Kinetics and mechanisms for the reactions of ozone with unsaturated oxygenated compounds

Article abstract of DOI:10.1002/cphc.201000404

Rate coefficients for the reaction of ozone with a series of unsaturated oxygenated compounds are determined in air at atmospheric pressure and (298±3) K. Rate data are obtained using both relative and absolute rate techniques, and the measured rate coefficients are found to be in good agreement. The results show that the reactivity of the compounds with respect to addition of ozone to the double bond is a function of the nature of the oxygenated substituent. Product distribution studies on the reactions provide information on the decomposition pathways for the primary ozonides, and on the effect of the oxygenated group on the relative importance of the degradation pathways. The results are discussed in terms of their importance in the atmospheric oxidation of unsaturated oxygenated compounds.

Full text of DOI:10.1002/cphc.201000404

DOI: 10.1002/cphc.201000404
Kinetics and Mechanisms for the Reactions of Ozone with
Unsaturated Oxygenated Compounds
Ismael Al Mulla,[c] Lisa Viera,[c] Rebecca Morris,[c] Howard Sidebottom,[c] Jack Treacy,[b] and
Abdelwahid Mellouki*[a]
Rate coefficients for the reaction of ozone with a series of un-
saturated oxygenated compounds are determined in air at at-
mospheric pressure and (298ꢀ3) K. Rate data are obtained
using both relative and absolute rate techniques, and the mea-
sured rate coefficients are found to be in good agreement. The
results show that the reactivity of the compounds with respect
to addition of ozone to the double bond is a function of the
nature of the oxygenated substituent. Product distribution
studies on the reactions provide information on the decompo-
sition pathways for the primary ozonides, and on the effect of
the oxygenated group on the relative importance of the deg-
radation pathways. The results are discussed in terms of their
importance in the atmospheric oxidation of unsaturated oxy-
genated compounds.
1. Introduction
Field measurements indicate that oxygenated volatile organic
compounds form a major component of the trace gases found
in the troposphere,[1] and hence play an important role in de-
termining the oxidising capacity of the troposphere both on a
regional and a global scale.[2] Unsaturated oxygenated com-
pounds, such as ethers, aldehydes, ketones, acids and esters,
are present in the atmosphere as a result of direct anthropo-
genic and biogenic emissions.[3] Anthropogenic sources include
their use as solvents, fuel additives and releases during indus-
trial manufacturing processes. They are also formed in the tro-
pospheric oxidation of alkenes, aromatic compounds and bio-
genic species, such as isoprene, terpenes and sesquiterpenes.
Oxidation of unsaturated oxygenated compounds in the at-
mosphere can be initiated by reaction with ozone,[4–7] ni-
trate[5,8,9] and hydroxyl[5,10–13] radicals. Reaction of O3 with an
alkene has been shown to involve addition to the double-
bond system to form an ozonide, which rapidly decomposes
to generate a carbonyl compound and a Criegee diradical in-
termediate (carbonyl oxide).[4–7] Formation of carbonyl oxide
species in the ozonolysis of alkenes in the gas phase has not
been directly observed; however, the stabilised CH2OO biradi-
cal has recently been identified in a study of the chlorine-
atom-initiated oxidation of dimethyl sulphoxide in the gas
phase.[14]
vide information on the tropospheric lifetimes of these com-
pounds with respect to reaction with ozone. The fate of unsa-
turated oxygenated compounds in the troposphere is not only
important in terms of oxidant formation; they may also be oxi-
dised to form polyfunctionalised species. These compounds
often have low volatility and relatively high solubility. Thus,
they can lead to the formation of secondary organic aerosols
or be susceptible to uptake in cloud and fog droplets. Hence,
the presence of unsaturated oxygenated compounds in the
troposphere can impact on wet deposition and on climate
forcing.[2]
Experimental Section
Kinetic and product studies on the reactions of ozone with a series
of unsaturated oxygenated compounds were performed at (298ꢀ
3) K in air at atmospheric pressure in a collapsible FEP Teflon reac-
tion vessel of volume approximately 50 L. The reaction vessel was
housed in a thermostatically controlled chamber and a uniform
temperature (ꢀ3 K) was maintained during the reactions by posi-
tioning fans on either side of the reaction vessel. The temperature
was monitored by a thermocouple placed in the centre of the
vessel. The reaction vessel was connected to the gas chromato-
graph and long-path Fourier transform infrared (FTIR) spectroscopy
1
analytical systems by = ’’ outside diameter (O.D.) Teflon tubing.
In the present work, rate coefficients for the reaction of O3
with a series of unsaturated ethers, carbonyls, acids and esters
have been determined using both relative and absolute rate
techniques. The reactivity of the compounds with respect to
reaction with O3 is discussed in terms of their structure. The
primary degradation products of the ozonides produced from
the reaction of O3 with a number of unsaturated oxygenated
compounds have been identified, and the influence of the
nature of the substituent oxygenated group on the relative
yields of carbonyl and Criegee compounds formed in the de-
composition of the ozonides is discussed. The kinetic data pro-
4
[a] Dr. A. Mellouki
ICARE/CNRS
45071 Orlꢀans (France)
Fax: (+33) 2386 960 04
E-mail: mellouki@cnrs-orleans.fr
[b] Prof. J. Treacy
School of Chemical and Pharmaceutical Sciences
Dublin Institute of Technology, Kevin Street, Dublin 8 (Ireland)
[c] Dr. I. Al Mulla, Dr. L. Viera, Dr. R. Morris, Prof. H. Sidebottom
School of Chemistry and Chemical Biology
University College Dublin, Belfield, Dublin 4 (Ireland)
ChemPhysChem 2010, 11, 4069 – 4078
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4069

A. Mellouki et al.
Kinetic Studies: Relative Rate Measurements: The reaction vessel
was filled to approximately half of its full capacity with the diluent
gas zero-grade air using a calibrated flowmeter. Measured pres-
sures (MKS Baratron capacitance manometer) of substrate and ref-
erence compounds were flushed from Pyrex bulbs of determined
volume into the reaction vessel by the diluent gas. Since previous
studies have shown that hydroxyl radicals are generally formed in
the reactions of ozone with alkenes,[4–7] experiments were carried
out in the presence of an excess of cyclohexane to scavenge OH
radicals. The Criegee intermediates generated in the reactions of
O3 with acrylic acid and methacrylic acid have been reported to
react with the parent acids,[15] and hence formic acid was added to
the reaction mixtures used to study the kinetics of the reactions of
O3 with these acids to scavenge the Criegee biradicals. Following
addition of the reactant gases, the reaction vessel was filled to its
full capacity with the diluent gas. Typical initial reactant mixing
ratios were [Substrate]0 =20–60 ppm, [Reference]0 =20–60 ppm,
[Cyclohexane]0 =300–700 ppm and in the reactions of acrylic and
of spectral stripping in which small fractions of calibrated reference
spectra were continuously subtracted from the sample spectra.
