Reactions of alkenes with ozone in the gas phase: A matrix-isolation study of secondary ozonides and carbonyl-containing reaction products

Article abstract of DOI:10.1016/S1386-1425(00)00368-1

Gas phase ozonolysis reactions of the alkenes ethene, cis- and trans-but-2-ene, isoprene and the monoterpenes α-pinene, β-pinene, 3-carene, limonene and β-myrcene have been carried out and the reaction products have been trapped in O₂-doped-argon matrices onto a CsI window held at 12 K. Products have been identified by IR spectroscopy. Comparison with previous matrix spectra, where secondary ozonides have been generated either in situ by annealing or in solution reactions allows a positive identification of the secondary ozonides of ethene and of cis and trans-but-2-ene to be made. These observations are backed up by experiments utilizing the isotopes ¹³C and ²H (D). It appears that secondary ozonides have also been formed from isoprene and the range of monoterpenes studied; this hypothesis is based upon the similarity of spectral features seen in the products of these reactions within those of the simpler alkenes. A number of other primary and secondary products are also identified from these reactions. Ethene gives formaldehyde as a primary product and acetaldehyde as a secondary product; it is found that the yield of acetaldehyde compared to formaldehyde increases as the reaction times are increased. Formaldehyde, one of the expected primary products, is formed by ozonolysis of β-pinene, although the other expected primary product, nopinone, is not seen. A range of secondary reaction products have been identified from the ozonolysis of the monoterpenes studied.

Full text of DOI:10.1016/S1386-1425(00)00368-1

Spectrochimica Acta Part A 56 (2000) 2605–2616
www.elsevier.nl/locate/saa
Reactions of alkenes with ozone in the gas phase: a
matrix-isolation study of secondary ozonides and
carbonyl-containing reaction products
Emma J. Feltham, Matthew J. Almond *, George Marston, Vivienne P. Ly,
Karen S. Wiltshire
Department of Chemistry, Uni6ersity of Reading, Whiteknights, Reading RG6 6AD, UK
Received 3 April 2000; accepted 13 April 2000
Abstract
Gas phase ozonolysis reactions of the alkenes ethene, cis- and trans-but-2-ene, isoprene and the monoterpenes
a-pinene, b-pinene, 3-carene, limonene and b-myrcene have been carried out and the reaction products have been
trapped in O2-doped-argon matrices onto a CsI window held at 12 K. Products have been identified by IR
spectroscopy. Comparison with previous matrix spectra, where secondary ozonides have been generated either in situ
by annealing or in solution reactions allows a positive identification of the secondary ozonides of ethene and of cis-
2
and trans-but-2-ene to be made. These observations are backed up by experiments utilising the isotopes 13C and H
(D). It appears that secondary ozonides have also been formed from isoprene and the range of monoterpenes studied;
this hypothesis is based upon the similarity of spectral features seen in the products of these reactions within those
of the simpler alkenes. A number of other primary and secondary products are also identified from these reactions.
Ethene gives formaldehyde as a primary product and acetaldehyde as a secondary product; it is found that the yield
of acetaldehyde compared to formaldehyde increases as the reaction times are increased. Formaldehyde, one of the
expected primary products, is formed by ozonolysis of b-pinene, although the other expected primary product,
nopinone, is not seen. A range of secondary reaction products have been identified from the ozonolysis of the
monoterpenes studied. © 2000 Elsevier Science B.V. All rights reserved.
Keywords: Matrix-isolation; Atmospheric chemistry; Ozonolysis; Biogenic volatile organic compounds; Infrared spectroscopy
1. Introduction
the importance of such reactions in the chemistry
of the troposphere. In particular, reactions of the
so-called biogenic volatile organic compounds
(Bio-VOCs) are of particular interest [1,2]. The
Bio-VOCs include methane, non-methane hydro-
carbons (NMHCs) and other organic compounds
including oxygen-containing species. The NMHCs
There is much current interest in the chemical
reactions of ozone with alkenes, in part because of
* Corresponding author. Fax: +44-1-189316332.
E-mail address: m.j.almond@rdg.ac.uk (M.J. Almond).
1386-1425/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S1386-1425(00)00368-1

