FTIR spectroscopic study of the OH-induced oxidation of two linear acetates: Ethyl and n-propyl acetates

Article abstract of DOI:10.1039/b101704g

OH-induced oxidation mechanisms of ethyl and n-propyl acetates have been investigated at room temperature (298 ± 5 K) and atmospheric pressure by photolysing CH₃ONO/acetate/NO mixtures with FTIR spectroscopy as analytical device. The main oxidation products and their yields were as follows: from ethyl acetate, acetic acid (0.75 ± 0.13), acetoxyacetaldehyde (0.15 ± 0.05), acetic anhydride (0.02 ± 0.01), formic acetic anhydride (0.02 ± 0.01) and peroxyacetyl nitrate (PAN); from n-propyl acetate, acetoxyacetaldehyde (0.22 ± 0.06), formic acetic anhydride (0.28 ± 0.03), acetic acid (0.15 ± 0.02), acetaldehyde (0.35 ± 0.10), peroxypropionyl nitrate (PPN) and probably acetoxypropionaldehyde (0.30 ± 0.10). From these data, oxidation schemes of these two acetates were elucidated. This study reveals in particular the specifc reactivity of acetates by confirming the novel α-ester rearrangement proposed recently by Tuazon et al. (J. Phys. Chem. A, 1998, 102, 2316) and then by showing that oxygenated alkoxyl radicals may not follow the same rules of reactivity as other alkoxyl radicals. This last observation shows the necessity for further experiments to understand the influence of the oxygenated function on alkoxyl reactivity.

Full text of DOI:10.1039/b101704g

View Article Online / Journal Homepage / Table of Contents for this issue
FTIR spectroscopic study of the OH-induced oxidation of two linear
acetates: ethyl and n-propyl acetates
Benedicte Picquet-Varrault,* Jean-Francois Doussin, Regine Durand-Jolibois and
Patrick Carlier
L aboratoire Interuniversitaire des Systemes Atmospheriques, UMR-CNRS 7583, Universites de
Paris 7 et Paris 12, 61 Avenue du General de Gaulle, 94010 Creteil Cedex, France
Received 22nd February 2001, Accepted 26th April 2001
First published as an Advance Article on the web 5th June 2001
OH-induced oxidation mechanisms of ethyl and n-propyl acetates have been investigated at room temperature
(298 ^ 5 K) and atmospheric pressure by photolysing CH ONO/acetate/NO mixtures with FTIR
3
spectroscopy as analytical device. The main oxidation products and their yields were as follows: from ethyl
acetate, acetic acid (0.75 ^ 0.13), acetoxyacetaldehyde (0.15 ^ 0.05), acetic anhydride (0.02 ^ 0.01), formic
acetic anhydride (0.02 ^ 0.01) and peroxyacetyl nitrate (PAN); from n-propyl acetate, acetoxyacetaldehyde
(0.22 ^ 0.06), formic acetic anhydride (0.28 ^ 0.03), acetic acid (0.15 ^ 0.02), acetaldehyde (0.35 ^ 0.10),
peroxypropionyl nitrate (PPN) and probably acetoxypropionaldehyde (0.30 ^ 0.10). From these data,
oxidation schemes of these two acetates were elucidated. This study reveals in particular the speciÐc reactivity
of acetates by conÐrming the novel a-ester rearrangement proposed recently by Tuazon et al. (J. Phys. Chem.
A, 1998, 102, 2316) and then by showing that oxygenated alkoxyl radicals may not follow the same rules of
reactivity as other alkoxyl radicals. This last observation shows the necessity for further experiments to
understand the inÑuence of the oxygenated function on alkoxyl reactivity.
a-ester rearrangement), as shown in Scheme 1. This reaction
1. Introduction
has also been observed by Christensen et al.10 and Cavalli et
Acetates are emitted into the atmosphere by several anthropic
and natural sources. Natural sources are mainly vegetation
and biomass combustion.1 Concerning anthropic sources,
ethyl and n-propyl acetates in particular are widely used as
solvents in paints and adhesives. Moreover, ethyl acetate is
emitted during the combustion of esteriÐed rapeseed oil used
recently as a substitution fuel.2 Acetates are also formed in the
atmosphere since they are the oxidation products of several
ethers used as automobile fuel additives (methyl tert-butyl
ether, ethyl tert-butyl ether, tert-amyl methyl ether, . . .).3h5
In order to evaluate the environmental impact of these
compounds, the kinetic and mechanistic knowledge of their
oxidation by OH, their main diurnal sink, is needed. Concern-
ing kinetic data on the oxidation of ethyl and n-propyl ace-
tates, several studies have been carried out,6h8 leading to
accurate rate constants and consequently to a good know-
ledge of their tropospheric lifetimes. Moreover, mechanistic
data are rare and often incomplete. Concerning ethyl acetate,
only one mechanistic study has been carried out, by Tuazon et
al.9 Those authors have identiÐed the main oxidation product,
acetic acid, with a molar formation yield of 0.96 ^ 0.08. They
explain its formation by a novel rearrangement of the alkoxyl
radical CH C(O)OCH(O~)CH proceeding by an H-atom
al.11 for methyl acetate and methyl propionate.
In the presence of NO this rearrangement leads to the for-
x
mation of peroxyacetyl nitrate (PAN), which has been
observed in signiÐcant amounts during their experiments. The
hypothesis of this rearrangement is reinforced by the facts
that the overall reaction is calculated to be exothermic by D5
kcal mol~1 and the potential intermediate radical
CH C~(OH)OC(O)CH is estimated to be more stable than
3
3
the CH C(O)OCH(O~)CH one by the di†erence in the OÈH
3
3
and CÈH bond dissociation energies of D10È15 kcal mol~1.
Other oxidation channels such as decomposition or reaction
with O of the CH C(O)OCH(O~)CH radical are possible but
seem to be negligible: no other product has been observed
and their acetic acid formation yield is unity within the experi-
mental uncertainties.
2
3
3
Kerr and Stocker12 have studied the mechanism of the
atmospheric oxidation of n-propyl acetate. Within this study,
the authors have observed the formation of acetaldehyde and
propionaldehyde and explain it by the classical channels of
VOC oxidation shown in Scheme 2.
They have not detected any formation of acetic acid sug-
gesting the existence of the a-ester rearrangement (see Scheme
3). However this is not surprising since acetic acid is unlikely
to be eluted from the gas chromatographic columns used
3
3
transfer through a Ðve-membered ring transition state (called
Scheme 1
DOI: 10.1039/b101704g
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606
This journal is ( The Owner Societies 2001
2595

