Product study of the gas-phase reactions of O3, OH and NO 3 radicals with methyl vinyl ether

Article abstract of DOI:10.1039/b315997c

Products of the atmospherically important reactions of methyl vinyl ether were investigated at room temperature and 1 bar of dry air using FTIR spectroscopy and a scanning mobility particle sizer instrument. In the presence of NO_x the reaction with OH radicals was found to lead to the formation of methyl formate (80.9 ± 8.2)% and formaldehyde (76.6 ± 7.9)% as the main products. In the absence of NO_x, the major products were methyl formate (50.2 ± 5.1)% and formaldehyde (57.0 ± 6.0)%. The main products of the reaction of methyl vinyl ether with O₃ were methyl formate (73.5 ± 7.5)%, formaldehyde (27.8 ± 6.6)%, hydroperoxymethyl formate (19.6 ± 6.2)%, methanol (5.4 ± 0.9)%, CO₂ (30.3 ± 6.1)% and CO (14.2 ± 1.5)%, the formation yield of OH radicals in this reaction was estimated to be (14 ± 7)%. The reaction of methyl vinyl ether with NO₃ radicals again led to formation of methyl formate (52.5 ± 6.3)% and formaldehyde (51.4 ± 6.2)% as the main products. All yields given here are on a molar basis, the fraction of carbon recovered in identified products was (79 ± 11)%C in the OH reaction, (88 ± 7)%C in the ozonolysis and (52 ± 9)%C in the NO₃ reaction. The formation of aerosol was observed in the ozonolysis, but not the OH or NO₃ radical reactions. The particle formation in the ozone reaction could be modelled assuming an aerosol consisting of a single component with a mass-based formation yield of (0.66 ± 0.13)% and a partitioning coefficient of (0.98 ± 0.27) m ³ μg^-1. Other products were observed in low yields, and residual FTIR spectra of unidentified products are presented. Simple mechanistic degradation schemes are proposed to explain the observed products, and the results are discussed with regard to their implications for atmospheric processes.

Full text of DOI:10.1039/b315997c

View Article Online / Journal Homepage / Table of Contents for this issue
R E S E A R C H P A P E R
Product study of the gas-phase reactions of O3 , OH and NO3
radicals with methyl vinyl ether
a
a
b
¨
Bjorn Klotz,*y Ian Barnes and Takashi Imamura
a
Bergische Universita¨t Wuppertal, Gauss Strasse 20, D-42097 Wuppertal, Germany.
E-mail: barnes@uni-wuppertal.de; Fax: þ49 202/439 2505; Tel: þ49 202/439 2510
National Institute for Environmental Studies, Atmospheric Environment Division,
16-2 Onogawa, Tsukuba-shi, Ibaraki-ken 305-8506, Japan. E-mail: imamura@nies.go.jp;
Fax: þ81 298/50 2579; Tel: þ81 298/50 2406
b
Received19th December 2003, Accepted17th February 2004
F|rst published as an AdvanceArticle on the web 9th March 2004
Products of the atmospherically important reactions of methyl vinyl ether were investigated at room
temperature and 1 bar of dry air using FTIR spectroscopy and a scanning mobility particle sizer instrument. In
the presence of NOx the reaction with OH radicals was found to lead to the formation of methyl formate
(80.9 ꢀ 8.2)% and formaldehyde (76.6 ꢀ 7.9)% as the main products. In the absence of NOx , the major products
were methyl formate (50.2 ꢀ 5.1)% and formaldehyde (57.0 ꢀ 6.0)%. The main products of the reaction of
methyl vinyl ether with O3 were methyl formate (73.5 ꢀ 7.5)%, formaldehyde (27.8 ꢀ 6.6)%, hydroperoxymethyl
formate (19.6 ꢀ 6.2)%, methanol (5.4 ꢀ 0.9)%, CO2 (30.3 ꢀ 6.1)% and CO (14.2 ꢀ 1.5)%, the formation yield of
OH radicals in this reaction was estimated to be (14 ꢀ 7)%. The reaction of methyl vinyl ether with NO3 radicals
again led to formation of methyl formate (52.5 ꢀ 6.3)% and formaldehyde (51.4 ꢀ 6.2)% as the main products.
All yields given here are on a molar basis, the fraction of carbon recovered in identified products was
(79 ꢀ 11)%C in the OH reaction, (88 ꢀ 7)%C in the ozonolysis and (52 ꢀ 9)%C in the NO3 reaction. The
formation of aerosol was observed in the ozonolysis, but not the OH or NO3 radical reactions. The particle
formation in the ozone reaction could be modelled assuming an aerosol consisting of a single component with a
mass-based formation yield of (0.66 ꢀ 0.13)% and a partitioning coefficient of (0.98 ꢀ 0.27) m3 mgꢁ1. Other
products were observed in low yields, and residual FTIR spectra of unidentified products are presented. Simple
mechanistic degradation schemes are proposed to explain the observed products, and the results are discussed
with regard to their implications for atmospheric processes.
1. Introduction
2. Experimental
Recent field and modelling studies1,2 have shown that there are
significant discrepancies between modelled and measured tro-
pospheric concentrations of solvents. These can be explained
either by inadequate emission databases or by discrepancies
in our understanding of the degradation chemistry of these
classes of compounds, which are mostly oxygenated hydrocar-
bons like ethers. Vinyl ethers are a class of ether that is being
used increasingly in manufacturing, as solvents and are being
considered as petrol additives.3 This and the volatility of the
simple vinyl ethers will result in significant emissions of these
compounds into the atmosphere. A better understanding of
the atmospheric chemistry of this class of compounds is there-
fore highly desirable.
Until now the atmospheric chemistry of vinyl ethers has not
been studied to a significant extent, only a few reaction rate
constants are known4–6 or have been estimated.7 Product stu-
dies are available only for ethyl vinyl ether.6
In the present study the products of the reactions of the sim-
=
plest vinyl ether, methyl vinyl ether (CH3–O–CH CH2), with
2.1 Reaction chamber
All experiments were performed in the evacuable and bakeable
reaction chamber at the National Institute for Environmental
Studies (NIES) in Tsukuba, Japan. The chamber has a volume
of 6.065 m3 and a surface/volume ratio of 3.7 mꢁ1. The light
source is a solar simulator consisting of 19 high pressure Xe
arc lamps with a power rating of 1 kW each. The lamps were
operated at 800 W and their light was filtered through 2 mm
Pyrex glass filters (l ꢂ 290 nm) or T-1130 quartz filters
(Toshiba, l ꢂ 220 nm). Using the Pyrex glass filter photolysis
frequencies of J(NO2) ¼ 4.25 ꢃ 10ꢁ3 sꢁ1 and J(CH3ONO) ¼
8.7 ꢃ 10ꢁ4 sꢁ1 were determined, with the quartz filters these
were J(NO2) ¼ 5.4 ꢃ 10ꢁ3 sꢁ1 and J(CH3ONO) ¼ 1.9 ꢃ 10ꢁ3
s
ꢁ1. A more detailed description of the chamber is given
elsewhere.8
An FTIR spectrometer Nicolet Nexus 670 was used to deter-
mine gas-phase concentrations of reactants and products. The
spectrometer is coupled to a White-type mirror system with an
optical absorption path of 221.5 m. An SMPS system (Scanning
Mobility Particle Sizer, TSI model 3934) was used to determine
aerosol size distributions. The instrument was connected to the
chamber through a stainless steel tube (diameter: 1/4 in, length:
60 cm). Aerosol mass concentrations are calculated from the
mobility diameters determined by this instrument assuming
the most important atmospheric reactants OH, NO3 and O3
have been investigated. To our knowledge this represents the
first product study on methyl vinyl ether reactions under atmo-
spheric conditions.