Absolute Rate Measurements: The experimental technique is based
on measuring the loss of O3 in the presence of a known excess
concentration of the substrate compounds. Reactants entered the
1
reaction vessel through a = ’’ O.D. Teflon tube, placed along the
4
centre of the vessel, which was plugged at the end and perforated
along its length. This allowed rapid mixing of the reactants (ꢂ
2 min), as determined by gas chromatographic analysis of a test
hydrocarbon. Ozone was produced as previously described. Before
the start of each experiment, the reaction vessel was purged and
then filled to a measured volume with ozonised air to give a final
O3 mixing ratio of less than 1 ppm. Accurate concentrations of the
oxygenated alkenes, (1–20)ꢁ1014 moleculecmꢁ3, were added to
the reaction vessel by expanding a measured pressure of the com-
pounds into a bulb of calibrated volume, and sweeping the con-
tents into the vessel with the diluent gas. The reaction mixtures
were allowed to mix for approximately 2 min before sampling. The
total volume of the diluent gas admitted into the reaction vessel
was determined by the use of a calibrated mass flow controller. In
most of the experiments the oxygenated alkene was added to the
reaction vessel just before the O3 decay experiments were initiated,
but the order of addition of the reactants had no detectable effect
upon the O3 decay rates. Background O3 decays in the absence of
the oxygenated compounds were determined periodically and
shown to be negligible (ꢂ10ꢁ4 sꢁ1) compared to the O3 decays in
the presence of the substrate compounds. Ozone concentrations
were continuously monitored during the course of the reactions
with a chemiluminescence ozone monitor (Monitor Labs, model
8410). The signal was fed to a 16-bit high-resolution analog–digital
converter (Pico, model ADC), which was interfaced to a personal
computer for data analysis. The oxygenated compounds were
quantitatively monitored by gas chromatography with flame ioni-
sation detection (Shimadzu, model GC-8A). The gas chromatograph
was calibrated with measured concentrations of the substrate prior
to the kinetic experiments.
methacrylic
acids
[HC(O)OH]0 =0–200 ppm
(1 ppm=2.46ꢁ
1013 moleculecmꢁ3 at 298 K and 760 Torr total pressure). The reac-
tants were allowed to mix for around 30 min prior to the addition
of O3 to the system. Ozone was generated by passing zero-grade
air through an ozone generator (Monitor Labs) directly into the re-
action vessel for about 1 min at a flow rate of 1 Lminꢁ1 giving an
O3 mixing ratio of around 5 ppm. During the course of the relative
rate experiments, approximately six further additions of O3 to the
vessel were made in the same manner, and the losses of substrate
and reference compounds were monitored after each O3 addition.
In general, substrate and reference compounds were analysed
during the course of the reaction using gas chromatography; how-
ever, the loss of acrylic and methacrylic acids was also determined
by infrared spectroscopy. Samples of the reaction mixtures were
drawn through a six-port Valco gas-sampling valve for quantitative
analysis by gas chromatography (Shimadzu Model GC-8A with
flame ionisation detector). Chromatographic separation was ach-
ieved using wide-bore capillary columns (Alltech) operated over
the temperature range 30–508C using nitrogen as the carrier gas.
A 15 m, 0.53 mm internal diameter (I.D.) Carbowax-coated column
was employed to separate the substrate and reference compounds
in the acrolein, methyl vinyl ketone, acrylic acid and methacrylic
acid studies, whereas a 30 m, 0.53 mm I.D. SE-54 column was used
to separate the components in the investigations on ethyl vinyl
ether, allyl ethyl ether, ethylene glycol divinyl ether, methyl acry-
late, methyl methacrylate and vinyl acetate. The resulting chroma-
tograms were recorded, stored and analysed using a Shimadzu CR-
6A Chromatopac integrator. Three replicate chromatograms with
standard deviations less than 5% were obtained for each datum
point. For the kinetic investigations on acrylic acid and methacrylic
acid, loss of the acids was also monitored using FTIR spectroscopy
(Mattson Research Series fitted with a wide-band MCT detector),
and of the reference compounds by gas chromatography. Spectra
were obtained using an evacuable 1 L Pyrex glass cell, containing a
multi-pass White mirror arrangement which gave a path length of
5 m, mounted in the sample compartment of the spectrometer.
Prior to sampling, the cell was evacuated and the reaction mixture
was then expanded into the cell. The spectra were recorded over
Product Studies: All the product distribution experiments on the
ozonolysis of unsaturated oxygenated compounds were carried
out at (298ꢀ3) K and atmospheric pressure with zero-grade air as
the diluent gas. At the beginning of each experiment, the reaction
vessel was filled with zero-grade air and the background IR spec-
trum recorded. The substrate compound was first flushed into the
vessel with diluent gas as described for the relative rate studies.
An excess of cyclohexane was added to the system to scavenge
any OH radicals generated in the ozonolysis reactions. Formic acid
was also added to the reaction mixtures for investigations on the
reactions of O3 with acrylic acid and methacrylic acid, to suppress
the reactions of the Criegee intermediates formed in the ozonolysis
of the parent acids. Typical reaction mixtures had mixing ratios of
[Substrate]0 =30–100 ppm, [Cyclohexane]0 =500–1000 ppm, and in
the reactions of acrylic acid and methacrylic acid [HC(O)OH]0 =
200 ppm. The reactants were allowed to mix for around 30 min
before the addition of O3 to the system. Approximately 5 ppm of
O3 was added to the reaction mixture at the start of the reaction
by passing zero-grade air through the ozoniser. During the course
of the oxidation reactions, up to ten further additions of O3 were
made in the same manner. The loss of substrate and formation of
products was monitored after each O3 addition and their concen-
trations corrected for dilution due to the addition of ozonised air.
Analysis of the reaction mixtures was by FTIR spectroscopy and
was carried out as described for the relative rate studies on acrylic
acid and methacrylic acid.
the wavelength range 400–4000 cmꢁ1 with a resolution of 2 cmꢁ1
.
The spectra were derived from the co-addition of 128 scans and
analysed by using Mattson WINFIRST software. Reference spectra
and calibration curves for the reactants and products were ob-
tained by expanding measured pressures of pure samples of the
compounds into the IR cell and recording their spectra. The con-
centrations of the substrates were determined through a process
4070
www.chemphyschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2010, 11, 4069 – 4078

Reactions of Ozone with Unsaturated Oxygenated Compounds
Materials: The materials employed in the kinetic investigations
were synthetic air, zero-grade, 99.95% (Air Products), ethyl vinyl
ether (CH2=CHOC2H5), 99% (Aldrich), allyl ethyl ether (CH2=
CHCH2OC2H5), 97% (Aldrich), ethylene glycol divinyl ether (CH2=
CHOCH2CH2OCH=CH2), 97% (Aldrich), acrolein (CH2=CHC(O)H),
99% (Aldrich), methyl vinyl ketone (CH2=CHC(O)CH3), 99% (Al-
drich), acrylic acid (CH2=CHC(O)OH), 99% (Aldrich), methacrylic
acid CH2=C(CH3)C(O)OH), 99% (Aldrich), methyl acrylate (CH2=
CHC(O)OCH3), 99% (Fluka), methyl methacrylate (CH2=C-
(CH3)C(O)OCH3), 99% (Aldrich), vinyl acetate (CH2=CHOC(O)CH3),
>99% (Aldrich), 1-methylcyclohexene (1-CH3-c-C6H9), 97% (Al-
drich), propene (CH2=CHCH3), 99% (Aldrich), cyclohexane (c-C6H12),
99.9% (Aldrich) and formic acid (HC(O)OH), 98% (Fluka). All materi-
als were used as received. The sources of materials employed in
the product studies were the same as those in the kinetic investi-
gations. The reaction products, ethyl formate (C2H5OC(O)H), 97%
(Aldrich), ethylene glycol diformate (HC(O)OCH2CH2OC(O)H), 98%
(Fluka), glyoxylic acid (HC(O)C(O)OH), 50% in water (Aldrich) and
pyruvic acid CH3C(O)C(O)OH), 98% (Aldrich) were used as received.