View Article Online
Scheme 2
(10% PEG on Chromosorb W-HP). Moreover, it must be
noted that the carbon balance in these experiments is incom-
plete and the sum of their yields does not exceed 0.5. Conse-
quently, further experiments are needed to elucidate the
OH-induced oxidation mechanism of n-propyl acetate.
Concerning the other acetates, many kinetic studies have
already been carried out8 but mechanistic data are still rare:
two studies have recently been published on the oxidation
mechanisms of methyl, ethyl, isopropyl and tert-butyl ace-
tates.9,10
solution of H SO to a mixture of NaNO and methanol.14
2
4
2
The initial concentrations of the reactants (acetate, methyl
nitrite and NO) were in the ppm range in a synthetic mixture
of 80% N and 20% O . All experiments were conducted at
2
2
298 ^ 5 K with continuous irradiation (for 1 or 2 h) and were
monitored by acquiring infrared spectra about every 5 min
(corresponding to 100 co-added interferograms) with
a
resolution of 0.7 cm~1 and a path length between 96 and
186 m.
Infrared spectra of reactants and products were calibrated
by Ñushing known amounts of the vapor of these compounds
(determined by measuring the pressure into a calibrated 0.55
L Pyrex bulb) into the chamber. To avoid errors due to the
formation of dimer in the vapor phase, acetic acid was cali-
brated by introducing weighed amounts of the liquid which
was vaporized into the chamber. Integrated band intensities of
the main infrared absorption band of acetates and their oxida-
tion products are presented in Table 1.
In order to improve our understanding of the reactivity of
acetates we report here the results of a mechanistic study of
the OH-induced oxidation of two linear acetates: ethyl and
n-propyl acetates.
2. Experimental section
Methyl and ethyl nitrates, formic acetic anhydride
[CH C(O)OCHO], propionic acetic anhydride [CH C(O)-
Apparatus and methods
3
3
OC(O)C H ], acetoxyacetaldehyde [CH C(O)OCH CHO]
Experiments were performed in an evacuable environmental
chamber comprising a Pyrex reactor of 977 L surrounded by
2 ] 40 Ñuorescent “black lampsÏ (Philips TL05 and TL03),
whose emissions are respectively centred on 360 and 420 nm,
and 16 arc lamps, which provide irradiation with a broad
spectral range from 300 to 600 nm. The reactor contains a
multiple reÑection optical system interfaced to an FTIR
spectrometer (Bomem DA8-ME). Details of the environmental
chamber have been described previously by Doussin et al.13
Hydroxyl radicals were generated by photolysing methyl
nitrite:
2
5
3
2
and acetoxyacetone [CH C(O)OCH C(O)CH ], which were
3
2
3
assumed to be oxidation products of ethyl and n-propyl ace-
tates, were synthesized and their reference spectra were
obtained. All other organic reagents were obtained from com-
mercial sources with purity higher than 99%.
Synthesis
Methyl nitrate (or ethyl nitrate) was synthesized following the
procedure of Desseigne15 by slowly adding 0.1 ml of methanol
(99%) (or ethanol) to a mixture containing 0.2 ml of H SO
2
4
(95%), 0.2 ml of HNO and a grain of urea, cooled at 0 ¡C.
CH ONO ] hl ] CH O~ ] NO
3
3
3
After a few minutes, the supernatant layer was separated and
CH O~ ] O ] HCHO ] HO
washed with a solution of 2% Na CO in water to remove
3
2
2
2
3
residual acid.
HO ] NO ] OH ] NO
2
2
Acetoxyacetaldehyde was synthesized in two steps accord-
ing to a procedure adapted from the one of Rothermel and
Hanack.16
and NO was added to limit the formation of nitrate radicals
and ozone. Methyl nitrite was prepared by adding a dilute
Scheme 3
2596
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606