2606
E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
via a primary ozonide, which then decomposes to
give a carbonyl product together with a so-called
Criegee intermediate. The Criegee intermediate
may react by one of several possible routes; it may
react with the carbonyl product to give a sec-
ondary ozonide; it may be stabilised by a third
body (M) to give a carbonyl oxide; last, it may
decompose to give OH radicals. It may be noted
that the formation of secondary ozonide is un-
likely under tropospheric conditions [1,2]. The
secondary ozonide or carbonyl oxide will, of
course, themselves react further to yield other
stable products. These reactions are summarised
for the alkene ethene in Scheme 1.
include isoprene (I) and a class of alkenes known
as monoterpenes, (e.g. II–VI) which are emitted,
in particular, by conifer trees [1].
The products of the reactions of BioVOCs with
ozone are involved in global tropospheric chem-
istry and their effects are far-reaching [3]. Some
products are involved in ozone formation and
destruction, affecting the oxidising capacity of the
atmosphere [3]. Others are involved in the forma-
tion of aerosols which may lead to clouds and
Monoterpenes are destroyed in the troposphere
by reaction with ozone (alongside other reactions
such as those with the hydroxyl radical and the
nitrate radical). Ozonolysis is known to proceed
Scheme 1.

E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
2607
photochemical smogs. At present little is known
about the detailed mechanism of BioVOC oxida-
tion [4,5].
tween ethene and ozone was observed, the pri-
mary and secondary ozonides of ethene could be
formed in a CO2 matrix at temperatures of 25 K
or above. Andrews et al. have extended this work
to the alkenes propene, trans-but-2-ene and
methylpropene (isobutene) where primary and
secondary ozonides are formed upon annealing
ozone-doped xenon matrices to 80–100 K [10]. In
a separate series of experiments Hawkins et al.
[11] and Kuehne et al. [12] have generated isoto-
pomers of the ethene secondary ozonides incorpo-
rating the isotopes 18O and D. These species were
prepared in solution prior to trapping in argon
matrices.
Previous gas phase studies of alkene ozonolysis
have not been linked to matrix isolation with the
exception of studies carried out by Kuehne et al.
to investigate the ozonolysis of both ethene and
cis-but-2-ene using microwave spectroscopy of the
gas phase mixture and subsequent matrix isola-
tion. The secondary ozonides of ethene and cis-
but-2-ene were successfully isolated in these
experiments and a range of primary and sec-
ondary products were seen [13–15]. Other gas
phase ozonolysis studies have mainly concerned
the kinetics of reaction and investigation of prod-
ucts. Atkinson has written several comprehensive
reviews of the current available kinetic data of
such reactions [16,17]. Several reports focus on
the generation of hydroxyl radicals [6,7,18–21].
Examples of other recent product studies include
the gas-phase ozonolysis of trans-but-2-ene, car-
ried out by Fajgar et al. [22], where the secondary
ozonide was detected by GC-MS, and a study of
the products formed in the gas-phase ozonolysis
of cis- and trans-but-2-ene carried out by Horie et
al. [23].
The aim of the work which we describe in this
paper was to carry out the gas phase ozone
reaction of ethene, cis- and trans-but-2-ene, iso-
prene (I) and the monoterpenes a-pinene (II),
b-pinene (III), 3-carene (IV), limonene (V) and
b-myrcene (VI), then to trap the products in a
low-temperature matrix. We anticipated that we
would not trap the primary ozonide products or
Criegee intermediates; these have lifetimes in the
gas phase which would be too short to allow
trapping by our experimental procedures [6].
However, we hoped to find infrared spectroscopic
evidence for the secondary ozonides, whose life-
times are known to be much longer [6,7], and for
other stable oxidised products. These include pri-
mary products (i.e. the aldehydes formed by sim-
ple ozonolysis of the double bond e.g.
formaldehyde from ethene) and secondary prod-
ucts (i.e. those formed by secondary reactions, e.g.
acetaldehyde from ethene). In this way we hoped:
(i) to obtain evidence for the formation of sec-
ondary ozonides of these alkenes in the gas phase;
(ii) to learn something of the product distributions
from these reactions; (iii) to identify primary and
secondary reaction products and hence to learn
something of the reaction mechanisms.
There have been a number of previous matrix
isolation studies of the ozonolysis of alkenes.
These have focused upon the simpler alkenes and
there has been no previous work on the monoter-
penes. Moreover, the methods used to generate
secondary ozonides in these experiments have cen-
tred upon two methods: (i) annealing of a solid
matrix containing the alkene and ozone and (ii)
prior formation of the ozonide by a solution
reaction followed by trapping in a low-tempera-
ture matrix. Kohlmiller et al. investigated the
ozonolysis of ethene in a xenon matrix warmed to
80–100 K and hence isolated the primary and
secondary ozonides of ethene alongside a range of
other products including formaldehyde and ethyl-
ene oxide [8]. Samuni et al. have carried out
similar experiments in argon and carbon dioxide
matrices [9]. These workers found that while an
argon matrix evaporated before any reaction be-
The strategy in our experiments was, therefore,
to isolate the secondary ozonides of ethene and
cis- and trans-but-2-ene where we could compare
our spectroscopic data with spectra previously
obtained for ozonides produced by solution reac-
tions or by annealing. Once we had found the
correct conditions to isolate ozonides of these
simple alkenes we could then move on to consider
the monoterpenes, where no previous spectro-
scopic data for the secondary ozonides exists.