Formaldehyde was generated in the reaction chamber by the addi-
tion of excess O3 to ethene. The yield of HC(O)H in this reaction
has been determined to be unity within experimental error, based
to the system resulting in dilution of the reactants. If the dilu-
tion factor at time t is Dt (Dt =ln[C]0/[C]t where [C]0 and [C]t are
the concentrations of a chemically inert species in the system
at time 0 and t, respectively), then Equation (I) is modified to
[Eq. (II)]:
lnf½Substrateꢃ0=½SubstrateꢃtgꢁDt
¼ k1=k2 ½lnf½Referenceꢃ0=½ReferenceꢃtgꢃꢁDt
ðIIÞ
Under the reaction conditions employed, the concentration
of the substrate and reference compounds were found to
decay by 5–95% during the experiments. At least three indi-
vidual runs were carried out with each compound. Plots of the
data in the form of Equation I) were found to be linear with
near-zero intercepts and representative data are shown in
Figure 1. Rate coefficients for the reactions of O3 with the un-
saturated oxygenated compounds were calculated from the
[6]
on the loss of CH2=CH2. Methyl glyoxal was prepared by the pho-
tolysis of molecular chlorine in air in the presence of hydroxyace-
tone. The yield of CH3C(O)C(O)H in this reaction has been shown
[16]
to be close to unity based on the loss of HOCH2C(O)CH3. Formic
acetic anhydride was prepared by the rapid addition of freshly dis-
tilled acetyl chloride to sodium formate in tetrahydrofuran.[17] After
vacuum filtration, the solution was distilled under reduced pressure
to remove excess acetyl chloride and tetrahydrofuran. The
CH3C(O)OC(O)H was further purified by distillation at reduced pres-
sure, and the resulting product was shown to be better than 99%
pure by infrared and NMR spectroscopy.
2. Results
2.1. Kinetic Studies
Figure 1. Relative rate plots for the reaction of O3 with unsaturated oxygen-
ated compounds at (298ꢀ3) K and atmospheric pressure.
2.1.1. Relative Rate Measurements
Relative rate coefficients for the reactions of O3 with the unsa-
turated oxygenated substrates were determined by comparing
their rates of decay with the following reference compounds:
1-methylcyclohexene, propene, methyl vinyl ketone and
methyl acrylate [reactions (1),(2)]:
measured rate coefficient ratios k1/k2 using the following
values of k2 (in units of cm3 moleculeꢁ1 sꢁ1): 1-methylcyclohex-
ene k2 =1.62ꢁ10ꢁ16 [18]
,
propene k2 =1.01ꢁ10ꢁ17 [5]
,
methyl vinyl
work]
ketone k2 =4.32ꢁ10ꢁ18 [this
,
and methyl acrylate k2 =1.10ꢁ
work]
10ꢁ18 [this
.
1-Methylcyclohexene was used as the reference
O3 þ Substrate ! Products
O3 þ Reference ! Products
ð1Þ
ð2Þ
compound for the reactions of O3 with ethyl vinyl ether and
ethylene glycol divinyl ether. As a check on the value of this
reference rate coefficient, relative rate experiments were car-
ried out on mixtures of 1-methylcyclohexene and cis-2-butene.
Concentration–time data plotted in the form of Equation (I)
Provided that both compounds are only lost by the reaction
with O3 and that dilution due to sampling is negligible, then
[Eq. (I)]:
gave
a slope of k2(O3+1-methylcyclohexene)/k2(O3+cis-2-
butene)=1.33ꢀ0.10, which, combined with the evaluated rate
coefficient for the reaction of O3 with cis-2-butene, k2(O3+cis-2-
lnf½Substrateꢃ0=½Substrateꢃtg
butene)=1.25ꢁ10ꢁ16 cm3 moleculeꢁ1 sꢁ1 [5]
yields a value of
,
ðIÞ
¼ k1=k2 lnf½Referenceꢃ0=½Referenceꢃtg
k2(O3+1-methylcyclohexene)=(1.66ꢀ0.13)ꢁ
10ꢁ16 cm3 moleculeꢁ1 sꢁ1, in good agreement with the rate coef-
ficient reported by Treacy et al.[18] for this reaction. A summary
of the reference compounds, the rate coefficient ratios k1/k2
and the derived values of k1 obtained from the experiments is
given in Table 1. It should be noted that methyl vinyl ketone
The subscripts 0 and t indicate concentrations at the begin-
ning of the reaction and at time t, respectively, and k1 and k2
are the rate coefficients for reactions (1) and (2). During the
course of an experiment, up to six additions of O3 were made
ChemPhysChem 2010, 11, 4069 – 4078
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
4071

A. Mellouki et al.
and methyl acrylate were used
as both reactant and reference
compounds. All the experimen-
tal rate coefficients obtained in
this work are shown in Table 2,
together with previously report-
ed results. The errors quoted are
twice the standard deviation
arising from the least-squares fit
of the data and do not include
an estimate of the error in the
reference rate coefficient k2. The
errors in k2 may add a further
20% to the reported rate coeffi-
cients from the relative rate
studies. It was found that the ad-
dition of formic acid to the reac-
tion mixtures decreased the
decays of both acrylic acid and
methacrylic acid, while the rate
of loss of the reference com-
pounds remained unaffected.
The decays of the other unsatu-
rated oxygenated compounds in
the presence of O3 were shown
to be unchanged following addi-
tion of formic acid to the mix-
tures. It was found that on in-
creasing the amounts of formic
acid added to the acrylic acid
and methacrylic acid systems,
the decay rates of the acids
Table 1. Rate data from relative rate studies on the reactions of O3 with a series of unsaturated oxygenated
compounds at (298ꢀ3) K and atmospheric pressure.
[a]
Compound
Reference
k1/k2
k1[a] [cmꢁ3 moleculeꢁ1 sꢁ1
]
ethyl vinyl ether
allyl ethyl ether
ethylene glycol divinyl ether
acrolein
methyl vinyl ketone
acrylic acid
methacrylic acid
methyl acrylate
methyl methacrylate
vinyl acetate
1-methylcyclohexene[b]
propene[c]
1-methylcyclohexene[b]
methyl acrylate[d]
propene[c]
methyl acrylate[d]
methyl acrylate[d]
methyl vinyl ketone[e]
methyl vinyl ketone[e]
propene[c]
0.926ꢀ0.102
0.86ꢀ0.12
(1.50ꢀ0.17)ꢁ10ꢁ16
(8.69ꢀ1.23)ꢁ10ꢁ18
(5.15ꢀ0.52)ꢁ10ꢁ16
(3.63ꢀ0.35)ꢁ10ꢁ19
(4.16ꢀ0.33)ꢁ10ꢁ18
(7.60ꢀ0.51)ꢁ10ꢁ19
(1.89ꢀ0.18)ꢁ10ꢁ18
(1.19ꢀ0.11)ꢁ10ꢁ18
(5.75ꢀ0.52)ꢁ10ꢁ18
(2.30ꢀ0.22)ꢁ10ꢁ18
3.18ꢀ0.32
0.330ꢀ0.032
0.412ꢀ0.033
0.691ꢀ0.046[f]
1.72ꢀ0.16[f]
0.275ꢀ0.026
1.33ꢀ0.12
0.228ꢀ0.022
[a] The errors quoted are twice the standard deviation arising from the least-squares fit of the data and repre-
sent precision only. Errors in the reference rate coefficients k2 may add a further 20% to the reported values of
the rate coefficients k1. [b] Value for k2 =1.62ꢁ10ꢁ16 cm3 moleculeꢁ1 sꢁ1 taken from Treacy et al.[18] [c] Value for
k2 =1.01ꢁ10ꢁ17 cm3 moleculeꢁ1 sꢁ1 taken from Atkinson.[5] [d] Value for k2 =1.10ꢁ10ꢁ18 cm3 moleculeꢁ1 sꢁ1 taken
from the average value from the relative and absolute rate studies in this work. [e] Value for k2 =4.32ꢁ
10ꢁ18 cm3 moleculeꢁ1 sꢁ1 taken from the average value from the relative and absolute rate studies in this work.
[f] Results obtained using an excess of formic acid (200 ppm) as a scavenger for Criegee intermediates.
Table 2. Rate coefficients for the reactions of O3 with a series of unsaturated oxygenated compounds at
(298ꢀ6) K and atmospheric pressure.