View Article Online
Table 1 Integrated band intensities of the main infrared absorption band of acetates and their oxidation products
Main absorption
band/cm~1
IBIa
/cm molecule~1
Compound
Reference
Ethyl acetate
n-Propyl acetate
1330È1188
1325È1170
1189È1098
1095È1005
1300È1170
1840È1712
1330È1255
1340È1250
(5.23 ^ 0.12) ] 10~17
(5.29 ^ 0.11) ] 10~17
(6.19 ^ 0.17) ] 10~17
(5.99 ^ 0.13) ] 10~17
(3.2 ^ 0.6) ] 10~17
(4.36 ^ 0.24) ] 10~17
2.08 ] 10~17
This work
This work
This work
This work
This work
This work
Ref. 28
Acetic anhydride
Formic acetic anhydride
Acetoxyacetaldehyde
Acetic acid
Peroxyacetyl nitrate
Peroxypropionyl nitrate
2.51 ] 10~17
Ref. 29
a Calculated using the natural logarithm.
First acetoxyacetaldehyde diethylacetal was synthesized by
mixing 15 g of bromoacetaldehyde diethylacetal with 14.5 g of
silver acetate in 40 ml of dimethyl formamide (DMF). The
mixture was stirred and boiled under reÑux until no appre-
ciable amount of bromoacetaldehyde diethylacetal was left in
the solution (around one week). The silver bromide was Ðl-
tered o† and the powder was rinsed with ether. The solution
was washed twice with water to remove DMF and dried with
magnesium sulfate. Then ether was distilled using a rotavapor
on a cold water-bath and the residue was puriÐed using a
chromatographic column Ðlled with silica. The eluant was
10/90 ether/petroleum ether. The success of the synthesis and
the purity of the product were veriÐed by acquiring IR and
NMR spectra.
Then acetoxyacetaldehyde was synthesized by mixing 4 g of
acetoxyacetaldehyde diethylacetal with 5 ml of formic acid for
2 h at 85 ¡C under a nitrogen atmosphere to prevent oxidation
by the oxygen of the air. Then formic acid and ethanol were
distilled under diminished pressure (17 mmHg) at 25 ¡C. The
residue was diluted in 10 ml of ether and washed twice with a
sodium carbonate solution. After the ethereal solution was
dried with magnesium sulfate, ether was evaporated under
vacuum. Because acetoxyacetaldehyde is very unstable, it was
impossible to purify it at the end of this procedure.
are acetic acid, peroxyacetyl nitrate (PAN), acetoxyacetal-
dehyde and to a lesser extent acetic anhydride and formic
acetic anhydride.
By plotting the formation of products vs. loss of ethyl
acetate (Fig. 2), it appears that all products whose plots have
initial slope di†erent from zero are primary products. The
slight decrease of the acetic acid slope during the experiment
indicates that this compound is subject to a secondary loss
process in the chamber. This loss is mainly due to the oxida-
tion of acetic acid by OH but other processes are also suspect-
ed. Indeed, the experimental points for acetic acid were
corrected for secondary reactions with OH radicals using the
expression given by Atkinson et al.18 This equation deÐnes a
factor F to be applied to the measured concentration of acetic
acid in order to obtain the concentration “without lossÏ:
A
k [ k BG
1 [ ([A] /[A] )
H
A
B
t
0
F \
k
([A] /[A] )kB@kA [ ([A] /[A] )
A
t t
0
0
where A is ethyl acetate and B is acetic acid, and k and k are
A
B
the rate constants of the OH-induced oxidation of A and B.
Acetoxyacetone was synthesized in one step following the
same procedure as the previous one (Ðrst step) by reacting
chloroacetone with silver acetate in dimethyl formamide.
Acetic formic anhydride was synthesized following the pro-
cedure of Schijf and Stevens.17 First 56 ml (1 mol) of freshly
distilled acetyl chloride were added rapidly to 68 g (1 mol) of
dried sodium formate in 100 ml of dry tetrahydrofuran. The
mixture was stirred for 24 h at 0 ¡C under a nitrogen atmo-
sphere. After Ðltration, tetrahydrofuran and unreacted acetyl
chloride were removed by distillation at reduced pressure
using a rotavapor. The residue was then distilled under 17
mmHg. The IR spectrum showed that no free acid was present
in the residue.
Acetic propionic anhydride was synthesized following the
same procedure as the previous one, by reacting acetyl chlo-
ride with sodium propionate.
3. Results
Ethyl acetate
To investigate the mechanism of OH-induced oxidation of
ethyl acetate, experiments were performed by photolysing
CH ONO/ethyl acetate/NO mixtures with analyses of reac-
3
tants and products by FTIR spectroscopy. Fig. 1 shows a
residual spectrum acquired after 3000 s obtained by subtrac-
Fig. 1 Residual infrared spectrum obtained after the photolysis of
tion of remaining reactants (ethyl acetate, NO, CH ONO) and
CH ONO/ethyl acetate/NO mixture: (a) residual spectrum obtained
3
3
products arising from CH ONO photolysis (HNO , NO ,
by subtraction of remaining reactants and products arising from
3
3
2
CH ONO , CO, HCHO, HCOOH, HONO, CO ). Remain-
CH ONO photolysis, compared with reference spectra of (b) acetic
3
2
2
3
ing products are attributed to reaction of ethyl acetate. They
acid, (c) acetoxyacetaldehyde and (d) PAN.
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606
2597

