:006Þ:
From the experiment done with only the two references pres-
ent, the following relative rate was obtained: k4=k3 ¼ 0:049
ðꢃ0:002Þ. The relative rate plots for the chlorine reaction of methyl
acetate are shown in Fig. 1a. The following relative rates were ob-
tained: k1=k3 ¼ 0:249ðꢃ0:013Þ; k1=k4 ¼ 3:931ðꢃ0:158Þ.
From the two relative rate expressions the average rate con-
stant of the reaction is
4.2. OH-experiments
For the reaction of methyl acetate with OH, CH2FCH2F and C2H6
were used as reference compounds. The absorption peaks from the
C–H stretching vibrations at 2895 and 2995 cmꢁ1 were used to
determine the concentrations of C2H6 and CH2FCH2F, respectively.
The expected ratio of k6 and k5 (Using the rate constants given
in Table 1) is k6=k5 ¼ 2:585ðꢃ1:091Þ.
k1 ¼ ð1:93 ꢃ 0:27Þ ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1
Table 2
Results for the reactions of methyl acetate with Cl and OH obtained in this study and in earlier studies.
Reaction
k ꢀ 1012 cm3 moleculeꢁ1 sꢁ1
Ref.
(1) CH3C(O)OCH3 + Cl ? products
(1) CH3C(O)OCH3 + Cl ? products
(1) CH3C(O)OCH3 + Cl ? products
(1) CH3C(O)OCH3 + Cl ? products
(1) CH3C(O)OCH3 + Cl ? products
(1) CH3C(O)OCH3 + Cl ? products
(2) CH3C(O)OCH3 + OH ? products
(2) CH3C(O)OCH3 + OH ? products
(2) CH3C(O)OCH3 + OH ? products
(2) CH3C(O)OCH3 + OH ? products
(2) CH3C(O)OCH3 + OH ? products
(2) CH3C(O)OCH3 + OH ? products
(2) CH3C(O)OCH3 + OH ? products
(2.17 0.23)
(1.69 0.14)
(1.93 0.27)
(2.85 0.35)
(2.2 0.3)
This study (relative to k3)
This study (relative to k4)
This study
[17]
[12]
[18]
(2.73 0.30)
(0.322 0.003)
(0.315 0.013)
(0.318 0.013)
(0.341 0.029)
(0.342 0.009)
(0.385 0.015)
(0.170 0.050)
This study (relative to k5)*
This study (relative to k6)
This study
[19]
[20]
[21]
[22]
*
Error derived using the standard error of the relative rate plot (two sigma) and does not include systematic error or the error in the rate of k5.

26
V.F. Andersen et al. / Chemical Physics Letters 472 (2009) 23–29
CH3C(O)OCH3
OH
CH3C(O)OCH2
CH2C(O)OCH3
O2
O2
.OOCH2C(O)OCH3
CH3C(O)OCH2OO
HO2
NO
CH3C(O)OCH2O
α-rearr.
½ CH3C(O)OCHO +
½CH3C(O)OCH2OH
OCH2C(O)OCH3
O2
HOOCH2C(O)OCH3/
HOCH2C(O)OCH3
O2
CH3C(O)OCHO
+ CH3C(O)OH
HCO
O2
HO2 + CO
CHOC(O)OCH3
hv
.OH
CHO + .C(O)OCH3
(O)C(O)OCH3
O2
CO + C(O)OCH3
OOC(O)C(O)OCH3
NO
CO2 + CO2 + HCHO
Fig. 2. The proposed mechanism for the OH initiated oxidation of methyl acetate [12,13].
From the experiment done with only the two references pres-
The possible OH initiated breakdown mechanism for methyl
acetate is illustrated in Fig. 2. As mentioned, Cl-initiated oxidation
is expected to proceed via a similar mechanism.
The reaction with Cl proceeds via hydrogen abstraction from
one of the C–H bonds in the molecule, leading to either of the
two radicals generated by reactions (7) and (8):
ent, the following relative rate was obtained: k6=k5 ¼ 2:615ðꢃ
0:059Þ
Together with the experiments done with the two reference
reactions for the chlorine reaction these experiments demonstrate
the reliability of the experimental system.
The relative rate plots from the reaction of methyl acetate with
OH shown in Fig. 1b were used to determine the following relative
rates: k2=k5 ¼ 3:667 ꢃ 0:015; k2=k6 ¼ 1:385 ꢃ 0:029.