Compound
k1[a] [cmꢁ3 moleculeꢁ1 sꢁ1
]
Technique[b]
Reference
ethyl vinyl ether
(CH2=CHOC2H5)
(1.54ꢀ0.30)ꢁ10ꢁ16
(2.0ꢀ0.2)ꢁ10ꢁ16
(1.50ꢀ0.17)ꢁ10ꢁ16
(1.43ꢀ0.08)ꢁ10ꢁ16
(8.69ꢀ1.23)ꢁ10ꢁ18
(8.95ꢀ0.50)ꢁ10ꢁ18
(5.15ꢀ0.52)ꢁ10ꢁ16
(5.65ꢀ0.30)ꢁ10ꢁ16
(2.8ꢀ0.5)ꢁ10ꢁ19
(3.0ꢀ0.4)ꢁ10ꢁ19
(2.6ꢀ0.5)ꢁ10ꢁ19
(3.63ꢀ0.35)ꢁ10ꢁ19
(3.36ꢀ0.22)ꢁ10ꢁ19
(4.77ꢀ0.59)ꢁ10ꢁ18
(4.72ꢀ0.09)ꢁ10ꢁ18
(4.2ꢀ0.4)ꢁ10ꢁ18
(5.84ꢀ0.39)ꢁ10ꢁ18
(5.4ꢀ0.6)ꢁ10ꢁ18
(4.16ꢀ0.33)ꢁ10ꢁ18
(4.48ꢀ0.20)ꢁ10ꢁ18
(6.5ꢀ1.3)ꢁ10ꢁ19
(7.60ꢀ0.51)ꢁ10ꢁ19
(7.87ꢀ0.70)ꢁ10ꢁ19
(4.1ꢀ0.4)ꢁ10ꢁ18
(1.89ꢀ0.18)ꢁ10ꢁ18
(2.25ꢀ0.74)ꢁ10ꢁ18
2.9ꢁ10ꢁ18
S-UV
S-FTIR
RR
S-CL
RR
Grosjean and Grosjean[19]
Thiault et al.[20]
This work
This work
This work
This work
This work
This work
Atkinson et al.[21]
Treacy et al.[22]
Grosjean et al.[23]
This work
allyl ethyl ether
(CH2=CHCH2OC2H5)
ethylene glycol divinyl ether
(CH2=CHOCH2CH2OCH=CH2)
acrolein
S-CL
RR
S-CL
S-CL
S-UV
S-UV
RR
S-CL
S-CL
S-UV
S-UV
S-UV
S-FTIR
RR
S-CL
S-FTIR
RR
S-CL
S-FTIR
RR
S-CL
DF-UV
S-UV
RR
S-CL
S-UV
RR
S-CL
S-UV
S-UV
RR
(CH2=CHC(O)H)
reached
a limiting value and
This work
hence the values of k1 calculated
from the slopes of the relative
rate plots, k1/k2, obtained in
these experiments also reached
a limiting value. Representative
data for the reaction of O3 with
mixtures of methacrylic acid and
methyl acrylate with variable
amounts of HC(O)H in the
system are shown in Figure 2.
methyl vinyl ketone
(CH2=CHC(O)CH3)
Atkinson et al.[21]
Grosjean et al.[24]
Treacy et al.[22]
Grosjean and Grosjean[19]
Neeb et al.[15]
This work
This work
acrylic acid
(CH2=CHC(O)OH)
Neeb et al.[15]
This work
This work
methacrylic acid
CH2=C(CH3)C(O)OH)
Neeb et al.[15]
This work
This work
methyl acrylate
(CH2=CHC(O)OCH3)
Munshi et al.[25]
Grosjean and Grosjean[19]
This work
(1.05ꢀ0.15)ꢁ10ꢁ18
(1.19ꢀ0.11)ꢁ10ꢁ18
(1.00ꢀ0.05)ꢁ10ꢁ18
(7.5ꢀ0.9)ꢁ10ꢁ18
(5.75ꢀ0.52)ꢁ10ꢁ18
(6.66ꢀ0.60)ꢁ10ꢁ18
(2.92ꢀ0.30)ꢁ10ꢁ18
(3.20ꢀ0.47)ꢁ10ꢁ18
(2.30ꢀ0.22)ꢁ10ꢁ18
(2.83ꢀ0.20)ꢁ10ꢁ18
2.1.2. Absolute Rate Measure-
ments
This work
methyl methacrylate
(CH2=C(CH3)C(O)OCH3)
Grosjean et al.[23]
This work
Second-order rate coefficients
for the reactions of O3 with the
unsaturated oxygenated com-
pounds were obtained by moni-
toring the increased rates of O3
decay in the presence of mea-
sured excess concentrations of
the substrates. Under these con-
ditions O3 may be lost via the
This work
vinyl acetate
(CH2=CHOC(O)CH3)
Grosjean et al.[23]
Grosjean and Grosjean[19]
This work
S-CL
This work
[a] The errors quoted are twice the standard deviation and represent precision only. [b] S-UV, static system-loss
of O3 determined by UV analysis; S-FTIR, static system-loss of unsaturate determined by FTIR analysis; RR, rela-
tive rate-loss of substrate and reference determined by GC or FTIR analysis; S-CL, static system-loss of O3 deter-
mined by chemiluminescence analysis; DF-UV, discharge flow-loss of O3 determined by UV analysis.
following
processes:
4072
www.chemphyschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2010, 11, 4069 – 4078

Reactions of Ozone with Unsaturated Oxygenated Compounds
the pseudo-first-order rate coefficients determined in this work
against the concentrations of the unsaturated oxygenated
compounds were linear as expected and the rate coefficients
k1, given in Table 2, were determined from least-squares analy-
sis of the slopes (the errors quoted are twice the standard de-
viation arising from the least-squares fit of the data). Typical
plots of the pseudo-first-order rate coefficients against the sub-
strate concentrations are shown in Figure 3.
Figure 2. Variation of the rate coefficient ratio k1/k2 from relative rate studies
on the reaction of O3 with mixtures of methacrylic acid and methyl metha-
crylate as a function of added formic acid.
[reactions (1),(3),(4)]:
O3 þ Substrate ! Products
O3 þ Impurities ! Products
O3 þ Wall ! Loss of O3
ð1Þ
ð3Þ
ð4Þ
Figure 3. Pseudo-first-order rate coefficients for the decay of O3 as a func-
tion of the concentrations of the unsaturated oxygenated compounds at
(298ꢀ3) K and atmospheric pressure.
and hence [Eq. (III)]:
ꢁd½O3ꢃ=dt ¼ ðk1½Substrateꢃ þ k3½Impuritiesꢃ þ k4Þ ½O3ꢃ
ðIIIÞ
The background decay of O3 was shown to be negligible
compared to the decay rates in the presence of the substrates.
All the compounds had purities higher than 97% and therefore
loss of O3 by reaction with impurities in the unsaturated oxy-
genated compounds is expected to be insignificant. Hence,
loss of O3 either at the wall or by reaction with impurities in
the substrate compounds is unlikely to be important and
[Eq. (IV)]:
2.2. Product Studies
Product distribution studies for the ozonolysis of a number of
the unsaturated oxygenated compounds were carried out at
(298ꢀ3) K in air at atmospheric pressure. Representative prod-
uct yield plots for the primary carbonyl compounds formed in
the reactions are shown in Figure 4, and the product yield data
are given in Table 3. For clarity the yields of HC(O)H are not
shown in Figure 4.
ꢁd½O3ꢃ=dt ¼ k1 ½Substrateꢃ ½O3ꢃ
ðIVÞ
All absolute rate kinetic experiments were carried out under
pseudo-first-order conditions with [Substrate]0/[O3]0 ratios ꢄ
10. Furthermore, the high [Substrate]0/[O3]0 ratio ensures negli-
gible contributions from reaction of O3 with primary reaction
products and hence [Eq. (V)]:
0
½O3ꢃt ¼ ½O3ꢃ0 eꢁk t
ðVÞ
where k’=k1[Substrate].