View Article Online
Fig. 2 Experimental plots of the formation of oxidation products vs.
loss of ethyl acetate. (The curves are drawn only to guide the eye.)
The following rate constants were used: k \ 8 ] 10~13 cm3
B
molecule~1 s~1 (ref. 19) and k \ 1.7 ] 10~13 cm3
A
molecule~1 s~1 (ref. 8). It was observed that calculation does
not entirely correct the decrease of the acetic acid plot. Conse-
quently, other processes such as losses and/or reactions on the
walls may also occur.
The formation of PAN as a primary product reveals that
acetyl (CH C~O) radicals, which are its precursors, are directly
3
formed from ethyl acetate oxidation. Primary formation of
acetic acid and PAN conÐrms the hypothesis made by Tuazon
et al.9 concerning the novel rearrangement of the alkoxyl
radical CH C(O)OCH(O~)CH described in Scheme 1.
3
3
Because acetyl radicals react with NO and NO , the pro-
2
duction yield of PAN depends directly on the ratio NO/NO .
2
Fig. 3 Experimental kinetic plots of concentrations of reactants and
Moreover, PAN is thermally unstable. So its concentration
products vs. time, following irradiation of
a
mixture of
depends on the temperature. Consequently, its formation yield
is not constant during the experiment as can be seen in Fig. 2.
At the beginning of the experiment, the temperature is 21È
CH ONO/ethyl acetate/NO.
3
22 ¡C and NO/NO ratio decreases rapidly, leading to a
2
strong conversion of acetyl radicals into PAN. At the end of
simpliÐed chemical system described from the previous obser-
vations:
the experiment, the temperature reaches approximately 30 ¡C
and NO/NO ratio increases signiÐcantly (see Fig. 3). So the
destruction of PAN becomes signiÐcant20 and irreversible
[CH C(O)OO~ radicals react with NO to form CH ~ and
2
k1
ethyl acetate ] OH ÈÈÈ acetic acid
(1)
(2)
(3)
3
CO ].
2
3
k2
Acetic acid can also be formed by a b-ester rearrangement
of CH C(O)OCH CH O~ radical (see Scheme 4) but glyoxal
was not observed in the spectra. Therefore, the upper limit of
the branching ratio of this channel was estimated from the
infrared detection limit of glyoxal.
ethyl acetate ] OH ÈÈÈ acetoxyacetaldehyde
3
2
2
k3
ethyl acetate ] OH ÈÈÈ acetic anhydride
Acetoxyacetaldehyde is formed as a primary product by an
k4
ethyl acetate ] OH ÈÈÈ formic acetic anhydride (4)
H-atom abstraction from the ÈCH group of the ethyl func-
3
tion followed by the reaction of CH C(O)OCH CH O~ rad-
3
2
2
icals with oxygen, as shown in Scheme 5.
k5
ethyl acetate ] OH ÈÈÈ products
k6
(5)
Traces of acetic and formic acetic anhydrides have also
been detected as products of ethyl acetate oxidation. They
may be formed by the reactions shown in Scheme 6.
Formic acetic anhydride may also be produced by the pho-
tolysis and/or OH-induced oxidation of acetoxyacetaldehyde
(Scheme 7), but its kinetic plots (Figs. 2 and 3) are not precise
enough to verify this secondary formation deÐnitively.
As has been seen before, most of the oxidation products are
subject to secondary loss processes which are not always well
identiÐed. In the case of acetic acid, for example, it is not pos-
sible to know the rate of these processes (because they may be
heterogeneous loss processes) and so to correct for its disap-
pearance. Therefore, the formation yields of oxidation pro-
ducts were not determined by calculating the slope of the
curves of [products] vs. [*[acetate] but by using numerical
simulation (FACSIMILE}, v 3.0, AEA Technology, UK) of a
acetoxyacetaldehyde ] OH ÈÈÈ
formic acetic anhydride (6)
k7
acetic acid ] OH ÈÈÈ products
(7)
where k to k are adjusted in order to allow the simulated
1
7
curves to Ðt the experimental points. The sum of the
OH ] acetate rate constants is Ðxed to the overall value of
1.7 ] 10~12 cm3 molecule~1 s~1 (ref. 8). Here it is important
to note that this system does not consider the formation of
PAN, which depends on the ratio NO/NO . Consequently it
is more uncertain to describe. However, there is no e†ect on
2
2598
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606

View Article Online
Scheme 4
the result because its formation is already implicitly taken into that is,
account by reaction (1) (see Scheme 1). The results of the com-
parison between simulated curves and experimental points are
shown in Fig. 4.
In order to conÐrm the co-formation of PAN and acetic
acid (Scheme 1), it is necessary to determine the formation
k [NO ]
8
2
r \
k [NO ] ] k [NO]
8
2
9
Then, the upper limit of the formation yield of PAN is given
by
yield of the acetyl CH C~(O) radicals, which should be equal
3
a
\ a
] (1/r)
to that of the acetic acid. This quantity represents the upper
PAN up. lim.
PAN exp.
limit of the yield of PAN, obtained if all peroxyacetyl radicals
where a
is the observed formation yield of PAN, esti-
PAN exp.
react with NO . This value was determined by calculating a
mated graphically from the initial slope of the plot of [PAN]
vs. [*[ethyl acetate]. This calculation was carried out for the
Ðrst three experimental points following the beginning of irra-
diation, for which we can consider that the temperature (21È
22 ¡C) is low enough to disregard the thermal decomposition
of PAN (see Table 2). According to Atkinson et al.21 the
2
factor r, which represents the ratio between the decay of per-
oxyacetyl radicals due to reaction (8) and its total decay,
k8
CH C(O)OO~ ] NO ÈÈÈ CH C(O)OONO
(8)
(9)
3
2
3
2
values of
k
and
k
are respectively 1.1 ] 10~11 and
2 ] 10~11 cm3 molecule~1 s~1.
k9
CH C(O)OO~ ] NO ÈÈÈ CH C(O)O~ ] NO
8
9
3
3
2
Scheme 5
Scheme 6
Scheme 7
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606
2599