spherical shape and a density of the aerosol of 1 g cmꢁ3
Reactants are transferred into the chamber by pre-mixing
the desired partial pressures in calibrated external bulbs, using
.
y Present address: Cognis Deutschland GmbH & Co KG, Henkel-
strasse 67, D-40551 Dusseldorf, Germany; E-mail: bjoern.klotz@
¨
cognis.com
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4
1725

View Article Online
pure N2 as the bath gas. The reactants are then flushed into the
chamber with pure N2 . Pressures are determined with a capa-
citance manometer.
package (version 5.1) supplied by Nicolet. The reference spec-
tra were obtained by vaporising the pure compound into a
bulb of known volume, the concentration was determined with
a capacitance manometer. The contents of the bulb were then
flushed into the chamber with pure N2 . Table 2 gives an over-
view of the absorption cross sections determined.
2.2 Experimental conditions
Three experiments each were conducted on the reactions of
methyl vinyl ether with OH radicals and ozone, and two
experiments on its reaction with NO3 radicals. The initial con-
ditions of all the experiments are listed in Table 1.
In two of the OH radical reaction experiments (MVH1,
MVH2) the photolysis (Pyrex filters) of CH3ONO/NO mix-
tures in air was used as the source of OH radicals, in the third
experiment (MVH3) photolysis (quartz filters) of H2O2 pro-
vided NOx-free conditions. In all three OH radical experiments
the reactants were pre-mixed in the chamber, opening the shut-
ters of the solar simulator started the reaction.
2.4 Corrections applied to concentrations
The concentrations measured cannot be evaluated directly in
some systems, when reactions other than those studied become
significant. This is mainly the case in the reaction of alkenes
with ozone, in which OH radicals are known to be formed.11
To determine the amount of OH formed, a tracer hydrocarbon
(di-n-butyl ether) that reacts exclusively with OH radicals was
added in the ozonolysis experiments. The OH concentration is
then given by eqn. (1):
ꢀ
ꢁ
In the first of the ozonolysis experiments (MVO1), ozone
was added to pre-mixed methyl vinyl ether and di-n-butyl ether
ln ½tracerꢄt =½tracerꢄt
ðtn ꢁ tnꢁ1ÞktracerþOH
Here, [tracer]t and [tracer]t are the concentrations of the tra-
nꢁ1
n
½OHꢄ
¼
ð1Þ
(kdi-n-butyletherþOH,298 ¼ 2.81 ꢃ 10ꢁ11 cm3 moleculeꢁ1
s
ꢁ1),9
tnꢁtnꢁ1
which was added as a tracer substance for OH radicals. Due
to the very fast reaction of methyl vinyl ether with ozone,
the other two experiments (MVO2, MVO3) were conducted
by multiple additions of between 1.2 and 2.2 ꢃ 1012 molecule
cmꢁ3 of methyl vinyl ether to ozone/di-n-butyl ether. The
chambers fans were used during the addition to ensure homo-
geneous mixing, but were switched off while spectra were
recorded.
nꢁ1
n
cer compound at times tnꢁ1 and tn , respectively, and ktracerþOH
is the rate constant of the reaction of the tracer with OH
radicals. The amount of a reactant (methyl vinyl ether,
kOH ¼ 6.4 ꢃ 10ꢁ11 cm3 moleculeꢁ1 sꢁ1 7
) reacted with OH is
then given by eqn. (2):
OH
D½reactantꢄt ꢁt ¼ ½reactantꢄt
ꢁ
n
nꢁ1
nꢁ1
The experiments on the reaction of methyl vinyl ether with
NO3 radicals were conducted by step-wise addition of methyl
vinyl ether to N2O5 , similar to the ozonolysis experiments.
N2O5 was synthesised in situ prior to each experiment by react-
ing an excess of NO2 with ozone. Due to the slow nature of this
reaction some ozone was still present when the experiment was
started by adding methyl vinyl ether, however, the amount was
low enough to ensure that > 99% of the methyl vinyl ether
reacted with NO3 .
In all experiments the FTIR spectrometer was operated at
180 co-added interferograms per spectrum, with a delay time
of 270 s between spectra. A total of 10 to 40 spectra with 1
cmꢁ1 resolution were recorded per experiment. The SMPS
instruments scan times were adjusted to match those of the
FTIR.
!!!
½tracerꢄt
kreactantþOH
nꢁ1
½reactantꢄt exp
ꢁ
ln
nꢁ1
ktracerþOH
½tracerꢄt
n
ð2Þ
This equation can be used for any second order loss process
and is also used for the correction of product concentrations
for secondary losses.
When a product is formed in two reaction channels (i.e., O3
and OH), the amount of the product formed in the OH reac-
tion can be calculated by eqn. (3):
OH
OH
product
reactantþOH
D½productꢄt ꢁt ¼ D½reactantꢄt ꢁt
Y
nꢁ1
ð3Þ
n
nꢁ1
n
To correct the raw data for first order secondary reactions of
products, e.g. photolysis or dilution rates, eqn. (4) is used:
All experiments were conducted at a pressure of (760 ꢀ 10)
Torr and a temperature of (298 ꢀ 2) K.