Excellent first-order plots were obtained from the measured
rates of O3 decay for all the unsaturated oxygenated com-
pounds investigated. This is in line with the previous studies of
Niki and co-workers,[26–28] in which it was established that the
optimum experimental conditions for determining rate coeffi-
cients for the reaction of O3 with alkenes involves carrying out
the reactions in the presence of >50 ppm O2, at high [Sub-
strate]0/[O3]0 ratios, and monitoring the decay of O3. Plots of
Figure 4. Formation yields of primary carbonyl compounds formed in the re-
actions of O3 with unsaturated oxygenated compounds at (298ꢀ3) K and at-
mospheric pressure. For clarity the yields of formaldehyde in the reactions
are not shown.
ChemPhysChem 2010, 11, 4069 – 4078
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
4073

A. Mellouki et al.
The reason for this discrepancy
is unclear. In the relative rate ex-
periments with acrylic acid and
methacrylic acid it was found
that addition of formic acid to
the reaction mixtures decreased
the decays of both acrylic acid
and methacrylic acid, while the
rate of decay of the reference
compound methyl acrylate re-
mained unchanged. Hence, the
rate coefficient ratio k1/k2 was
shown to decrease when in-
creasing amounts of formic acid
were added to the system.
Series of relative rate experi-
ments were performed in which
the amounts of formic acid
added to the reaction mixtures
were increased, and it was
shown that the decay rates of
acrylic acid and methacrylic acid
reached a limiting value, and
therefore the ratio k1/k2 deter-
mined in the relative rate studies
also became constant (see
Figure 2) for the case of metha-
crylic acid. Neeb et al.[15] have
previously suggested that the
Criegee intermediates generated
Table 3. Formation yields of the primary carbonyl compound products from the gas-phase ozonolysis of a
series of unsaturated oxygenated compounds carried out in an excess of cyclohexane at (298ꢀ3) K and atmos-
pheric pressure.
Compound
Primary carbonyl products
Yield [%]
Reference
CH2=CHCH3
HC(O)H
CH3C(O)H
HC(O)H
CH3OC(O)H
HC(O)H
C2H5OC(O)H
HC(O)H
C2H5OC(O)H
HC(O)H
C2H5OC(O)H
HC(O)H
HC(O)OCH2CH2OC(O)H
HC(O)H
CH3C(O)C(O)H
HC(O)H
CH3C(O)C(O)H
HC(O)H
C2H5C(O)C(O)H
HC(O)H
HC(O)C(O)OH
HC(O)H
CH3C(O)C(O)OH
HC(O)H
CH3C(O)OC(O)H
HC(O)H
65ꢀ5
45ꢀ9
28ꢀ7
74ꢀ8
49ꢀ5
>39
19ꢀ4
87ꢀ6
15ꢀ4
80ꢀ8
47ꢀ4
50ꢀ4
5
Tuazon et al.[29]
CH2=CHOCH3
CH2=CHOC2H5
Klotz et al.[30]
Grosjean and Grosjean[3]
Thiault et al.[20]
This work
CH2=CHOCH2CH2OCH=CH2
CH2=CHC(O)CH3
This work
Grosjean and Grosjean[31]
This work
87
44ꢀ8
50ꢀ8
5
44
148ꢀ20
<10
77ꢀ7
74ꢀ10
45ꢀ5
not measured
30ꢀ4
70ꢀ8
CH2=CHC(O)C2H5
CH2=CHC(O)OH[a]
CH2=C(CH3)C(O)OH[a]
CH2=CHOC(O)CH3
Grosjean and Grosjean[31]
This work
This work
Grosjean and Grosjean[3]
This work
CH3C(O)OC(O)H
[a] Results obtained using an excess of formic acid (200 ppm) as a scavenger for Criegee intermediates.
3. Discussion
in the reactions of O3 with acrylic acid and methacrylic acid
react with the parent acids to form the corresponding hydro-
peroxy compounds [reactions (5) and (6)]. In the presence of
excess formic acid the Criegee intermediates will be scavenged
[reaction (7)], and the loss of acrylic acid and methacrylic acid
arises solely from their reactions with O3:
3.1. Rate Coefficient Data
Rate coefficients for the reactions of ozone with a series of un-
saturated oxygenated compounds measured in this work at
(298ꢀ3) K, together with the available literature data, are
given in Table 2. Rate data were determined using relative and
absolute rate techniques, and rate coefficients obtained from
both methods were in agreement within the combined experi-
mental errors. Relative rate experiments were carried out in
excess cyclohexane to scavenge OH radicals generated in the
ozonolysis reactions. The rate coefficients determined in this
study for the reaction of O3 with ethyl vinyl ether are in good
agreement with the value reported by Grosjean and Gros-
jean,[19] who investigated the reaction under pseudo-first-order
conditions with an excess of the ether. The somewhat higher
rate coefficient found by Thiault et al.[20] was derived from fit-
ting the ozone and ether concentration–time profiles mea-
sured in experiments carried out under second-order condi-
tions with similar initial concentrations of reactants. The agree-
ment between the rate coefficients determined in this work for
the reactions of O3 with the carbonyl compounds acrolein and
methyl vinyl ketone, and the esters methyl acrylate, methyl
methacrylate and vinyl acetate, and the data previously report-
ed in the literature is good with the exception of the rate coef-
ficient determined by Munshi et al.[25] for the reaction of O3
with methyl acrylate determined in a discharge flow system.
RCOO þ CH2¼CHCðOÞOH ! RCðOOHÞOCðOÞCH¼CH2
RCOO þ CH2¼CðCH3ÞCðOÞOH ! RCðOOHÞOCðOÞCðCH3Þ¼CH2
ð6Þ
ð7Þ
ð5Þ
RCOO þ HCðOÞOH ! RðOOHÞOCðOÞH
In the absolute rate investigations on the reactions of O3
with acrylic acid and methacrylic acid, the reactions were car-
ried out with [O3]0 ![Substrate]0 and loss of O3 was monitored
by chemiluminescence. Since the substrate compound was in
large excess, small losses of the acids by reaction with OH radi-
cals or Criegee intermediates will be unimportant. The rate co-
efficient determined in this study for the reaction of O3 with
acrylic acid is in good agreement with the value previously re-
ported by Neeb et al.[15] These workers investigated the reac-
tion under pseudo-first-order conditions by monitoring the
loss of the acid by FTIR spectroscopy in an excess of O3. The
reaction was studied in excess cyclohexane to trap OH radicals.
The rate coefficient was shown to decrease markedly when the
reaction was carried out in the presence of formic acid, and
4074
www.chemphyschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2010, 11, 4069 – 4078

Reactions of Ozone with Unsaturated Oxygenated Compounds
reached a limiting value at high concentrations of formic acid.