From the two relative rate expressions the average rate constant
of the reaction is
CH3CðOÞOCH3 þ ClÅ ! CH3CðOÞOCHÅ þ HCl
ð7Þ
ð8Þ
2
CH3CðOÞOCH3 þ ClÅ ! ÅCH2CðOÞOCH3 þ HCl
H-abstraction from the alkoxy-end has been observed to be the
more probable of the two reactions [12,13]. This was confirmed
by the present study. In the absence of oxygen two products were
observed, one being primary and the other secondary. Spectra of
both compounds are shown in Fig. 3. Comparison with calibrated
reference spectra recorded at the same conditions as for the reac-
tants showed that the two products could not be attributed to di-
methyl succinate (CH2O(O)C(CH2)2C(O)OCH3), ethylene glycol,
diacetate (CH3C(O)O(CH2)2O(O)CCH3) or chloro acetic acid
(ClCH2C(O)OCH3). The most likely explanation is, therefore, that
the primary product is CH3C(O)OCH2Cl formed by reaction (9),
and the secondary product is CH3C(O)OCH(Cl)2 formed from the pri-
mary product by reaction (10) and (11) [12,13].
k2 ¼ ð3:18 ꢃ 0:13Þ ꢀ 10ꢁ13 cm3 moleculeꢁ1 sꢁ1
The quoted error corresponds to two standard deviations.
The results obtained using the two reference compounds agree
with each other to within the uncertainty limits. All results includ-
ing those from previous studies are summarized in Table 2.
The rate constant determined for reaction (2) is in good agree-
ment with the rate constants determined by Wallington et al., El
Boudali et al., and Smith et al. [19–21] for this reaction (see Table
2). The rate constant determined by Campbell and Parkinson is sig-
nificantly lower than the rate constants determined in all other
studies including the one determined in the present study [22].
CH3CðOÞOCHÅ þ Cl2 ! CH3CðOÞOCH2Cl þ ClÅ
ð9Þ
ð10Þ
ð11Þ
2
CH3CðOÞOCH2Cl þ ClÅ ! CH3CðOÞOCHðꢄÞCl þ HCl
CH3CðOÞOCHðꢄÞCl þ Cl2 ! CH3CðOÞOCHðClÞ2
5. Results for the product studies
The IR spectra of the proposed primary product CH3C(O)OCH2Cl
and the secondary product CH3C(O)OCHCl2 were calculated at the
B3LYP/6-31+G(d) level of theory for cis and trans optimised geom-
etries of the chlorinated group relative to the carbonyl group. The
calculations showed that the carbonyl vibration of the secondary
product will shift to higher frequency and have a reduced intensity
relative to the primary product. This calculation, at a rather basic
level of theory, fits the experimental observation very well and fur-
The reaction of methyl acetate with Cl was studied in the pres-
ence and absence of oxygen and NO. The following four experi-
ments were performed:
(1) Methyl acetate + Cl in 980 mbar N2.
(2) Methyl acetate + Cl in 980 mbar air.
(3) Methyl acetate + Cl in 980 mbar air, containing NO.
(4) Methyl acetate + Cl in 980 mbar O2, containing NO.

V.F. Andersen et al. / Chemical Physics Letters 472 (2009) 23–29
27
2CH3CðOÞOCH2OOÅ ! CH3CðOÞOCHO þ CH3CðOÞOCH2OH
2CH3CðOÞOCH2OOÅ ! 2CH3CðOÞOCH2OÅ þ O2
ð14Þ
ð15Þ
ð16Þ
ther supports our assignment. The calculated spectra are available
as electronic supplementary material. In addition to these two spe-
cies there are other Cl-substituted analogues of CH3C(O)OCH3 that
will contribute to the measured peaks.
CH3CðOÞOCH2OOÅ þ NO ! CH3CðOÞOCH2OÅ þ NO2
The peak from the primary product has a shoulder that is pre-
served, but decreases, as the secondary product is formed. A possi-
ble explanation is the formation of another primary product,
ClCH2C(O)OCH3 according to reaction (12). Comparison with a cal-
ibrated spectrum of this compound did not however confirm its
presence. The shoulder is therefore most likely due to other chlori-
nated compounds.
The alkoxy radicals formed by reaction (15) and (16) can either
rearrange to form acetic acid (CH3C(O)OH) and the formyl radical
(ÅCHO) (reaction (17)), or they can react with O2 to form acetic for-
mic anhydride (CH3C(O)OCHO) (reaction (18)).