firstꢁorder
D½reactantꢄt ꢁt
¼ ½reactantꢄt ꢁ
n
nꢁ1
nꢁ1
ꢀ
ꢁ
ð4Þ
½reactantꢄt expðꢁðtn ꢁ tnꢁ1Þkfirstꢁorder
Þ
nꢁ1
2.3 Evaluation of FTIR-spectra
The FTIR spectra were evaluated by computer-aided subtrac-
tion of calibrated reference spectra using the OMNIC software
Finally, it is interesting to calculate the total number of OH
radicals formed in the ozonolysis of alkenes. The amount of
Table 1 Initial conditions of the experiments
[Methyl vinyl ether]/ [NO]/
[CH3ONO]/
[H2O2]/
[O3]/
[Di-n-butyl ether]c / [N2O5]/
Code
molecule cmꢁ3 molecule cmꢁ3 molecule cmꢁ3 molecule cmꢁ3 molecule cmꢁ3 molecule cmꢁ3
molecule cmꢁ3
Reaction with OH radicals:
MVH1
MVH2
MVH3
1.2 ꢃ 1013
2.5 ꢃ 1013
2.2 ꢃ 1013
1.3 ꢃ 1013
2.5 ꢃ 1013
—
7.5 ꢃ 1012
2.5 ꢃ 1011
—
—
—
—
—
—
—
—
—
—
—
—
7.5 ꢃ 1013
Reaction with O3 :
MVO1a
MVO2b
MVO3b
2.4 ꢃ 1013
—
—
—
—
—
—
—
—
—
ca. 5 ꢃ 1013
2.6 ꢃ 1013
4 ꢃ 1013
5 ꢃ 1012
5 ꢃ 1012
5 ꢃ 1012
—
—
—
(1.2–2.2) ꢃ 1012
(1.2–2.2) ꢃ 1012
Reaction with NO3 radicals:
MVN1b
MVN2b
(1.2–2.0) ꢃ 1012
—
—
—
—
—
—
1.8 ꢃ 1012
—
—
1.4 ꢃ 1013
(1.3–2.7) ꢃ 1012
9 ꢃ 1011
1.25 ꢃ 1013
Ozone added to methyl vinyl ether. b ca. 10 additions of methyl vinyl ether to ozone or N2O5 . c Di-n-butyl ether was added as a tracer for OH
a
radicals
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1726
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4

View Article Online
Table 2 Integrated absorption cross sections determined in this
studya
Integration range/cmꢁ1
s
10/cm moleculeꢁ1
Methyl vinyl ether:
1710–1554
1360–1285
(8.36 ꢀ 0.84) ꢃ 10ꢁ18
(1.75 ꢀ 0.18) ꢃ 10ꢁ18
(8.50 ꢀ 0.85) ꢃ 10ꢁ18
(4.05 ꢀ 0.41) ꢃ 10ꢁ18
1280–1163
930–750
Methyl formate:
3210–2770
1890–1660
(6.50 ꢀ 0.65) ꢃ 10ꢁ18
(1.88 ꢀ 0.19) ꢃ 10ꢁ17
(2.21 ꢀ 0.22) ꢃ 10ꢁ17
1315–1055
Formaldehyde:
1845–1625
Hydroperoxymethyl formate:b
1870–1630
1240–970
Formic anhydride:b
1890–1715
1144–853
(5.42 ꢀ 0.55) ꢃ 10ꢁ18
(1.84 ꢀ 0.54) ꢃ 10ꢁ17
(2.86 ꢀ 0.98) ꢃ 10ꢁ17
(2.64 ꢀ 0.52) ꢃ 10ꢁ17
(4.04 ꢀ 1.06) ꢃ 10ꢁ17
Base 10. See Thamm et al.10
a
b
OH radicals formed between times tnꢁ1 and tn is given by:
X
total
½OHꢄt ꢁt ¼ ½OHꢄt ꢁt
OH
tnꢁtnꢁ1
þ
D½reactantiꢄ
ð5Þ
n
nꢁ1
n
nꢁ1
i
In eqns. (1) to (5), the times tn and tnꢁ1 represent the recording
times of the spectra taken. The values calculated for each time
interval tn ꢁ tnꢁ1 can then be added to give total concentrations
over any given time range.
3. Results and discussion
3.1 OH reaction
Fig. 1 Product spectra recorded during the reaction of methyl vinyl
ether with OH radicals. (a) Product spectrum in the presence of NOx
(MVH2), after subtraction of the absorptions of reactants. (b) Product
spectrum in the absence of NOx (MVH3), after subtraction of the
absorptions of reactants. (c) scaled reference spectrum of methyl for-
mate. (d) scaled reference spectrum of formaldehyde. (e) Spectrum
(a) after subtraction of the absorptions of all identified compounds.
(f) Spectrum (b) after subtraction of the absorptions of all identified
compounds. Note that for easier comparison (e) and (f) are multiplied
by a factor of 10.
Product spectra recorded during the reaction of methyl vinyl
ether with OH radicals and selected reference spectra are
shown in Fig. 1. For better comparison all spectra are offset.
The main difference evident between panels (a) and (b) is the
absence of NOx in (b); (b) represents the NOx-free experiment
MVH3, (a) the NOx-containing experiment MVH2. The
absorptions of the main products methyl formate (c) and for-
maldehyde (d) are clearly evident in both spectra. It is also evi-
dent that somewhat less methyl formate and formaldehyde are
formed in the NOx-free experiment (b). In both spectra, the
same amount of methyl vinyl ether has reacted. The residual
spectra shown in panels (e) and (f) represent unidentified pro-
ducts, see below. Here, the intensities are higher for the NOx-
free system, reflecting the lower amounts formed of the main
products.
Plots of the product formation as a function of reacted
methyl vinyl ether are shown in Fig. 2. Formaldehyde concen-
trations are corrected for secondary losses through photolysis
and OH radical reaction, see section 2.4. Due to the low reac-
tivity,12,13 no correction was required for methyl formate. The
formation yields obtained are summarised in Table 3. The
observation of methyl formate as the main product is in agree-
ment with the ethyl vinyl ether study of Thiault et al.6 where
ethyl formate was observed as the main product of the OH
initiated oxidation. Scheme 1 shows simple pathways proposed
for the formation of these products. Two basic pathways are
possible in the reaction of methyl vinyl ether 1 with OH radi-
cals. One is addition of OH to the vinyl double bond (left side
of Scheme 1), the other one is abstraction of an H atom from
the methyl group (right side of Scheme 1). In the OH addition
pathway 2 sites are available for adding the OH radical, see
intermediates 2 and 3 in Scheme 1. In both cases, the initial
addition will be followed by addition of molecular oxygen to
give the radical intermediates 5 and 6, respectively. In the pre-
sence of NOx these will react with NO to give the oxyl radicals
8 and 9. The observed main products formaldehyde 11 and
methyl formate 14 are then formed by fragmentation of the
C–C bonds of 8 or 9, followed by reaction of the radical frag-
ment with O2 to give the corresponding carbonyl and HO2 .
Other possible pathways are direct abstraction of H from 8
or 9 to give 2-hydroxy-3-oxabutanal 15 or glycolic acid methyl
ester 16, respectively. A reference spectrum of 16 but not 15
was available. However, the product spectra recorded in the
presence of NOx did not show any absorptions of OH groups,
which both of these compounds should exhibit. Another pos-
sible reaction pathway of 8 and 9 is cleavage of the C–O bond
instead of the C–C bond. This would lead to the formation of
2-hydroxyacetaldehyde from 9 and possibly glyoxal from 8.
Again, no absorptions that can be attributed to these com-
pounds were observed. This is in line with expectations, as
C–C bonds are known to be significantly weaker than C–O
bonds and ethers generally react through C–C bond
cleavage.14
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4
1727

View Article Online
of peroxyl radicals with HO2 , RO2 þ HO2 ! ROOH þ O2 .
The absorptions do not conform to a well characterised hydro-
peroxide, i.e. methyl hydroperoxide, ethyl hydroperoxide,
hydroxymethyl hydroperoxide or hydroperoxymethyl formate.
It may be speculated that the hydroperoxides corresponding
to the peroxyl radicals 5 and/or 6 are formed. The formation
of hydroperoxides, with additional contributions from
RO2 þ RO2 cross- and self-reactions, can explain the lower
yields of the main products observed in the absence of NOx .