2608
E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
2. Experimental
3. Results and discussion
The matrix isolation apparatus used at Read-
ing has been described elsewhere [24,25]. Matrices
were deposited onto a CsI window which was
cooled to ꢀ12 K by an Air Products 2020SP
‘Displex’ closed-cycle helium refrigerator. Pres-
sures of ꢀ1×10−7 Torr were maintained by
means of an Edwards model E02 oil–vapour dif-
fusion pump backed by an Edwards model
ED100 oil-sealed rotary pump. Infrared spectra
were recorded using a Perkin–Elmer 983 spec-
trometer with a typical resolution of 2 cm−1 or
better.
There were two major aims of the work de-
scribed in this paper. First, to identify the sec-
ondary ozonides from
a range of alkenes
including isoprene and a number of monoterpe-
nes, generated by the gas phase reaction of the
alkene with ozone and trapped in a low-tempera-
ture matrix. Second, to identify reaction products
and hence to learn something of the mechanisms
of these reactions. These two areas will be consid-
ered in separate sections.
3.1. Formation of secondary ozonides
Gaseous samples of ethene (purity 99.5+%)
and cis- and trans-but-2-ene (purities 99+%)
were obtained from Aldrich and were used as
supplied. Liquid samples of isoprene (99%), a-
pinene (98%), b-pinene (99%), 3-carene (90%),
limonene (97%), b-myrcene (85%), methyl vinyl
ketone (99%), methacrolein (95%) and nopinone
(98%) were obtained from Aldrich and were de-
gassed by the ‘freeze-pump-thaw’ method prior to
use. The isotopically-labelled compounds C2D4
(99 at.% D) and 13C2H4 (99 at.% 13C) were ob-
tained from Aldrich and CKGas Products respec-
tively. The matrix gases argon (99.998+%) and
oxygen (99.6+%) were purchased from Aldrich.
Ozone was generated by an electric discharge in a
flow of oxygen using a Fischer Ozone Generator
model 502. Ozone was obtained as a 5% mixture
in O2.
Gaseous mixtures were prepared on an all-glass
vacuum line fitted with Young’s greaseless taps.
The line was maintained at a pressure of ꢀ1×
10−5 Torr by means of a water-cooled mercury
vapour diffusion pump backed by an oil-sealed
rotary pump. Pressures of gaseous samples were
measured using a mercury manometer. For liquid
samples the temperature was adjusted to give a
vapour pressure of 1 Torr, the vapour was trans-
ferred on the line to a 1 l Pyrex bulb and an
excess of argon gas was added.
Mixtures of ozone (5%) in oxygen and ethene
(1%) in argon were allowed to mix for a period of
1 h, then deposited onto a CsI window held at 12
K via a Pyrex tube fitted with a needle valve
which was opened to 1/3 of its maximum opening
to allow a slow rate of deposition. A number of
bands were observed in the infrared spectrum of
the resulting condensate which are not seen in the
spectra of either of the starting materials. In
order to identify which bands result from
ozonides the deposit was subject to broad-band
UV–vis photolysis. It is known that such
ozonides are photolabile [8,11]; thus their concen-
tration will decrease upon photolysis. In Fig. 1
is illustrated the region 1250–900 cm−1 of the
infrared spectrum of the matrix generated in our
experiments before and after 20 min broad-band
UV–vis photolysis. The decay in intensity of
bands assigned to the secondary ozonide can
clearly be seen. As mentioned in the introduction
it is unlikely that primary ozonides could be iso-
lated in this way. Thus, it seems most likely that
the observed bands arise from the secondary
ozonide. Further confirmation that these bands
do indeed belong to the secondary ozonide comes
from a comparison with literature data for this
matrix-isolated species. In these previous studies
the ozonide was generated in situ in argon ma-
trices containing ethene and ozone by annealing
of the matrix. In our experiments the ozonides
were generated in the gas phase and were trapped
in a mixed argon/oxygen matrix (oxygen was
necessarily an ingredient of our matrices as we
Samples were deposited onto the cooled matrix
window by slow spray on. The Pyrex spray on
tube was fitted with a needle valve; this could be
closed to allow gaseous samples to mix in a
mixing chamber for a fixed period of time, then
opened to deposit the sample.