CH3CðOÞOCH2OÅ ! CH3CðOÞOH þ ÅCHO
ð17Þ
ð18Þ
CH3CðOÞOCH2OÅ þ O2 ! CH3CðOÞOCHO þ HOÅ
2
CH3CðOÞOCHÅ þ Cl2 ! ClCH2CðOÞOCH3 þ ClÅ
ð12Þ
2
The only products observed in the reaction of methyl acetate
with Cl in 980 mbar air and in the absence of NO were acetic
formic anhydride, CO, and CO2, which excludes reaction (14).
The formation of acetic formic anhydride very likely arises from
the oxy radicals formed in reaction (15), and the formation of
the product was observed to correlate linearly with the decrease
in methyl acetate concentration (see Fig. 4). The formation of
acetic formic anhydride and CO account for 85.5 9.2% of the
loss of methyl acetate. The molar yield of CO was 23.2 3.3%.
Since three CO molecules can be formed from one methyl ace-
tate molecule, this accounts for 7.7 1.1% of the total methyl
acetate loss. The 77.8 9.3% yield of acetic formic anhydride is
in agreement to within experimental error with the results from
Christensen et al., 66 7% [12]. The quoted errors correspond to
two standard deviations. Wall reactions creating CO may also be
present but their contribution could not be quantified in the
present work. The experiment shows that hydrogen abstraction
occurs primarily from the alkoxy end of the molecule, supporting
the assumption that the primary product formed in the absence
of oxygen is CH3C(O)OCH2Cl.
Apart from CH3C(O)OCH(Cl)2, formation of other secondary
products such as ClCH2C(O)OCH2Cl and (Cl)2C(O)OCH3 is possible
by reactions similar to reactions (10) and (11). A shoulder seen
on the secondary product peaks could be due to either of these
compounds.
In the presence of oxygen, peroxy radicals will be formed by
addition of O2 to the radicals formed by reaction (7) and (8),
illustrated in reaction (13) for the radical generated by reaction
(7):
CH3CðOÞOCHÅ þ O2 ! CH3CðOÞOCH2OOÅ
ð13Þ
2
These peroxy radicals can either rearrange to form anhydrides
and hydroxyl compounds (reaction (14)), or they can form alkoxy
radicals by self reaction (reaction (15)) or by reaction with NO
(reaction (16)) [12,13].
In the presence of NO, formic acid, acetic acid and formaldehyde
were produced as well as acetic formic anhydride, CO and CO2.
Acetic acid can be formed via the a-ester rearrangement from alk-
oxy radicals generated by either reaction (15) or (16). Since the
compound was only formed in the presence of NO, it is concluded
that it is formed by the alkoxy radicals generated via reaction (16).
In the studies done by Christensen et al. and Tyndall et al. acetic
acid was generated to a higher extent when NO was present, con-
firming the assumption that oxy radicals generated by reaction
with NO are more likely to undergo
a-ester rearrangement than
oxy radicals generated by the self reaction (15) [12,13]. Christen-
sen et al. explained this phenomenon by the formation of excited
alkoxy radicals on reaction with NO (reaction (16)). These excited
radicals would have enough energy to overcome the exothermic
barrier for the a-ester rearrangement [12]. Acetic acid was not ob-
served in the absence of NO in the present study. The acetic acid
yield in the study by Christensen et al. was 11 3%. An estimate
of the lower limit for detection of CH3COOH in our system corre-
sponds to 0.05 ppm or a yield of ca. 0.5%. Tyndall et al. used Cl2 ini-
tial concentrations that were 2–3 times higher and NO
concentrations that were 2 times higher than the initial concentra-
tions of methyl acetate whereas Christensen et al. used initial Cl2
concentrations that were 1–25 times the concentration of methyl
acetate, and NO initial concentrations that were equal to or half
of the concentrations of methyl acetate [12,13]. Initial concentra-
tions of methyl acetate in the two experiments were similar to
the ones used in the present study. Thus, the observed differences
were not caused by differences in the concentrations of reactants
used in the different studies.
A few percent of formaldehyde were formed, either through
hydrolysis of acetic formic anhydride or reactions involving form-
aldehyde produced by the decomposition of alkoxy radicals
according to (19) [12].
Fig. 3. Spectra obtained for the reaction of methyl acetate with Cl in 980 mbar N2 at
293 K. The first spectrum was obtained before UV-irradiation. The second spectrum
was obtained after 34 s of irradiation and shows the primary product. The last
spectrum was obtained after 17.5 min irradiation and shows the secondary product.

28
V.F. Andersen et al. / Chemical Physics Letters 472 (2009) 23–29
CH3CðOÞOCH2OÅ ! CHÅ þ CO2 þ HCHO
Formaldehyde can be converted to formic acid by processes involv-
ing the walls, OH, HO2 or RO2, but note that HO2 will be suppressed
in the presence of NO [23].