Another factor contributing to the lower methyl formate and
formaldehyde formation yields in the absence of NOx may
be the increased stability of the oxyl radicals 8 and 9 formed
via RO2 þ RO2 reactions compared to those formed from
RO2 þ NO reactions. This would be expected from previous
studies of such reactions,16 as discussed above.
Here, the carbon balance is only (53 ꢀ 8)%C, which is also
evident in the significantly more intense absorptions in the resi-
dual spectrum.
Given the high reactivity of the double bond, which is
further activated by the adjacent O-atom,3 the abstraction
channel is expected to be only of very minor importance. From
structure-reactivity relationships,17 it is estimated that 3.4% of
the reaction of methyl vinyl ether with OH radicals occurs
through the abstraction channel. Consequently the expected
main product from that channel, formic anhydride 10, is not
observed. No reference spectrum of vinyl formate 7 was avail-
able, but this compound is expected to react rapidly with OH
to give formic anhydride 10.
Fig. 2 Plot of the amount of product formed as a function of the
amount of methyl vinyl ether reacted in the OH radical reaction of
methyl vinyl ether. The formaldehyde concentrations are offset by
5 ꢃ 1012 molecule cmꢁ3
.
In the absence of NOx , the peroxyl radicals 5 and 6 will
undergo cross- and self-reactions. These can either lead to
the formation of the oxyl radicals 8 and 9, respectively, or,
via the molecular channel, to 2-hydroxy-3-oxabutanal 15, gly-
colic acid methyl ester 16 and 1,2-dihydroxy-3-oxabutane 17.
When the oxyl radicals 8 and 9 are formed in the absence of
NOx they are less thermally exited compared to the same radi-
cals when formed from the RO2 þ NO reaction. Therefore they
are less likely to undergo C–C bond cleavage to yield formal-
dehyde and methyl formate.15 Instead reaction with O2 will
gain significant importance, leading to the experimentally
observed reduced formation yields of the main products in
the absence of NOx .
Overall, (79 ꢀ 11)%C of the carbon in the reaction of methyl
vinyl ether with OH radicals has been accounted for in the pre-
sence of NOx . The residual spectrum, see Fig. 1(e), exhibits
absorptions in the carbonyl region around 1750 cmꢁ1. The
bands visible in the regions around 1680 cmꢁ1, 1290 cmꢁ1
and 830 cmꢁ1 are typical of nitrates, which would be expected
to form in significant yields. It is thus concluded that nitrates
and unidentified carbonyls, possibly carbonyl nitrates, prob-
ably form a significant portion of the missing carbon. Addi-
tional minor contributions from formic anhydride 10 and the
products of the reaction of the oxyl radicals 8 and 9 with
O2 , 2-hydroxy-3-oxabutanal 15 and glycolic acid methyl ester
16, respectively, can not be excluded, though no indications for
their formation are evident in the spectra.
3.2 O3 reaction
Product and reference spectra recorded during the ozonolysis
of methyl vinyl ether are shown in Fig. 3. The main products
are again methyl formate and formaldehyde, compare Fig. 1,
panels (c) and (d). Again, there is qualitative agreement with
the ethyl vinyl ether behaviour observed by Thiault et al.6
Also, formation of CO2 and CO has been observed in high
yields. After subtraction of the absorptions of the reactants
and identified products, absorptions belonging to an addi-
tional product are observed, see panel (b) of Fig. 3. This pro-
duct is identified as hydroperoxymethyl formate (HPMF), a
reference spectrum of which is shown in panel (c). In addition,
small amounts of formic anhydride 10 are observed, which is a
decomposition product of hydroperoxymethyl formate. Fig. 4
shows the concentrations of the main products formed as a
function of the amount of methyl vinyl ether reacted with
O3 . Corrections for reaction with OH radicals, which are
formed in the ozonolysis, have been applied to the concentra-
tions of methyl vinyl ether and formaldehyde. These cor-
rections accounted for less than 10% of the observed values
in all cases, average OH radical concentrations were
(3 ꢀ 1) ꢃ 106 molecule cmꢁ3. The concentrations of methyl for-
mate and formaldehyde shown in Fig. 4 are corrected for the
formation of these compounds in the OH reaction of methyl
vinyl ether, using the yields determined in section 3.1. All cor-
rections were performed as described in section 2.4. In addition
to the main products shown in Fig. 4, the formation of formic
acid (0.75 ꢀ 0.13)% and of formic anhydride (0.80 ꢀ 0.25)%
It is noteworthy that the residual spectrum of the reaction of
methyl vinyl ether with OH radicals under NOx-free conditions
shows absorptions of an OH-group around 3600 cmꢁ1. This
absorption is similar to those of known hydroperoxides,
though it cannot be excluded that alcohols such as 15 or 17
are responsible. Hydroperoxides may be formed in the reaction
Table 3 Product formation yields in reactions of methyl vinyl ethera
Reaction with
Methyl formate
HCHO
HPMF
Methanol
CO2
CO
OHb
OHc
O3
80.9 ꢀ 8.2
50.2 ꢀ 5.1
73.5 ꢀ 7.5
52.5 ꢀ 6.3
76.6 ꢀ 7.9
57.0 ꢀ 6.0
27.8 ꢀ 6.6
51.4 ꢀ 6.2
—
—
—
—
—
—
—
—
19.6 ꢀ 6.2
5.4 ꢀ 0.9
30.3 ꢀ 6.1
14.2 ꢀ 1.5
NO3
—
—
—
—
a
b
c
molar yields, in percent of methyl vinyl ether reacted. in the presence of NOx . in the absence of NOx .
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1728
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4

View Article Online
Scheme 1 Proposed pathways for the reaction of methyl vinyl ether with OH radicals. The products positively identified in the present study are
marked by frames.
has been observed. Formic anhydride was formed only after
the formation of hydroperoxymethyl formate and is probably
a product of its decomposition.10 The low formation yield of
formic acid is somewhat surprising for an ozonolysis reaction,
but can be explained by the reaction of HCOOH with Criegee
biradicals, which is known to give hydroperoxymethyl for-
mate.10 The fact that the observed yield of formic acid was less
than 4% of that of hydroperoxymethyl formate indicates that
this reaction must be very fast. More than 96% of the initially
formed formic acid apparently reacted with Criegee interme-
diates under the present experimental conditions. For form-
aldehyde, the formation yield in the initial phase of the
experiments, (27.8 ꢀ 6.6)%, is substantially higher than at later
stages, where a value of (16.4 ꢀ 2.6)% is determined, see inset
in Fig. 4. This behaviour may be qualitatively explained by
reaction of formaldehyde with Criegee intermediates, see
discussion below.
Scheme 2 shows a proposed degradation mechanism for the
ozonolysis of methyl vinyl ether, based on the general discus-
sion of alkene ozonolysis reactions given by Calvert et al.18
The reaction of methyl vinyl ether with O3 will initially lead
Fig. 3 Product spectra recorded during the ozonolysis of methyl
vinyl ether. (a) Product spectrum (MVO1). (b) Product spectrum after
subtraction of the absorptions of the reactants and all products but
hydroperoxymethyl formate (HPMF). (c) Scaled reference spectrum
of hydroperoxymethyl formate. (d) Residual spectrum after subtrac-
tion of all identified absorptions. Note that compared to (a), spectra
(b) and (d) are scaled up by factors of three and five, respectively.