Neeb et al.[15] also determined the rate coefficient for the reac-
tion of O3 with methacrylic acid using the same technique as
they employed for the study of O3 with acrylic acid. The rate
coefficient determined for the reaction of O3 with methacrylic
acid by Neeb et al.[15] is around a factor of two higher than the
value obtained in the present work. It seems likely that suffi-
cient formic acid was added to the system to quench the reac-
tions of the Criegee intermediates with the acid, and it is diffi-
cult to rationalise the difference in the rate coefficient ob-
tained and that determined in this work. No previous kinetic
data for the reactions of O3 with allyl ethyl ether and ethylene
glycol divinyl ether have been reported.
the electron density at the double bond due to the presence
of the electron-withdrawing inductive effect of the carbonyl
group. The rate coefficient for the reaction of O3 with CH2=
CHC(O)H is around five times smaller than that for the corre-
sponding alkene, ethene, whereas the rate coefficient for reac-
tion with CH2=CHC(O)CH3 is approximately a factor of two
lower than for reaction with propene. Hence, it seems that the
ꢁC(O)H group is significantly more electron-withdrawing than
the ꢁC(O)CH3 group. The reactivity of CH2=CHC(O)CH3 is con-
siderably greater than that of CH2=CHC(O)OH and CH2=
CHC(O)OCH3, which suggests that replacement of the ꢁCH3
group in methyl vinyl ketone by either a ꢁOH or a ꢁOCH3
group reduces the reactivity of these compounds. Hence, the
ꢁC(O)OH and ꢁC(O)OCH3 groups appear to be more electron-
withdrawing than the ꢁC(O)CH3 group. In a similar manner,
CH2=C(CH3)C(O)OH and CH2=C(CH3)C(O)OCH3 are both signifi-
cantly less reactive than CH2=C(CH3)2. The rate coefficient for
The reaction of O3 with alkenes has been shown to take
place by electrophilic addition to the double-bond system to
produce an energy-rich primary ozonide.[4–7] The rate coeffi-
cient for the reaction of O3 with propene is around a factor of
six times higher than that for reaction with ethene, k(O3+CH2= reaction of O3 with CH2=CHOC(O)CH3 shows that the presence
CHCH3)=1.01ꢁ10ꢁ17 cm3 moleculeꢁ1 sꢁ1[5]
and k(O3+CH2= of the carbonyl group in this ester causes a significant reduc-
CH2)=1.59ꢁ10ꢁ18 cm3 moleculeꢁ1 sꢁ1 [5]
Hence, as expected,
.
tion in its reactivity compared to that observed for the corre-
sponding vinyl ether and alkene. Thus, the ꢁOC(O)CH3 group is
electron-withdrawing, whereas both the ꢁOC2H5 and ꢁCH3
groups are electron-donating.
substitution of a hydrogen atom in ethene by an electron-do-
nating methyl group enhances reaction with the electrophilic
ozone molecule. Further increases in the chain length of the
substituent alkyl group in going from propene to 1-decene do
not significantly change the O3 rate coefficient.[5] The available
data show that in general, successive methyl substitution of
ethene tends to lead to an increase in the rate of addition of
O3 to the double-bond system.[5] However, the rate coefficients
for the alkenes have been shown to depend strongly on their
molecular structure. Thus, the rate coefficients for the reaction
of O3 with CH2=CHCH3 and CH2=C(CH3)2 are approximately
equal, whereas the rate coefficients for reaction with both cis-
and trans-CH3CH=CHCH3 are over an order of magnitude
higher than for reaction with CH2=CHCH3, with the trans
isomer being significantly more reactive than the cis isomer.[5]
The similarity in the rate coefficients for reaction of O3 with
propene and 2-methylpropene, and the difference in reactivity
of cis- and trans-2-butene appear to be due mainly to steric
effects.[22]
Interestingly, the reduction in the rate coefficients for the re-
actions of OH radicals with methyl vinyl ketone, k(OH+CH2=
CHC(O)CH3 =2.0ꢁ10ꢁ11 cm3 moleculeꢁ1 sꢁ1,[32] and acrylic acid,
k(OH+CH2=CHC(O)OH)=1.8ꢁ10ꢁ11 cm3 moleculeꢁ1 sꢁ1,[33] com-
pared to that for reaction with propene, k(OH+CH2=CHCH3)=
2.6ꢁ10ꢁ11 cm3 moleculeꢁ1 sꢁ1 [5]
are significantly smaller than
,
the corresponding reductions in the rate coefficients for the re-
actions of O3 with these compounds. It is suggested that the
transition state for the addition of OH radicals to the alkene
double-bond system in unsaturated carbonyl compounds is
stabilised by H-bond formation between the attacking OH radi-
cal and the carbonyl group. For example, in the reaction of OH
radicals with acrylic acid, the transition state can be stabilised
by formation of a six-membered
ring structure.
The rate coefficient for the reaction of O3 with CH2=
CHOC2H5 determined in this work (Table 2) is about 15 times
higher than that for reaction with CH2=CHCH3. This enhance-
ment in reactivity indicates that the ꢁOC2H5 group increases
the electron density at the double-bond site to a much greater
extent than a ꢁCH3 group. The reactivity of CH2=CHCH2OC2H5
towards O3 is similar to that of CH2=CHCH3. In this case the in-
fluence of the alkoxy group on the electron density at the
Stabilisation by H-bond forma-
tion is not possible for the reac-
tions of O3 with unsaturated car-
bonyl compounds. It has previous-
ly been proposed that H-atom abstraction from saturated ke-
tones by OH radicals involves the initial formation of a weakly
bound hydrogen-bonded adduct between the OH radical and
the carbonyl group, which leads to intramolecular H-atom
double bond is minimal due to the presence of the ꢁCH2ꢁ transfer from sites in the alkyl group via cyclic transition
group, which separates the two functional groups. As expect-
ed the rate coefficient for reaction of O3 with the nonconjugat-
ed divinyl ether, CH2=CHOCH2CH2OCH=CH2, is about twice that
observed for reaction with CH2=CHOC2H5. The reported rate
coefficients for the reactions of O3 with vinyl aldehydes, ke-
tones, acids and esters, in which the substituent group con-
tains a carbonyl moiety, are considerably lower than for the re-
actions with the analogous unsubstituted alkenes (Table 2).
These results can be rationalised in terms of the reduction in
states.[2,34]
3.2. Product Studies
The yields of the primary carbonyl products formed in the ozo-
nolysis of the unsaturated carbonyl compounds investigated in
this study are given in Table 3 together with the available liter-
ature data. All the experiments reported herein were carried
out in the presence of excess cyclohexane to scavenge OH rad-
ChemPhysChem 2010, 11, 4069 – 4078
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
4075

A. Mellouki et al.
icals, which may have been generated in the reactions. In the
experiments with acrylic acid and methacrylic acid, excess
formic acid was also added to the reaction mixtures to inhibit
the reactions of the Criegee intermediates with the parent
acids. Primary products from the ozone-initiated oxidation of
methyl vinyl ether[30] and ethyl vinyl ether[20] have previously
been reported. Thiault et al.[20] observed ethyl formate and
formaldehyde as the major products with yields of 87 and
19%, respectively, from the ozonolysis of ethyl vinyl ether, in
reasonable agreement with the present work. In an earlier
study on this system by Grosjean and Grosjean,[3] the yields of
ethyl formate and formaldehyde, measured by chromatograph-
ic analysis coupled with derivatisation techniques, appeared to
be about equal at around 50%. It is possible that the derivati-
sation procedures employed in the experiments caused prob-
lems with the quantification of the products. The yields of the
primary products methyl formate (74%) and formaldehyde
(28%) found in the reaction of O3 with methyl vinyl ether in
the presence of cyclohexane by Klotz et al.[30] are in line with
those found in this work and by Thiault et al.[20] for the ozonol-
ysis of CH2=CHOC2H5. Studies on the ozone-initiated oxidation
of methyl vinyl ketone[31] and vinyl acetate[3] have previously
been reported. The yields of methyl glyoxal and formaldehyde
of approximately 50% determined in this work from the reac-
tion of O3 with CH2=CHC(O)CH3 are significantly different from
those reported by Grosjean and Grosjean,[31] who found yields
of 87 and 5% for CH3C(O)C(O)H and HC(O)H, respectively. How-
ever, the near equal amounts of C2H5C(O)C(O)H and HC(O)H
obtained by Grosjean and Grosjean[31] from the ozonolysis of
ethyl vinyl ketone are consistent with the product distributions
obtained from CH2=CHC(O)CH3 in this work. The yield of
formic acetic anhydride of 70% measured in this study from
the reaction of O3 with vinyl acetate suggests that the primary
ozonide decomposes mainly to produce CH3C(O)C(O)H and the
Criegee intermediate CH2OO, whereas Grosjean and Grosjean[3]
reported a yield of HC(O)H close to 50%, thus indicating that
the decomposition of the ozonide occurs almost equally
between the two possible reaction channels.