View Article Online
residual spectrum acquired at t D 2500 s obtained by subtrac-
tion of remaining reactants (n-propyl acetate, NO, CH ONO)
3
and products arising from CH ONO photolysis. Remaining
3
products, which are attributed to reaction of n-propyl acetate,
are formic acetic anhydride, acetoxyacetaldehyde, per-
oxyacetyl and peroxypropionyl nitrates, acetaldehyde and
acetic acid (these last two products are not visible on the scale
of the Ðgure). In addition an unknown product(s) which
absorbs at 1120 and 1780 cm~1 is observed. These bands are
very close to those of acetoxyacetaldehyde, indicating that its
chemical structure may also be very similar. From the clas-
sical process of VOC oxidation, three compounds may be
formed: acetoxypropionaldehyde [CH C(O)OCH CH CHO],
3
2
2
acetoxyacetone [CH C(O)OCH C(O)CH ] and propionic
3
2
3
acetic anhydride [CH C(O)OC(O)CH CH ] (see Scheme 9).
3
2
3
Fig. 4 Simulated plots (full curves) and experimental points for the
OH-induced oxidation of ethyl acetate.
Therefore, in order to identify this “unknown compoundÏ, syn-
thesis and infrared spectra of acetoxyacetone and propionic
acetic anhydride were performed, but their bands did not
match with the unidentiÐed ones. Consequently it is probable
that the “unknown compoundÏ is acetoxypropionaldehyde.
Our attempt to synthesize it was not successful. So, this
hypothesis remains unveriÐed.
Final formation yields of all oxidation products are present-
ed in Table 3. Indicated errors of yields take into account the
estimated uncertainties in reactant and product IR absorption
band calibration and the uncertainties in the numerical simu-
lation estimated by FACSIMILE by the following procedure:
after having calculated the best value of each constant for the
Ðt, the solver makes some further iterations by slightly
modifying each Ðtted constant in order to estimate the sensi-
tivity of the Ðt to each of them. Using this information, it
calculates Ðnal uncertainties for a conÐdence range of 95%
(see FACSIMILE v3.0 Technical Reference). Errors due to the
quantiÐcation of compounds in infrared spectra have been
disregarded because they are not signiÐcant compared to the
previous ones.
The formation yields of PAN and acetic acid are the same,
conÐrming that they are formed conjointly. Moreover, the
sum of the yields, which is near unity within the experimental
uncertainties, reveals that the main products have been identi-
Ðed. From these data, the complete mechanism of OH-
induced oxidation of ethyl acetate was deduced (see Scheme
8).
By plotting the formation of products vs. loss of n-propyl
acetate (Fig. 6), we can observe that all of them except PAN
are primary products. But they are also secondary products as
revealed by the increase of the slopes during the experiment. It
should be noted that the curve of the unknown compound
was plotted in arbitrary units by following its two infrared
band intensity changes. Primary formation of acetoxyacetal-
dehyde is attributed to the process in Scheme 10.
By considering the classical processes of VOC oxidation
and photolysis, it appears that the only primary product that
can lead to acetoxyacetaldehyde is acetoxypropionaldehyde
(see Scheme 11). Therefore, this observation reinforces the
probability that the unknown product is acetoxypropional-
dehyde. Moreover, the inspection of the plot of [unknown
product] vs. [*[n-propyl acetate] shows that this compound
is subject to secondary loss processes. Primary formic acetic
anhydride is produced jointly with acetaldehyde by the
process described in Scheme 12 while its secondary formation
may be due to photolysis and/or OH-induced oxidation of
acetoxyacetaldehyde.
n-Propyl acetate
Primary formation of acetic acid and peroxypropionyl
nitrate indicates that the a-ester rearrangement of the
CH C(O)OCHO~CH CH radical described in Scheme 3 may
As for ethyl acetate, the mechanism of OH-induced oxidation
of n-propyl acetate was investigated by photolysing
3
2
3
occur. It leads to the formation of CH CH C(O)OO~ radicals,
3
2
CH ONO/n-propyl acetate/NO mixtures and by following
which, depending on the ratio NO/NO , produce PPN
and/or acetaldehyde (see Scheme 13) and, to a lesser extent,
3
2
their evolution with FTIR spectroscopy. Fig. 5 shows a
ethyl nitrate. Because the NO/NO ratio quickly becomes low
2
(see Fig. 7), the formation of PPN is predominant and its con-
centration does not decrease as quickly as observed for PAN
in the previous experiment. In the same way, secondary for-
mation of PAN, which is due to OH-induced oxidation of
acetaldehyde, is signiÐcant until the end of the experiment.
Finally, secondary formation of acetic acid may be due to
the hydrolysis of formic acetic anhydride on the wall of the
reactor (where water vapor may have condensed) and to a
lesser extent to the hydrolysis in the gas phase.
Table 2 Calculation of the theoretical formation yield of PAN
obtained in the absence of NO
Points
[NO]/ppm [NO ]/ppm
r
a
a
2
PAN exp.
PAN up. lim.
1(1455 s) 2.5
2(1720 s) 1.2
3(1990 s) 1.1
2.1
4.1
4.9
0.34 0.25
0.67 0.39
0.73 0.45
0.75
0.58
0.62
Formation yields of oxidation products were calculated
using numerical simulation (FACSIMILE) of the following
simpliÐed chemical system described from the previous obser-
vations:
Table 3 Oxidation products of ethyl acetate and their formation
yields
k10
Products
Formation yields
n-propyl acetate ] OH ÈÈÈ formic acetic anhydride
Acetic acid
0.75 ^ 0.13
0.15 ^ 0.05
0.02 ^ 0.01
0.02 ^ 0.01
0.75 ^ 0.20
(10)
Acetoxyacetaldehyde
Acetic anhydride
Formic acetic anhydride
PAN upper limit
k11
n-propyl acetate ] OH ÈÈÈ acetoxyacetaldehyde (11)
2600
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606