Fig. 4 Plot of the amount of product formed as a function of the
amount of methyl vinyl ether reacted with O3 .
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4
1729

View Article Online
Scheme 2 Proposed pathways for the reaction of methyl vinyl ether with O3 . Of the products plotted, methyl formate 14, formaldehyde 11,
hydroperoxymethyl formate 22, CH3OH, HCOOH, CO2 and CO were actually observed.
to the formation of the primary ozonide, which can then
decompose through channels (1a) through (1d). This decom-
position can either lead to formation of methyl formate 14
and the Criegee biradical CH2OO or formaldehyde 11 and
the methoxylated Criegee biradical CH3O–CHOO. Both Crie-
gee intermediates can be formed in either excited (18 and 20) or
stabilised (19 and 21) forms, resulting in a total of 4 decompo-
sition channels for reaction (1). The branching ratio between
channels (1a) and (1b) on the one hand and (1c) and (1d) on
the other can be calculated based on the formation yields of
the stable products methyl formate 14 and formaldehyde 11
observed in the experiments. For formaldehyde only the for-
mation yield in the initial phase of the experiments was used,
see Fig. 4. Formaldehyde is a reaction partner for Criegee bi-
radicals, see reactions (3a), (3b), (5a) and (5b), which become
important sinks once a sufficient amount of the Criegee biradi-
cals and formaldehyde have been formed. Reactions (5b) and
(5e) may form additional methyl formate at higher concentra-
tions of HCHO and H2O, respectively, and lead to a higher
formation yield of that compound than from reactions (1a)
and (1b) alone. Given the linear formation of methyl formate
as a function of the reacted methyl vinyl ether, see Fig. 4, reac-
tions (5b) and (5e) are concluded to be of minor importance
only, though a small change in the methyl formate yield would
be hard to distinguish. Thus channels (1a) and (1b) are calcu-
lated to account for (73.5 ꢀ 7.5)%, channels (1c) and (1d) for
(27.8 ꢀ 6.6)% of the decomposition of the primary ozonide,
with a total yield of (101 ꢀ 10)%. This branching ratio is
usually close to 50:50 for n-alkyl substituted alkenes, only with
bulky or branched-chain substituents is there a preference for
the formation of the longer carbonyl, i.e. methyl formate 14,
and the shorter Criegee biradical, i.e. 18/19.18 Thus it seems
that methoxy substitution has an effect similar to that of a
bulky or branched substituent.
the proposed pathways for this reaction lead to the re-forma-
tion of formaldehyde, see Scheme 2. However, reaction (5b)
would lead to the formation of methyl formate and conse-
quently the formation yield of this compound should be lower
in the initial stage of the experiments. This is not observed.
Two explanations for this are conceivable. One is that the
expected reduction of the methyl formate yield from
(73.5 ꢀ 7.5)% to (62 ꢀ 10)% is barely statistically significant
given the error limits and may therefore not be observable.
Alternatively, another reaction pathway that gives neither
formaldehyde nor methyl formate might be operative here.
Branching ratios for the formation of the excited or stabi-
lised Criegee biradicals cannot be inferred from the present
study. In the literature branching ratios are available for some
Criegee biradicals, however, these appear to depend not only
on the chemical nature of the biradical, but also on the alkene
precursor.18
The four types of Criegee intermediates formed in the ozo-
nolysis of methyl vinyl ether can undergo a variety of further
reactions, which are summarised in Scheme 2. In general, the
excited species are proposed to undergo decomposition (chan-
nels 2a–f and 4a–f), while the stabilised species react with var-
ious atmospheric trace gases (channels 3a–e and 5a–e). This
reaction scheme has been compiled based on the overview
given by Calvert et al.18 The degradation pathways for the
substituted Criegee intermediates 20 and 21 have been drawn
up based on the known chemistry of 18 and 19 and must
therefore be regarded as highly speculative. Regarding branch-
ing ratios it can be concluded from the present study that,
based on their distinct products, channels (3c) and (4b) appear
to be of significance, in addition to the numerous channels
leading to the formation of either HCOOH, CO or CO2 .
These products are formed in multiple reaction channels,
and consequently no information on their formation pathways
can be given here. The other channels that lead to stable,
detectable products appear to be of minor importance only.
Reference FTIR-spectra exist for hydrogen peroxide (3e) and
(5e), hydroxymethyl hydroperoxide (3e) and methyl hydro-
peroxide (4d).
As described above, a higher yield of formaldehyde is
observed in the beginning of the experiments compared to
the later stages, the difference being (11.4 ꢀ 7.1)%. While this
can be qualitatively explained by reactions of formaldehyde
with Criegee intermediates, with the exception of (5b) all of
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1730
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4

View Article Online
A total of (88 ꢀ 7)%C of the carbon reacted has been recov-
ered as identified products, the remaining (12 ꢀ 7)%C are prob-
ably responsible for the residual absorptions shown in Fig. 3,
panel (d). Thus, as was the case in the OH radical reaction,
a reasonably complete carbon balance is achieved in the
ozonolysis of methyl vinyl ether.
As described in section 2.4 the amount of OH radicals
formed in the ozonolysis reaction can be calculated from the
degradation of the OH tracer compound di-n-butyl ether and
the known OH radical reaction rate constants of the various
trace gases in the reaction mixture. Using this procedure
results in an OH formation yield of (14 ꢀ 7)%. This value
has to be regarded as a lower limit because not all possible
sinks for OH radicals can be included in the calculation. The
compounds considered were di-n-butyl ether, methyl vinyl
ether and formaldehyde, methyl formate, HCOOH and CO
were disregarded due to their very low reactivity towards
OH radicals.12 A comparison of the measured OH formation
yield of (14 ꢀ 7)% with literature values is not very useful
since this study represents the first investigation of alkoxy
substituted alkenes. The OH yields determined for other
alkenes, even ethene, vary considerably between different
research groups, even when the same method is used. Conse-
quently, structure activity relationships regarding the forma-
tion of OH radicals in ozonolysis systems are currently not
available.
Fig. 6 Plot of the amount of products formed as a function of the
amount of methyl vinyl ether reacted with NO3 radicals.
formed as a function of the amount of methyl vinyl ether
reacted. The yields determined from the slopes of these plots
are given in Table 1, together with those of the other reactions
studied. It is evident from the data that the fraction of carbon
recovered as products is only (52 ꢀ 9)%C, much lower than in
the OH radical or O3 reactions. As a consequence, the residual
FTIR absorptions are significantly stronger, see panel (d) of
Fig. 5. Note that while the residual spectra shown in Figs. 1
and 3 have been scaled up to make them discernible next to
the product spectra, no such multiplication was necessary for
the residual spectra of the NO3 reaction. The residual spectrum
exhibits strong absorptions around 1675, 1300, 1285 and 840
cmꢁ1, which are indicative of nitrate functionalities (R–
ONO2). The absorptions at 1675, 1300, 1285 and 790 cmꢁ1
indicate the presence of a peroxy nitrate, and the absorption
at 1730 cmꢁ1 shows the presence of a carbonyl compound,
probably a carbonyl nitrate or peroxy nitrate.