In 980 mbar oxygen with NO, the same products were ob-
served as in 980 mbar air with NO. In both experiments loss of
methyl acetate was observed to correlate linearly with the for-
mation of acetic formic anhydride, acetic acid, and formic acid
(shown in Fig. 4). Molar yields of acetic acid in the two experi-
ments were 36.8 3.0% and 38.0 3.0%, respectively. The molar
yields of acetic formic anhydride in the two experiments were
42.8 4.0% and 50.0 6.0%, respectively, while those for formic
acid were 3.7 0.7% and 6.0 1.0%, respectively. Quoted errors
ð19Þ
correspond to two standard deviations. Branching ratios for the
3
formation of acetic formic anhydride and acetic acid were not
significantly different for the two experiments, and therefore it
can be concluded that there is no oxygen dependence of the
a-ester rearrangement in the pressure range used in the present
experiments.
The formation of the mentioned products account for
78.9 7.4% of the total loss of methyl acetate. Apart from the ob-
served products, unknown primary products that absorb at 1250,
1350, 1675 and 1700 cmꢁ1, and secondary products that absorb
at 1050 and 1850 were observed. The absorption peaks at 1350
and 1675 cmꢁ1 are most likely to be due to N-O stretching vibra-
tions. The intensity of these peaks were observed to increase as
Fig. 4. Correlations between the formation of products formed and methyl acetate loss for the reaction in 980 mbar air in the absence (left) and presence (right) of NO.
Fig. 5. Residual spectra of the reaction of methyl acetate with Cl in 980 mbar air containing NO, after subtraction of all known products. The first spectrum was recorded after
5 min of irradiation, and the last spectrum was recorded after irradiation of 1 h and 10 min. Peaks attributed to the formation of NOCl and NO2 are indicated.

V.F. Andersen et al. / Chemical Physics Letters 472 (2009) 23–29
29
the peak from NO2 decreased. This is illustrated in the residual
Acknowledgements
spectra in Fig. 5. One of the primary products could be CH3C(O)O-
CH2ONO2 formed from addition of NO to peroxyl radicals [12]. The
yield of acetic acid and acetic formic anhydride in the experiments
performed in the presence of NO do not change over the range of
O2 concentrations used (206 to 980 mbar), which is consistent with
the work of Christensen et al. [12].
The authors thank the Copenhagen Center for Atmospheric Re-
search, sponsored by the Danish Natural Science Research Council
and the Villum Kann Rasmussen Fund, for its generous support.
Furthermore, we thank Dr. Timothy Wallington and the reviewers
for helpful advice.
6. Implications for atmospheric chemistry
Appendix A. Supplementary material
The degradation of methyl esters in the atmosphere is primarily
initiated by hydroxyl radicals because of the combination of high
reactivity with the concentration of this radical. Reaction with
NO3 as well as photolytic breakdown are additional possible re-
moval processes for methyl acetate. These processes have been
studied by Langer et al. and Benayada et al., and both were found
to be negligible compared to reaction with OH [24,25]. Atmo-
spheric lifetimes of the esters can, therefore, be determined from
their reaction rates with the hydroxyl radical. From the rate con-
stants determined in this study and an OH concentration of
1 ꢀ 106 molecule cmꢁ3, the following lifetime is expected for
methyl acetate:
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.cplett.2009.02.066.
References
[1] P.J. Crutzen, A.R. Mosier, K.A. Smith, W. Winiwarter, Atmos. Chem. Phys. 8
(2008) 389.
[2] M.Z. Jacobsen, Environ. Sci. Technol. 41 (2007) 4150.
[3] K. Bendtz, EU-25 – Oilseeds and products: Biofuels situation in the European
Union 2005, GAIN report, Washington DC: USDA and Foreign Agricultural
Service, 2005.
[4] Worldwatch Institute, Biofuels for transport, Earthscan, London, 2007.
[5] E. Adamska, T. Cegielska-Taras, Z. Kaczmarek, L. Szala, J. Appl. Genet. 45 (4)
(2004) 419.
[6] T.J. Wallington, E.W. Kaiser, J.T. Farrell, Chem. Soc. Rev. 35 (2006) 335.
[7] J.R. Pedersen, Å. Ingemarsson, J.O. Olsson, Chemosphere 11 (1999) 2467.
[8] J.P. Szybist, J. Song, M. Alam, A.L. Boehman, Fuel Process. Technol. 88 (2007) 679.