kenes gives rise to the formation of two possible carbonyl
compounds and the reactions of O3 with 1-alkenes, where the
substituent is an n-alkyl group, have been shown to yield
almost equal amounts of the two primary carbonyl com-
pounds RC(O)H and HC(O)H.[6] In a number of cases, it has
been reported that the yield of HC(O)H in these reactions is
slightly higher than 50%, which suggests that the reactions of
the Criegee intermediates may provide an additional source of
HC(O)H.[6] Investigations on the reactions of O3 with 1-alkenes
in which the substituent alkyl group is highly branched or is a
bulky group seem to indicate that there is preferential forma-
tion of the products RC(O)H and HCOO compared to HC(O)H
and RCOO.[6] The decomposition of the primary ozonides
formed in the reactions of O3 with unsymmetrical alkenes of
general structure CH2=CR1R2 and R1CH=CR2R3 leads mainly to
the production of the disubstituted Criegee intermediates with
typical yields of around 65%, with corresponding yields of the
stable carbonyls R1R2C(O) and R2R3C(O) of 35%.[6]
The primary product distribution yields obtained in this
work from the ozonolysis of ethyl vinyl ether and vinyl acetate,
which both have an oxygen atom bound to a carbon atom of
the double-bond system, clearly show that the primary ozo-
nides decompose mainly to give C2H5OC(O)H (80%) and
CH3C(O)OC(O)H (70%), respectively, and the energy-rich biradi-
cal HCOO. Thus, it appears as though the ꢁOC2H5 and ꢁ
OC(O)CH3 groups in ethyl vinyl ether and vinyl acetate have
the same effect on the relative importance of the decomposi-
tion of the primary ozonides as found for the ozonides formed
in the addition of O3 to 1-alkenes containing highly branched
or bulky groups. The major primary products observed in the
reaction of O3 with ethylene glycol divinyl ether were the difor-
mate HC(O)OCH2CH2OC(O)H and HC(O)H with near equal yields
of approximately 50%. The initial step in the reaction is addi-
tion of O3 to one of the double bonds to form the primary
ozonide. Assuming that this ozonide decomposes in a similar
manner to that found for the ozonide formed by addition of
O3 to CH2=CHOC2H5, then the yields of the unsaturated for-
mate HC(O)OCH2CH2OCH=CH2 and HC(O)H are expected to be
around 70 and 30%, respectively. Rapid addition of O3 to the
unsaturated formate gives rise to the production of the difor-
mate and formaldehyde. If the yields of HC(O)OCH2CH2OC(O)H
and HC(O)H are again 70 and 30%, respectively, based on the
loss of the unsaturated formate, then the overall yields of di-
formate and HC(O)H in the reaction of O3 with ethylene glycol
divinyl ether would be predicted to be approximately equal
with yields close to 50%. The results of the present work on
the ozone-initiated oxidation of this compound provide sup-
port for these arguments.
All of the available evidence indicates that the reaction of O3
with alkenes proceeds by the initial electrophilic addition of O3
to the carbon–carbon double bond to produce an energy-rich
primary ozonide.[4–7] The ozonide rapidly decomposes to gener-
ate a carbonyl compound and an energy-rich biradical, the
Criegee intermediate [reactions (8) and (9)]:[4–7]
The observed yields for CH3C(O)C(O)H and HC(O)H of
around 50% from the reaction of O3 with CH2=CHC(O)CH3 indi-
cate that decomposition of the ozonide occurs almost equally
between the two possible reaction channels. Hence, the sub-
stituent groups ꢁCH3 and ꢁC(O)CH3 appear to have a similar
effect on the relative importance of the decomposition path-
ways for the ozonides produced from the reactions of O3 with
propene and methyl vinyl ketone. The relative yields of the pri-
mary carbonyl compounds observed in the ozonolysis of acryl-
where [ ]* denotes an energy-rich species and R represent alkyl
groups.
Symmetrical alkenes have been shown to lead to the forma-
tion of a single carbonyl compound with a yield close to unity,
and hence it appears that that formation of carbonyl com-
pounds from the Criegee intermediates in these systems is
only of minor importance.[6] Ozonolysis of unsymmetrical al-
4076
www.chemphyschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2010, 11, 4069 – 4078

Reactions of Ozone with Unsaturated Oxygenated Compounds
ic acid and methacrylic acid were found to differ greatly.
Whereas the reaction of O3 with CH2=CHC(O)OH gave only
small amounts of glyoxylic acid (<10%), the reaction with
CH2=C(CH3)C(O)OH produced pyruvic acid in high yields (74%).
Formaldehyde was formed in both these reactions with yields
of 148 and 77%, respectively. The high yields of HC(O)H ob-
tained in these reactions are probably due to its production in
the subsequent reactions of the Criegee intermediates. It is in-
teresting to compare the decomposition of the ozonides
formed in the reactions of O3 with CH2=C(CH3)2 and in the re-
action with CH2=C(CH3)C(O)OH. In the former reaction, the
yield of CH3C(O)CH3 has been reported to be close to 35%[6]
while the yield of CH3C(O)C(O)OH obtained in this study from
the ozonolysis of methacrylic acid was around 75%. It is diffi-
cult to rationalise why the yield of HC(O)C(O)OH from the ozo-
nolysis of acrylic acid is so much lower than the yield of
CH3C(O)C(O)OH from the ozonolysis of methacrylic acid. It is
possible that either glyoxylic acid is unstable under the reac-
tion conditions employed, or the Criegee intermediates react
rapidly with this acid even though an excess of formic acid
was present in the system. The high yield of formaldehyde
produced in the reaction of O3 with acrylic acid may provide
some support for these possibilities.
(SCF-MO) calculations that both concerted and stepwise de-
composition processes are equally probable for the cleavage
of the primary ozonide, which lead to the carbonyl and Crie-
gee products. The activation energies for the reactions were
estimated to be less than around 100 kJmolꢁ1. Although the
ozonides formed by addition of O3 to the oxygenated alkenes
studied in this work are energy-rich, it is likely that the relative
importance of the possible decomposition pathways for the
ozonides produced in the reactions of O3 with the oxygenated
alkenes of structures CH2=CHR and CH2=CR1R2 will be depen-
dent to some extent on the relative stabilities of the two Crie-
gee intermediates and carbonyl compounds formed by cleav-
age of the ozonides. It is also possible that the dynamics of
the processes may play an important role in the rate of the de-
compositions. It seems reasonable to assume that the differ-
ence in the relative importance of the decomposition channels
of the ozonides produced in two different reaction systems,
such as from the reaction of O3 with CH2=CHR and CH2=CHR’,
is at least in part a function of the relative stabilities of the
Criegee intermediates RCHOO and R’CHOO, and the primary
carbonyl compounds RC(O)H and R/C(O)H formed on decom-
position of the ozonides, since HCOO and HC(O)H are formed
in both systems. It is suggested, for example, that the relatively
higher yield of ethyl formate generated in the ozonolysis of
CH2=CHOC2H5 compared to the yield of acetaldehyde formed
in the reaction of O3 with CH2=CHCH3 will partly be due to the
lower stability of the C2H5OCHOO Criegee intermediate com-
pared to the CH3CHOO intermediate. The main factor in deter-
mining the stability of the Criegee intermediate is expected to
be the ability of the intermediate to delocalise the two un-
paired electrons in the molecule. Hence, it appears that an
oxygen atom bonded to the electron-bearing carbon atom of
the Criegee intermediate destabilises the intermediate com-
pared to the intermediate CH3CHOO, in which the electron-
bearing carbon atom is bonded to a carbon atom.