3.3 NO3 reaction
A product spectrum recorded during the reaction of methyl
vinyl ether with NO3 radicals is shown in Fig. 5, panel (a). A
comparison with the reference spectra shown in panels (b)
and (c) shows that the main products are again methyl formate
14 and formaldehyde 11, with significant absorptions that
could not be associated with a product, see panel (d). These
residual absorptions are typical for organic nitrates and carbo-
nyl nitrates, the formation of which is common in NO3 reac-
tions of alkenes, see for example Barnes et al.19 Fig. 6 shows
plots of the amounts of methyl formate and formaldehyde
Scheme 3 shows a proposed chemical mechanism for the
reaction of methyl vinyl ether 1 with NO3 radicals. The main
degradation pathway is addition of NO3 to the double bond
of the vinyl group to give the nitrooxyl radicals 23 and 24,
respectively. This is followed by secondary addition of molecu-
lar oxygen, resulting in the formation of the peroxyl radicals 25
and 26. The observed main products suggest that the main
degradation pathway of 25 and 26 is reaction with another
RO2 species to form the corresponding oxyl radicals 27 and
28, which can then undergo C–C bond scission to give the
observed main products methyl formate 14 and formaldehyde
11 and NO2 .
The intermediates 27 and 28 may also form the dinitrate 35
by addition of NO2 to the oxyl group, or the carbonyl nitrates
33 and 34 by reaction with O2 . Peroxy nitrates 29 and 30 may
be formed by reversible addition of NO2 to the peroxyl groups
of 25 and 26, which may also react with R–CH2OO radicals,
i.e. 26, to form the hydroxy nitrates 31 and 32, O2 and a car-
bonyl, i.e. 34. Which of these nitrates 29 through 35 are actu-
ally formed in the system cannot be inferred from the results of
this study.
A formation yield for these nitrates can be approximated
based on the absorptions observed in the FTIR spectra, see
Fig. 5(d). Averaging absorption cross sections of 14 R–
ONO2 compounds Barnes et al.19 estimated an absorption
cross section of s10 ¼ (6.4 ꢀ 1.4) ꢃ 10ꢁ19 cm2 moleculeꢁ1 for
the –ONO2 symmetric stretching band located around 1285
cmꢁ1. Using this method, a nitrate formation yield of close
Fig. 5 Product spectra recorded during the reaction of methyl vinyl
ether with NO3 radicals. (a) Product spectrum after subtraction of
the absorptions of the reactants, H2O and NOx . (b) Scaled reference
spectrum of methyl formate. (c) Scaled reference spectrum of formal-
dehyde. (d) Residual product spectrum after subtraction of all identi-
fied absorptions.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4
1731

View Article Online
Scheme 3 Proposed mechanism for the reaction of methyl vinyl ether with NO3 radicals. The abstraction channel is not shown, see text for
details.
to 90% is obtained for the initial stages of the experiments.
However, this procedure is obviously subject to substantial
errors, not only because of the estimated absorption cross sec-
tion but also due to the fact that multiple –ONO2 groups can
be present in the compounds formed, see Scheme 3. The
nitrate yield obtained in this way is an indication for the for-
mation of difunctional nitrates and must be regarded as an
upper limit. From the relation of the magnitude of the absorp-
tion at 1283 cmꢁ1 to the amount of degraded methyl vinyl
ether as a function of time it is also evident that the formed
nitrates are probably highly reactive. This is consistent with
the expected formation of peroxy nitrates like 29 and 30.
The formation of the dinitrate 35 can be expected to depend
strongly on the NO2 concentration. Consequently, its forma-
tion yield should rise with the concentration of NO2 during
the course of the experiment. The expected increase of the yield
of 35 must lead to a corresponding decrease of the formation
yield of the fragmentation products of the oxyl radical, methyl
formate 14 and formaldehyde 11. Due to the reactions of NO3
radicals the NO2 concentration is increasing by about a factor
of two between the start and the end of the runs, however, no
effect on the yield of methyl formate and formaldehyde is seen,
see Fig. 6. From this it may be concluded that formation of the
dinitrate 35 is a minor process, even with the high NO2 concen-
trations employed in the present study. It should be unim-
portant under atmospheric conditions.
Next to the addition channel of NO3 radicals to methyl vinyl
ether there is also an abstraction channel, in which the nitrate
radical abstracts an H atom from the methyl group. The pro-
ducts of this channel, which are not shown in Scheme 3, would
be the same as shown in Scheme 1 for the OH radical abstrac-
tion channel, i.e. the highly reactive vinyl formate 7 and formic
anhydride 10. For alkylated alkenes H atom abstraction is gen-
erally much slower than the addition pathway, and given the
additional activation of the double bond through the methoxy
group, abstraction is assumed to be of very minor importance
in the present study. This is confirmed by the absence of the
expected main product of this channel, formic anhydride 10.
Another class of product frequently observed in nitrate radi-
cal reactions is the oxiranes. In the present study no evidence
for the formation of oxiranes was found in the spectra; this
result is in line with expectations from earlier studies where
oxiranes were primarily formed at low pressures and in the
absence of O2 ,20 though for highly reactive alkenes their for-
mation has also been observed at 1 bar of air.21
3.4 Aerosol formation
Particle measurements were conducted for all the experiments
presented in this study; however, formation of aerosol was
only detected in the ozonolysis of methyl vinyl ether. In the
OH and NO3 radical reactions no formation of aerosol was
observed.
The observed aerosol profiles showed typical behaviour asso-
ciated with homogeneous nucleation. In the beginning of the
experiments aerosol number concentrations are very low, i.e.
around 20 particles cmꢁ3; at some point a burst of small parti-
cles is seen. The particles initially have mean diameters <13 nm,
but appear to rapidly grow to diameters > 13 nm, which are
detectable by the instrument used. Particle number concentra-
tions measured immediately after they entered the detectable
size range were in the range from 140 000 to 210 000 particles
cmꢁ3. Given the detection limit of 13 nm of the SMPS instru-
ment, it cannot be excluded that ultra-fine seed particles are
present in the chamber prior to the experiments, however, it
appears extremely unlikely that their concentration can reach
values of several 100 000 particles cmꢁ3, as seen here. Thus it
is concluded that homogeneous nucleation is taking place.
The initially observed small particles subsequently grew due
to coagulation and/or evaporation of small particles followed
by re-condensation on larger ones. Towards the end of the
experiments mean diameters of around 60 nm were measured.
The aerosol profiles obtained can be evaluated in several
ways, the most common of which is the adsorption/partition-
ing model developed by Odum et al.22 In this model semi-vola-
tile compounds are proposed to partition between the gas- and
aerosol phase. Partitioning coefficients Kae,i and formation
yields ai for semi-volatile compounds i are obtained by fitting
eqn. (6) to a plot of the observed aerosol yield Y as a function
of the aerosol mass M0 .