[9] R. Prinn et al., Science 269 (5221) (1995) 87.
s
ðCH3CðOÞOCH3 þ OHÞ ¼ 29—46 days
The relatively long lifetime and the fact that esters have low sol-
[10] J. Stutz, A.A. Ezell, B.J. Finlayson-Pitts, J. Phys. Chem. A 102 (1998) 8510.
[11] H.B. Singh et al., J. Geophys. Res. 101 (1996) 1907.
[12] L.K. Christensen, J.C. Ball, T.J. Wallington, J. Phys. Chem. A 104 (2000) 345.
[13] G.S. Tyndall, A.S. Pimentel, J.J. Orlando, J. Phys. Chem. A 108 (2004) 6850.
[14] E.J.K. Nilsson, C. Eskebjerg, M.S. Johnson, Atmos. Environ., 2009, doi:10.1016/
j.atmosenv.2009.02.034.
ubility in the aqueous phase [12,26] imply that long distance trans-
port of methyl acetate will be significant.
In the marine boundary layer where reactions with atomic chlo-
rine can be important, the expected lifetime for methyl acetate at
peak chlorine concentrations of 106 radicals cmꢁ3 is:
[15] T.J. Wallington, M.D. Hurley, J.C. Ball, T. Ellermann, O.J. Nielsen, J. Seehested, J.
Phys. Chem. 98 (1994) 5435.
ðCH3CðOÞOCH3 þ ClÞ ¼ 4:5—9 days
[16] K.L. Feilberg, M.S. Johnson, C.J. Nielsen, J. Phys. Chem.
7393.
A 108 (2004)
Since this is based on a maximum Cl concentration, removal by
Cl is contingent on specific local circumstances. The fate of formal-
dehyde formed by the oxidation of methyl acetate includes photol-
ysis and reaction with OH, Cl, or other radicals.
[17] A. Notario, G. Le Bras, A. Mellouki, J. Phys. Chem. A 102 (1998) 3112.
[18] C.A. Cuevas, A. Notario, E. Martínez, J. Albaladejo, Atm. Env. 39 (2005)
5091.
[19] T.J. Wallington, P. Dagaut, R. Liu, M.J. Kurylo, Int. J. Chem. Kinet. 20 (1988)
177.
The acids formed react slowly with oxidants in the atmo-
sphere. Hence, the most likely fate of these is wet or dry depo-
sition or long distance transport. In the atmosphere acetic formic
anhydride will react rapidly with water to form formic and ace-
tic acid [27].
Methyl acetate reacts slowly with OH radicals, making atmo-
spheric impact by products possible far from the emission source.
Reactivity with OH radicals increases with increasing chain length
of the ester because of the presence of a greater number of hydro-
gens. For longer chained methyl esters, reaction is expected to be
increasingly faster making long distance transport less important.
Products for longer chained saturated methyl esters are expected
to resemble the products observed in the present study. The pres-
ent study serves as a model for product branching ratios for longer
chained methyl esters.
[20] A. El Boudali, S. Le Calvé, G. Le Bras, A. Mellouki, J. Phys. Chem. 100 (1996)
12364.
[21] D.F. Smith, C.D. McIver, T.E. Kleindienst, Int. J. Chem. Kinet. 27 (1995) 453.
[22] I.M. Campbell, P.E. Parkinson, Chem. Phys. Lett. 53 (2) (1978) 385.
[23] E. Nilsson, M.S. Johnson, F. Taketani, Y. Matsumi, M.D. Hurley, T.J. Wallington,
Atmos. Chem. Phys. 7 (2007) 1.
[24] S. Langer, E. Ljungström, I. Wängberg, J. Chem. Soc., Faraday Trans. 89 (1993)
425.
[25] B. Benayada, D.O. Leila, Desalination 126 (1999) 83.
[26] F. Cavalli, I. Barnes, K.H. Becker, T.J. Wallington, J. Phys. Chem. A 104 (2000)
11310.
[27] B.J. Finlayson-Pitts, J.R. Pitts, Chemistry of the Upper and Lower Atmosphere,
Academic Press, San Diego, 2000.
[28] J. Shi, T.J. Wallington, E.W. Kaiser, J. Phys. Chem. 97 (1993) 6184.
[29] K.G. Kambanis, Y.G. Lazarou, P. Papagiannakopoulos, J. Phys. Chem. A 101
(1997) 8496.
[30] J.W. Wilson, A.M. Jacoby, S.J. Kukta, L.E. Gilbert, W.B. DeMore, J. Phys. Chem. A
107 (2003) 9357.
[31] S.P. Sander et al., JPL Publication 06-2 (2006).