View Article Online
Scheme 8
k12
k15
n-propyl acetate ] OH ÈÈÈ acetoxypropionaldehyde ?
acetoxyacetaldehyde ] OH ÈÈÈ formic acetic anhydride
(15)
(12)
k16
k13
formic acetic anhydride ÈÈÈ acetic acid
(16)
(17)
n-propyl acetate ] OH ÈÈÈ acetic acid
(13)
(14)
k17
acetic acid ] OH ÈÈÈ products
k14
acetoxypropionaldehyde ÈÈÈ acetoxyacetaldehyde
where k to k are adjusted in order to allow the simulated
10 17
curves to Ðt the experimental points. The sum of the
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606
2601

View Article Online
Fig. 5 Residual infrared spectrum obtained after the photolysis of
CH ONO/n-propyl acetate/NO mixture: (a) residual spectrum
3
obtained by subtraction of remaining reactants and products arising
from CH ONO photolysis, compared with reference spectra of (b)
3
formic acetic anhydride, (c) acetoxyacetaldehyde, (d) PAN and (e)
PPN.
Fig. 6 Experimental plots of the formation of oxidation products vs.
loss of n-propyl acetate. (The curves are drawn only to guide the eye.)
OH ] acetate rate constants is Ðxed to the overall value of
3.6 ] 10~12 cm3 molecule~1 s~1 (ref. 8). In this system, forma-
tions of PAN, PPN and acetaldehyde, which depend on the
Table 4 Oxidation products of n-propyl acetate and their formation
yields
ratio NO/NO , were not explicitly taken into account. Con-
cerning the unknown product, which may be acetoxypropion-
Products
Formation yields
2
Acetoxyacetaldehyde
Formic acetic anhydride
Acetic acid
PPN upper limit
Acetaldehyde
0.22 ^ 0.06
0.28 ^ 0.03
0.15 ^ 0.02
0.26 ^ 0.10
0.35 ^ 0.10
0.30 ^ 0.10
aldehyde, the formation yield was calculated by assuming in
the simulation that it accounts for the balance of the n-propyl
acetate loss. Results are shown in Fig. 8 and compared to the
experimental points. The excellent agreement between Ðtted
curves and experimental points reveals that the simulation
describes the chemical system correctly.
Acetoxypropionaldehyde?
Scheme 9
Scheme 10
2602
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606

View Article Online
Scheme 11
The PPN formation yield was determined by the same
method as the one used for PAN in the mechanistic study of
ethyl acetate oxidation. The acetaldehyde one was estimated
using the plot of [acetaldehyde] vs. [*[n-propyl acetate]
corrected for secondary reaction with OH radicals. Because
the formation of CH CH C(O)OO~ should not be taken into
account twice, we did not consider the Ðrst experimental
points for which formation of acetaldehyde due to the reac-
tion of these radicals with NO is signiÐcant (see Scheme 3).
Final formation yields of all oxidation products are present-
ed in Table 4. From these data, the OH-induced oxidation
mechanism of n-propyl acetate was deduced (see Scheme 14).
A comparison of this study with the results of Tuazon et
al.9 reveals that the observed branching ratios of the
rearrangement channel are signiÐcantly di†erent: 0.75 ^ 0.13
against 0.96 ^ 0.08. Moreover, in this study the formation of
another major product, acetoxyacetaldehyde, was observed.
These di†erences are probably due to the fact that the absorp-
tion bands of ethyl acetate and acetoxyacetaldehyde are very
close. Consequently, Tuazon et al.9 may have confused the IR
spectra of these two compounds and so they may have under-
estimated the disappearance of the reactant. Therefore, this
may also explain why they overestimate the formation yield of
acetic acid.
3
2
To a much lesser extent, three other channels were identi-
Ðed which proceed via decomposition and/or reaction with O
of the CH C(O)OCH O~ and CH C(O)OCHO~CH radicals
4. Discussion
2
3
2
3
3
and lead to the formation of formic acetic anhydride and
acetic anhydride. Because two of them lead to the same
product, they could not be distinguished. According to the
mechanism proposed in Scheme 8, some other reactions could
also occur. The products to which they lead were not
observed and the carbon balance is complete. Consequently,
these channels seem to be negligible. Upper limits of their for-
mation yields were estimated from the infrared detection limit
of the products (after taking secondary reactions into
account).
Ethyl acetate
The two main oxidation products of ethyl acetate are acetic
acid and PAN with formation yields of 0.75 ^ 0.13 and
0.75 ^ 0.20 respectively. This observation is consistent with
the a-ester rearrangement of CH C(O)OCHO~CH radicals
reported recently by Tuazon et al.9 Acetic acid could also be
produced by other processes shown in Scheme 4, but the
yields of PAN and acetic acid are the same within experimen-
tal uncertainties, indicating that these processes are negligible.
3
3
Scheme 12
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606
2603