X
A corresponding plot is shown in Fig. 7, the curve represents a
aiKae;i
1 þ Kae;iM0
Y ¼ M0
ð6Þ
i
non-weighted least squares fit according to eqn. (6), assuming
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1732
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4

View Article Online
It may be speculated that impurities of larger molecules are
present in the reaction mixture, and the ozonolysis of these is
producing the aerosol. To be responsible for the observed
aerosol concentrations, the concentration of any impurity
must be well in excess of 1%. Impurities may, for example,
be formed by condensation reactions of methyl vinyl ether in
analogy to those of vinyl chloride. Tests were performed to
assess the influence of impurities. Spectra were recorded both
of the sample as provided by the supplier and after several
steps of condensation in liquid nitrogen followed by partial
evaporation. The spectra of the raw and purified samples were
identical, indicating that no significant quantities of such
impurities were present. The only impurity detected in the sam-
ple of methyl vinyl ether was methanol (1.2 ꢀ 0.3)%, the pre-
sence of additional impurities in mixing ratios as required to
explain the observed aerosol concentrations is concluded to
be unlikely. However, even with the tests conducted an influ-
ence of impurities cannot be excluded with absolute certainty.
It can only be speculated on the nature of the semi-volatile
compound(s) responsible for the observed aerosol formation.
One possibility is oxalic acid, which has been previously
observed in secondary organic aerosol.25 Condensation reac-
tions, possibly involving the Criegee intermediate, are also fea-
sible. Recently, Ziemann26 presented experimental evidence
suggesting that diacyl peroxides are the nucleating agents in
the ozonolyis of cyclohexene and other cyclic alkenes. These
acyl peroxides were proposed to be formed by cross-reactions
of acyl peroxyl radicals. In any event, if condensation reactions
are responsible for aerosol formation, the aerosol yield should
be a strong function of the radical concentrations in the experi-
ments. Changing the added concentrations of ozone and the
alkene could be a good test for this. If condensation plays a
significant role, the evaluation according to Odum et al.22
should give higher aerosol yields when the added concentra-
tions are raised. It is planned to conduct such experiments in
the future.
Fig. 7 Aerosol formation in the ozonolysis of methyl vinyl ether. The
main figure shows a plot of the aerosol mass yield as a function of the
aerosol mass formed. The inset shows the aerosol mass formed as a
function of the amount of methyl vinyl ether reacted with ozone. All
calculations were done assuming a spherical shape of the aerosol and
a density of 1 g cmꢁ3
.
only one partitioning semi-volatile compound. For this
hypothetical compound
formation yield of a ¼
a
(0.66 ꢀ 0.13)% by mass and a partitioning coefficient of
Kae ¼ (0.98 ꢀ 0.27) m3 mgꢁ1 is obtained. The inset in Fig. 7
shows a plot of the mass of aerosol formed as a function of
the mass of methyl vinyl ether reacted with ozone. It is clearly
evident that a certain amount of methyl vinyl ether must react
before aerosol is formed. From the slope of this plot, ignoring
the part until aerosol is formed, an aerosol formation yield of
(0.66 ꢀ 0.07)% is calculated. A total of (162 ꢀ 38) mg mꢁ3 of
methyl vinyl ether is consumed before the onset of aerosol for-
mation. This can be interpreted as the amount that needs to be
consumed to produce enough of the semi-volatile compound
to reach its vapour pressure. Thus in combination with the for-
mation yield taken from the slope the saturation vapour pres-
sure is estimated to be (1.07 ꢀ 0.27) mg mꢁ3. Obviously, this
approach is ignoring effects like supersaturation and the value
obtained should be treated as an estimate. The formation
yields of the condensable product calculated by the two meth-
ods are identical, indicating near-ideal behaviour of this
hypothetical compound. The calculated vapour pressure indi-
cates a very non-volatile compound. Even when a lower limit
of about 60 g molꢁ1 is assumed for the molecular weight of
the aerosol forming compound(s), a vapour pressure of 1.6
mbar is derived.
4. Conclusions and atmospheric implications
The experiments conducted in the present study were per-
formed under conditions that are close to, but not identical
with those found in the troposphere. Consequently corrections
have to be applied to the yield data presented when they are to
be used in the modelling of atmospheric processes.
In the OH radical reactions, the main difference to atmo-
spheric conditions is the substantially higher concentrations
of NOx in the experiments. This is thought not to have a sig-
nificant effect on the data since vinyl ethers are expected to
be emitted in conjunction with NOx , and therefore enough
NOx will always be present to ensure that the primary fate
of peroxyl radicals is reaction with NO, not with other RO2 .
In the presence of NOx about 20% of the carbon balance could
not be identified as products. It is assumed that these additional
products are the hydroxy carbonyls 15 and 16, which are
formed by reaction of the oxyl radicals 8 and 9 with O2 . The
abstraction channel, which gives formic anhydride 10, is
assumed to be negligible. The following simplified reaction
scheme is suggested for use in chemical models of the OH
initiated oxidation of methyl vinyl ether:
One of the three experiments shown in Fig. 7 exhibits a dis-
tinctly different behaviour compared to the other two. Also,
this experiment shows a falling aerosol yield with aerosol mass,
which is inconsistent with accepted aerosol formation beha-
viour.22 In this experiment (MVO1) a large amount of ozone
was added to methyl vinyl ether pre-mixed in the chamber,
see section 2.2. As a consequence all methyl vinyl ether reacted
within a few minutes and a homogeneous mixing of the cham-
ber could not be assured. Thus the two instruments used for
the aerosol study, the FTIR to detect methyl vinyl ether and
the SMPS to measure aerosol size distributions, may not have
sampled the same air parcel. Due to these problems this experi-
ment is not considered in the above evaluation.
MVE þ OH ! 0:8 methyl formate þ 0:8 HCHO
þ 0:2 hydroxy carbonyls 15=16
In the ozonolysis experiments, the most important difference
to atmospheric conditions is the absence of water vapour in the
experiments. As a consequence it is expected that under atmo-
spheric conditions, hydroperoxymethyl formate will not be
formed in significant yield. Instead, the major degradation
pathway of the stabilised Criegee intermediates will be reaction
with water vapour, leading to the formation of unstable
Methyl vinyl ether, with only three carbon atoms, is the
smallest compound studied until now that produces aerosol
during its atmospheric degradation. Usually seven or more
C-atoms are required for aerosol formation from non-cyclic
hydrocarbons, six for cyclic species.23,24
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4
1733

View Article Online
hydroxy hydroperoxides, see channels (3e) and (5e) in Scheme
2. The other products observed result largely from the decom-
position of the excited Criegee intermediate and their forma-
tion yields are taken directly from the experiments. The
following simplified scheme is therefore recommended for
modelling:
Acknowledgements
Financial support of this work by the European Commission
(project MOST, contract EVK2-CT-2001-00114) and
Grant-in-Aid for Scientific Research (13127105) from the Min-
istry of Education, Science, Sports, and Culture of Japan is
gratefully acknowledged.
a
MVE þ O3 ! 0:75 MF þ 0:3 HCHO þ 0:3 CO2
þ 0:15 CO þ 0:2 HCOOH þ 0:05 CH3OH
þ 0:14 OH þ 0:35 hydroxy hydroperoxides
þ 0:01 aerosol
References
1
2
3
M. Placet, C. O. Mann, R. O. Gilbert and M. J. Niefer, Atmos.
Environ., 2000, 34, 2183.