It is apparent from the yields of the stable primary carbonyl
products produced in the gas-phase ozonolysis of the series of
unsaturated oxygenated compounds investigated in this study
that the relative yield of the two possible carbonyl compounds
is a function of the structure of the oxygenated substituent
group (Table 3). The formation of the carbonyl and Criegee
products from the ozonide may occur in a one-step concerted
process involving simultaneous cleavage of one of the OꢁO
bonds and the CꢁC bond [reactions (8) and (9)]. Alternatively,
it is possible that the carbonyl and Criegee products are gener-
ated via two consecutive reactions. The first step leads to ring
opening of the ozonide by breaking of one of the OꢁO bonds
to form a peroxy alkoxy diradical [reactions (10) and (11)], fol-
lowed by decomposition of the biradical giving the carbonyl
and Criegee products [reactions (12) and (13)]:
4. Atmospheric Implications
The kinetic and product yield data determined for the reac-
tions of O3 with a series of unsaturated oxygenated com-
pounds in this investigation provide information on the atmos-
pheric fate and impact of emissions of these compounds. With
a
typical tropospheric O3 concentration of 1.2ꢁ
1012 moleculecmꢁ3 (50 ppb), the rate coefficients from this
work can be used to estimate the tropospheric lifetimes with
respect to reaction with O3. The lifetime of ethyl vinyl ether is
about 3 h, which is similar to that reported for removal by re-
action with OH radicals.[20] The lifetimes for the unsaturated
oxygenated compounds containing a carbonyl group are con-
siderably longer and vary from approximately 4 days for
methyl vinyl ketone to 50 days for acrolein. For these carbonyl
compounds, the lifetimes for loss by reaction with OH radicals
are less than 1 day,[2] and hence OH-initiated oxidation will
dominate. The relatively short lifetimes of the unsaturated oxy-
genated compounds show that once released or formed in the
troposphere, they will be rapidly removed in the gas phase
and contribute to local or regional oxidant formation. Reac-
The relative yields of the two primary carbonyl products
R1R2C(O) and R3R4C(O) in the two-step process will be deter-
mined by the relative yields of the biradicals formed in reac-
tions (10) and (11).
The reaction of O3 with alkenes is sufficiently energetic to
produce an energy-rich ozonide, which will rapidly decompose
by a facile ring-opening process. Hiberty and Devidal[35] have
shown from ab initio self-consistent field molecular orbital
ChemPhysChem 2010, 11, 4069 – 4078
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
4077

A. Mellouki et al.
[13] R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson, Jr., J. A. Kerr, M. J.
Rossi, J. Troe, J. Phys. Chem. Ref. Data 1999, 28, 191.
[14] C. A. Taatjes, G. Meloni, T. M. Selby, A. J. Trevitt, D. L. Osborn, C. J. Perciv-
al, D. E. Shallcross, J. Am. Chem. Soc. 2008, 130, 11883.
[15] P. Neeb, A. Kolloff, S. Koch, G. K. Moortgat, Int. J. Chem. Kinet. 1998, 30,
769.
[16] R. Meller, W. Raber, J. N. Crowley, M. E. Jenkin, G. K. Moortgat, J. Photo-
chem. Photobiol. A 1991, 62, 163.
[17] B. Picquet-Varrault, J. F. Doussin, R. Durand-Jolibois, P. Carlier, Phys.
Chem. Chem. Phys. 2001, 3, 2595.
[18] J. Treacy, M. Curley, J. Wenger, H. Sidebottom, J. Chem. Soc. Faraday
Trans. 1997, 93, 2877.
[19] E. Grosjean, D. Grosjean, Int. J. Chem. Kinet. 1998, 30, 21.
[20] G. Thiault, R. Thꢂvenet, A. Mellouki, G. Le Bras, Phys. Chem. Chem. Phys.
2002, 4. 613.
tions of the unsaturated oxygenated compounds with O3 have
been shown to lead to the formation of formaldehyde, esters,
dicarbonyls, acids and anhydrides. Formaldehyde and the di-
carbonyls will be rapidly lost mainly by photolysis, while esters
will be removed by reaction with OH radicals to form acids or
anhydrides.[2] The acids and anhydrides produced in these re-
actions are highly soluble, and may be incorporated into cloud
droplets or form secondary organic aerosols by homogeneous
nucleation. Reactions of the unsaturated oxygenated com-
pounds with OH radicals are expected to give rise to similar
products to those observed in the ozonolysis reactions, and
therefore will lead to similar atmospheric effects to those dis-
cussed for the ozonolysis of these compounds.
[21] R. Atkinson, S. M. Aschmann, A. M. Winer, J. N. Pitts, Jr., Int. J. Chem.
Kinet. 1981, 13, 1133.
[22] J. Treacy, M. El Hag, D. O’Farrell, H. Sidebottom, Ber. Bunsen-Ges. 1992,
96, 422.
[23] D. Grosjean, E. Grosjean, E. L. Williams II, Int. J. Chem. Kinet. 1993, 25,
783.
Keywords: atmospheric chemistry · kinetics · ozone · reaction
mechanisms · unsaturated oxygenated compounds
[24] D. Grosjean, E. L. Williams II, E. Grosjean, Environ. Sci. Technol. 1993, 27,
830.
[1] H. Singh, Y. Chen, A. Standt, D. Jacob, D. Blake, B. Heikes, J. Snow,
Nature 2001, 410, 1078.
[25] H. B. Munshi, K. V. S. R. Rao, R. M. Iyer, Atmos. Environ. 1989, 23, 1971.
[26] D. H. Stedman, C. H. Wu, H. Niki, J. Phys. Chem. 1973, 77, 2511.
[27] S. M. Japar, C. H. Wu, H. Niki, J. Phys. Chem. 1974, 78, 2318.
[28] S. M. Japar, C. H. Wu, H. Niki, J. Phys. Chem. 1976, 80, 2057.
[29] E. C. Tuazon, S. M. Aschmann, J. Arey, R. Atkinson, Environ. Sci. Technol.
1997, 31, 3004.
[30] B. Klotz, I. Barnes, T. Imamura, Phys. Chem. Chem. Phys. 2004, 6, 1725.
[31] E. Grosjean, D. Grosjean, J. Atmos. Chem. 1997, 27, 271.
[32] T. Gierczak, J. B. Burkholder, R. K. Talukdar, A. Mellouki, S. B. Barone, A. R.
Ravishankara, J. Photochem. Photobiol. A 1997, 110, 1.
[33] M. A. Teruel, M. B. Blanco, G. R. Luque, Atmos. Environ. 2007, 41, 5769.
[34] I. W. M. Smith, A. R. Ravishankara, J. Phys. Chem. A 2002, 106, 4798.
[35] P. C. Hiberty, J. P. Devidal, Tetrahedron 1979, 35, 1015.
[2] A. Mellouki, G. Le Bras, H. Sidebottom, Chem. Rev. 2003, 103, 5077.
[3] E. Grosjean, D. Grosjean, J. Atmos. Chem. 1999, 32, 205.
[4] R. Atkinson, W. P. L. Carter, Chem. Rev. 1984, 84, 437.
[5] R. Atkinson, J. Phys. Chem. Ref. Data 1997, 26, 215.
[6] 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.
[7] D. Johnson, G. Marston, Chem. Soc. Rev. 2008, 37, 699.
[8] R. Atkinson, J. Phys. Chem. Ref. Data 1991, 20, 459.
[9] R. P. Wayne, I. Barnes, P. Biggs, J. P. Burrows, C. E. Canosa-Mas, J. Hjorth,
G. Le Bras, G. K. Moortgat, D. Perner, G. Poulet, G. Restelli, H. Sidebot-
tom, Atmos. Environ. 1991, 25A, 1.
[10] R. Atkinson, Chem. Rev. 1986, 86, 69.
[11] R. Atkinson, J. Phys. Chem. Ref. Data 1989, Monograph 1.
[12] R. Atkinson, J. Phys. Chem. Ref. Data 1994, Monograph 2.
Received: May 21, 2010
Published online on August 30, 2010
4078
www.chemphyschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2010, 11, 4069 – 4078