View Article Online
Scheme 13
Primary formation of acetic acid and peroxypropionyl
nitrate was observed indicating that a-ester rearrangement of
CH C(O)OCHO~CH CH (see Scheme 3) may occur. PPN
3
2
3
may also be produced by another process since its formation
yield is higher than that of acetic acid (0.26 ^ 0.10 against
0.15 ^ 0.02). This additional source may be the OH-induced
oxidation of propionaldehyde, produced by decomposition of
CH C(O)OCHO~CH CH . Indeed, even if this compound
3
2
3
was not detected in infrared spectra since it reacts very rapidly
with OH radicals [for the propionaldehyde ] OH reaction,
k \ 2 ] 10~11 cm3 molecule~1 s~1 (ref. 21)], it could be pro-
duced during the experiment. According to FTIR analyses,
any formation of propionaldehyde had a molar yield \0.15
(after taking into account secondary reaction with OH).
Therefore, the primary formation yield of PPN may be about
0.10È0.15 and would be in good agreement with that of acetic
acid.
According to the mechanism proposed in Scheme 14, some
other reactions could also occur but seem to be negligible
because none of the products to which they lead were
observed. The upper limits of their formation yields were esti-
mated from the infrared detection limit of the products
(corrected for secondary reactions).
This study agrees quite well and completes the previous one
performed by Kerr and Stocker12 who identiÐed acetaldehyde
and propionaldehyde as primary oxidation products with for-
mation yields of 0.35 and 0.07 respectively. They also observed
formation of PAN and PPN but, according to the authors,
these nitrates are secondary products whereas this study
reveals that PPN is a primary as well as a secondary product.
Moreover, new oxidation products, i.e. acetoxyacetaldehyde,
formic acetic anhydride, acetic acid and acetoxypropional-
dehyde, were identiÐed in our study.
Fig. 7 Experimental kinetic plots of concentrations of reactants and
products vs. time, following irradiation of a mixture of CH ONO/n-
3
propyl acetate/NO.
n-Propyl acetate
All the main oxidation products of n-propyl acetate were
identiÐed and quantiÐed, except acetoxypropionaldehyde for
which formation is strongly suspected but not deÐnitely
proved. They are formic acetic anhydride, acetic acid, PPN,
acetaldehyde and acetoxyacetaldehyde. Nevertheless it is quite
difficult to elucidate completely its oxidation mechanism since
some of these products are formed by several channels. For
example, formic acetic anhydride and acetaldehyde can be
formed by decomposition of CH C(O)OCHO~CH CH and
Reactivity of linear acetates
From this mechanistic study, some information concerning
mainly the Ðrst and the fourth steps of the VOC oxidation
process, i.e. H-abstraction by hydroxyl radicals and evolution
of alkoxyl radicals, can be deduced.
3
2
3
CH C(O)OCH CHO~CH radicals. The same problem occurs
3
2
3
for acetoxyacetaldehyde (see Scheme 14). Consequently,
several channels cannot be distinguished from each other.
For ethyl acetate, it is shown in Scheme 8 that OH radicals
attack predominantly the ÈCH È function (D80%) and to a
2
lesser extent the ÈCH one (D15%), conÐrming that H-
3
abstraction occurs preferentially at the secondary carbon
instead of the primary carbon. Moreover, H-abstraction on
the acetate function seems to be negligible because none of the
product to which it could lead [acetaldehyde and/or ethyl 2-
oxoacetate, CH(O)C(O)OCH CH ] was observed. From the
2
3
carbon balance, the molar yield of this channel was estimated
to be \0.05. These observations are in good agreement with
previous mechanistic studies carried out on acetates.9,10 From
SAR calculations,22 H-atom abstraction from the ÈCH and
2
ÈCH groups and from the acetate function are estimated to
3
account for 87%, 10% and 3% of the overall reaction respec-
tively. So the SAR estimations are also in agreement with the
results of this study within experimental uncertainties. For n-
propyl acetate, it was not possible to distinguish H-
abstractions on the three carbons of the alkyl function since
they lead to the same products.
Concerning evolution processes of alkoxyl radicals, we con-
Ðrm here the hypothesis made by Tuazon et al.9 that radicals
Fig. 8 Simulated plots (full curves) and experimental points for the
OH-induced oxidation of n-propyl acetate.
2604
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606

View Article Online
Scheme 14
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606
2605