R. F. Sawyer, R. A. Harley, S. H. Cadle, J. M. Norbeck, R. Slott
and H. A. Bravo, Atmos. Environ., 2000, 34, 2161.
For the reaction of methyl vinyl ether with NO3 radicals, it
has to be taken into account that the experiments conducted in
the present study were carried out under conditions of high
NO2 concentrations that are not representative of the tropo-
sphere. Hence, the formation of the di-nitrate 35 and the
unstable peroxy nitrates 29 and 30 is probably negligible under
atmospheric conditions. To derive product formation yields
for atmospheric conditions several assumptions need to be
made. Here it is assumed that the nitro peroxyl radicals 25
and 26 behave similar to their hydroxyl derivatives 5 and 6.
Under this assumption the yields of formaldehyde and methyl
formate in the NO3 initiated oxidation of methyl vinyl ether
are expected to be similar to those observed in the OH initiated
oxidation under NOx free conditions, i.e. (54 ꢀ 7)%. This is
due to the absence of RO2 þ NO reactions in both systems.
Then, a branching ratio of 0.33:0.67 can be assumed for the
molecular versus the peroxyl radical formation channels of
the RO2 þ RO2 reaction shown in Scheme 3, which is an aver-
age value for such reactions.27 Thus the hydroxy nitrates 31/32
should be formed with a molar yield of 16%. Considering that
the carbonyl nitrates 33/34 will be formed through both reac-
tion of O2 with the oxyl radicals 27/28 and through the mole-
cular channel of the self reaction of the peroxyl radicals 25/26,
their yield is estimated to be 30%. The following simplified
scheme is proposed:
´ ´
M. Joly, Federation des Industries de la Peinture, Encres et Colles,
FIPEC, personal communication; S. Lemoine, European Solvents
Industry Group, ESIG, personal communication.
R. A. Perry, R. Atkinson and J. N. Pitts, Jr., J. Chem. Phys., 1977,
67, 611.
E. Grosjean and D. Grosjean, Int. J. Chem. Kinet., 1998,
30, 21.
´
G. Thiault, R. Thevenet, A. Mellouki and G. Le Bras, Phys.
Chem. Chem. Phys., 2002, 4, 613.
D. Grosjean and E. L. Williams, Atmos. Environ., 1992, 8, 1395,
BASF Development Product Information, http://www.basf.de/,
2001.
4
5
6
7
8
9
H. Akimoto, M. Hoshino, G. Inoue, F. Sakamaki, N. Washida
and M. Okuda, Environ. Sci. Technol., 1979, 13, 471.
R. Volkamer, B. Klotz, I. Barnes, T. Imamura, K. Wirtz, N.
Washida, K. H. Becker and U. Platt, Phys. Chem. Chem. Phys.,
2002, 4, 1598.
10 J. Thamm, S. Wolff, W. V. Turner, S. Ga¨b, W. Thomas,
F. Zabel, E. H. Fink and K. H. Becker, Chem. Phys. Lett.,
1996, 258, 155.
11 B. J. Finlayson, J. N. Pitts, Jr. and H. Akimoto, Chem. Phys.
Lett., 1972, 12, 495.
12 S. LeCalve, G. LeBras and A. Mellouki, J. Phys. Chem. A, 1997,
101, 5489.
13 T. J. Wallington, P. Dagaut, R. Liu and M. J. Kurylo, Int. J.
Chem. Kinet., 1988, 20, 177.
14 T. Maurer, H. Geiger, I. Barnes and K. H. Becker, J. Phys. Chem.
A, 2000, 104, 11 087.
MVE þ NO3 ! 0:16 hydroxy nitrate
þ 0:3 carbonyl nitrate
15 T. J. Wallington, M. D. Hurley, V. Fedotov, C. Morrell and
G. Hancock, J. Phys. Chem. A, 2002, 106, 8391.
16 M. Bilde, J. J. Orlando, G. S. Tyndall, T. J. Wallington,
M. D. Hurley and E. W. Kaiser, J. Phys. Chem. A, 1999, 103,
3963.
17 E. S. C. Kwok and R. Atkinson, Atmos. Environ., 1995, 29,
1685.
18 J. G. Calvert, R. Atkinson, J. A. Kerr, S. Madronich, G. K.
Moortgat, T. J. Wallington, G. Yarwood, The Mechanisms of
Atmospheric Oxidation of the Alkenes, Oxford University Press,
New York, 2000.
þ 0:54 formaldehyde þ 0:54 methyl formate
Due to the limited number of kinetic measurements avail-
able1,2 no statement can currently be made regarding the rela-
tive importance of these different degradation channels under
atmospheric conditions. However, given the high reactivity
of methyl vinyl ether apparent from the experiments presented
here it appears likely that all three channels, including reaction
with NO3 radicals, will be of significance under atmospheric
conditions.
19 I. Barnes, V. Bastian, K. H. Becker and Z. Tong, Chem. Phys.
Lett., 1990, 94, 2413.
With the present study most of the products of the three
important atmospheric degradation reactions of methyl vinyl
ether have been identified. Carbon balances are fairly com-
plete, with the exception of the NO3 radical reaction, where
a significant amount of unidentified nitrates is formed. Apart
from completing the carbon balances some additional ques-
tions remain, these mostly concern the ozonolysis. One is on
the chemical nature of the semi-volatile intermediate(s) respon-
sible for the observed formation of aerosol and on the nature
of the aerosol itself. Another question is whether the
ozonolysis in the presence of water vapour gives the same pro-
ducts as those observed here, and the influence of water vapour
on aerosol formation. These questions should be addressed in
future studies.
20 T. Berndt and O. Bo¨ge, J. Atmos. Chem., 1995, 21, 275.
21 T. Benter, M. Liesner, R. N. Schindler, H. Skov, J. Hjorth and
G. Restelli, J. Phys. Chem., 1994, 98, 10 492.
22 J. R. Odum, T. Hoffmann, F. Bowman, D. Collins, R. C. Flagan
and J. H. Seinfeld, Environ. Sci. Technol., 1996, 30, 2580.
23 S. N. Pandis, R. H. Harley, G. R. Cass and J. H. Seinfeld, Atmos.
Environ., Part A, 1992, 26A, 2269.
24 J. H. Seinfeld and S. N. Pandis, Atmospheric Chemistry and Phy-
sics: From Air Pollution to Climate Change, John Wiley & Sons
Inc., New York, 1998.
25 E. D. Baboukas, M. Kanakidou and N. Mihalopoulos, J. Geo-
phys. Res., 2000, 105, 14 459.
26 P. J. Ziemann, J. Phys. Chem. A, 2002, 106, 4390.
27 P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau, G. D.
Hayman, M. E. Jenkin, G. K. Moortgat and F. Zabel, Atmos.
Environ., Part A, 1992, 26A, 1806.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
1734
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 7 2 5 – 1 7 3 4