View Article Online
of the type CH C(O)OCH(O~)R proceed predominantly by
an a-ester rearrangement leading to acetic acid and PANs as
alkoxyl radicals and consequently on the nature of the oxida-
tion products.
3
primary products. In particular, we show for the Ðrst time that
this rearrangement occurs with CH C(O)OCH(O~)CH CH
In conclusion, this study reveals in particular the speciÐc
reactivity of acetates, which must be taken into account in
chemical mechanisms for the tropospheric degradation of oxy-
genated VOCs.
3
2
3
radicals. On the other hand, we observe that b-ester
rearrangement, which could potentially proceed by an H-
transfer through a six-membered ring transition state, is a
minor loss process for CH C(O)OCH CH(O~)R radicals
3
2
Acknowledgements
since the main fate of CH C(O)OCH CH O~ and
3
2
2
CH C(O)OCH CHO~CH is reaction with O and decompo-
We thank the French Ministry of the Environment for sup-
porting this research as part of the PRIMEQUAL pro-
gramme. We also gratefully thank Prof. Francoise Heymans
and her collaborators for their help in the synthesis of ace-
toxyacetaldehyde and acetoxyacetone.
3
2
3
2
sition, respectively. This observation is in good agreement
with those made by Tuazon et al.9 for ethyl and isopropyl
acetates.
Alkoxyl radicals other than CH C(O)OCHO~R do not
3
react by a-ester rearrangement. So we studied the branching
ratios of the classic channels: unimolecular decomposition,
References
reaction with O and isomerization. From the oxidation pro-
2
ducts and their formation yields, we Ðrst deduce that primary
1
T. E. Graedel, D. T. Hawkins and L. D. Claxton, Atmospheric
Chemical Compounds: Sources, Occurrence and Bioassay, Aca-
demic Press, Orlando, FL, 1986.
alkoxyl (RÈCH O~), for example CH C(O)OCH CH O~, pro-
2
3
2
2
ceeds predominantly by reaction with oxygen. Here, isomer-
ization and decomposition are minor processes. The
2
3
4
5
6
7
C. Ferrari, PhD Thesis, Universite Joseph FourierÈGrenoble 1,
1995.
predominance of reaction with O was already observed in
2
E. C. Tuazon, W. P. L. Carter, S. M. Aschmann and R. Atkinson,
Int. J. Chem. Kinet., 1991, 23, 1003.
D. F. Smith, T. E. Kleindienst, E. E. Hudgens, C. D. McIver and
J. J. Bufalini, Int. J. Chem. Kinet., 1992, 24, 199.
D. F. Smith, C. D. McIver and T. E. Kleindienst, Int. J. Chem.
Kinet., 1995, 27, 453.
T. J. Wallington, P. Dagaut, R. Liu and M. J. Kurylo, Int. J.
Chem. Kinet., 1988, 20, 177.
A. El Boudali, S. Le Calve, G. Le Bras and A. Mellouki, J. Phys.
Chem., 1996, 100, 12364.
B. Picquet et al., Int. J. Chem. Kinet., 1998, 30, 839.
E. C. Tuazon, S. M. Aschmann, R. Atkinson and W. P. L. Carter,
J. Phys. Chem. A, 1998, 102, 2316.
many cases especially for alkoxyl produced by oxidation of
alkanes.23h25
Concerning n-propyl acetate, FTIR analysis shows that any
formation of acetoxyacetone and propionic acetic anhydride
had a molar yield \0.01 and \0.03 respectively. This lack of
observed products suggests that CH C(O)OCHO~CH CH
3
2
3
and CH C(O)OCH CHO~CH predominantly decompose
3
2
3
instead of reacting with oxygen. Therefore it appears that sec-
ondary alkoxyl radicals [R ÈCH(O~)ÈR ] undergo mainly
8
9
1
2
decomposition. This observation is quite surprising since
atmospheric chemists used to consider that reaction with O
is the dominant fate of alkoxyl radicals. However, it is inter-
2
10 L. K. Christensen, J. C. Ball and T. J. Wallington, J. Phys. Chem.
A, 2000, 104, 345.
11 F. Cavalli, I. Barnes, K. H. Becker and T. J. Wallington, J. Phys.
Chem., 2000, 104, 11310.
esting to note that this observation was made for alkoxyl rad-
icals produced by OH-induced oxidation of ethers.5,26,27
Therefore, it seems that oxygenated functions play a role of
prime importance on the reactivity of alkoxyl radicals and
consequently on the nature of the oxidation products. To
understand this phenomenon, further experiments on oxygen-
ated alkoxyl radicals should be carried out.
12 J. A. Kerr and D. W. Stocker, J. Atmos. Chem., 1986, 4, 263.
13 J. F. Doussin, D. Ritz, R. Durand-Jolibois, A. Monod and P.
Carlier, Analusis, 1997, 25, 236.
14 W. D. Taylor et al., Int. J. Chem. Kinet., 1980, 12, 231.
15 G. Desseigne, Memorial des Poudres, 1948, 30, 58.
16 R. Rothermel and M. Hanack, L iebigs Ann. Chem., 1991, 1013.
17 R. Schijf and W. Stevens, Rec. T rav. Chim., 1966, 85, 627.
18 R. Atkinson, S. M. Aschmann, W. P. L. Carter, A. M. Winer and
J. N. Pitts, Jr., J. Phys. Chem., 1982, 86, 4563.
5. Conclusion
The atmospheric oxidation of ethyl and n-propyl acetates was
studied. The major reaction products are acetic acid, PAN
and acetoxyacetaldehyde from ethyl acetate; and acetoxyace-
taldehyde, formic acetic anhydride, acetic acid, acetaldehyde,
PPN and probably acetoxypropionaldehyde from n-propyl
acetate. From these observations, the oxidation schemes of
these two acetates were elucidated and mechanistic informa-
tion concerning the reactivity of alkoxyl radicals was deduced.
This study reveals in particular the speciÐc reactivity of ace-
tates by conÐrming the novel a-ester rearrangement proposed
recently by Tuazon et al.9 and then by showing that oxygen-
ated functions may play a signiÐcant role on the reactivity of
19 R. Atkinson, J. Phys. Chem. Ref. Data, Monograph 2, 1992, 134.
20 C. Tsalkani, PhD Thesis, Universite Paris 7, 1987.
21 R. Atkinson et al., J. Phys. Chem. Ref. Data, 1999, 28, 191.
22 E. S. C. Kwok and R. Atkinson, Atmos. Environ., 1995, 29, 1685.
23 R. A. Cox, K. F. Patrick and S. A. Chant, Environ. Sci. T echnol.,
1981, 15, 587.
24 L. Batt, Int. Rev. Phys. Chem., 1987, 6, 53.
25 R. Atkinson, Int. J. Chem. Kinet., 1997, 29, 99.
26 T. J. Wallington and S. M. Japar, Environ. Sci. T echnol., 1991, 25,
410.
27 K. Stemmler, W. Mengon and J. A. Kerr, J. Chem. Soc., Faraday
T rans., 1997, 93, 2865.
28 F. Zabel, private communication, 1999.
29 N. Tsalkani and G. Toupance, Atmos. Environ., 1989, 23, 1849.
2606
Phys. Chem. Chem. Phys., 2001, 3, 2595È2606