E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
2609
Fig. 1. The region 1250–900 cm−1 of the transmission infrared spectrum of a matrix initially containing products of the gas phase
reaction of ethene and ozone: (a) after mixtures of ethene (1%) in argon and ozone (5%) in oxygen had been allowed to mix for 1
h and the resulting mixture deposited onto a CsI window held at 12 K; (b) after the condensate had been subject to 20 min
broad-band UV–vis photolysis.
obtained O3 as a dilute mixture with O2). In Table
1 are listed the bands that we assign to the
secondary ozonide of ethene in our experiments.
Here these bands are compared with literature
data from three previous experiments [8,11,13]. It
may be seen that there is a good correlation
between the different sets of data. The maximum
observed shift in position of any of the bands is
3%; most bands show shifts of 1% or less. These
modest shifts may be accounted for by the change
in matrix material from pure argon to mixed
argon/oxygen.
This is good evidence that we have indeed
isolated the secondary ozonide of ethene from a
gas phase reaction. Further evidence is supplied
by the results of isotopic substitution experiments.
We have generated the same ozonide product
from 13C2H4 and from C2D4. The bands observed
in these experiments are listed in Tables 2 and 3
where they are compared with previous work on
the same species [8,11,12]. Again a very good
correlation with the literature spectra is obtained
with discrepancies in band position from one ex-
periment to another of 3% or less; for most bands
less than 1%.
We then moved on to study the reactions of cis-
and trans-but-2-ene with ozone. These com-
pounds are expected to be more reactive towards
ozone than is ethene. The rate constants for the
gas phase reactions of ethene and cis- and trans-
but-2-ene with ozone [17] are: kozone (cm3
molecule−1
s
−1): ethene 1.59×10−18; cis-but-2-
ene 1.25×10−16; trans-but-2-ene 1.90×10−16
.
Indeed some reaction occurred if mixtures of
ozone (5%) in O2 and cis- or trans-but-2-ene (1%)
in argon were codeposited via a tube fitted with a
constriction ꢀ0.5 mm in diameter. However, a
larger yield of the secondary ozonide was seen if
the same procedure as used for ethene was
adopted, i.e. that the reagents were allowed to mix

2610
E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
Table 1
Positions of infrared bands (in cm−1) assigned to the secondary ozonide of ethene
This work
Kuehne et al. [13]
Kohlmiller et al. [8]
Hawkins et al. [11]
1126
1090
1021
1002
915
1129
1078
1029
1001
926
1130
1079
1030
–
1129
1078
1029
–
–
–
833
808
806
808
Table 2
Table 4
Positions of infrared bands (in cm−1) assigned to the sec-
ondary ozonide of 13C2H4
Positions of infrared bands (in cm−1) assigned to the sec-
ondary ozonides of cis- and trans-but-2-ene
This work
Kohlmiller et al. [8] Hawkins et al. [11]
This worka
Kuehne et
al.a [15]
This workb
Andrews et
al.b [10]
2845
2710
1119
1089
1000
893
–
–
1119
1059
1009
–
2889
2706
1119
1058
1008
912
1391
1144
1116
1076
889
1393
1147
1119
–
893
–
1392
–
1115
1075
–
1392
1146
1117
1075
895
847
–
800
805
805
852
523
852
523
522
a Data from cis-but-2-ene experiments.
b Data from trans-but-2-ene experiments.
Table 3
Positions of infrared bands (in cm−1) assigned to the sec-
ondary ozonide of C2D4
erated by annealing of an argon matrix. We may
therefore conclude that we have successfully gen-
erated the secondary ozonides of cis- and trans-
but-2-ene in these experiments.
This work
Hawkins et al. [11]
Ku¨hne et al. [12]
2175
2148
2085
1158
1148
1112
1053
1021
979
2181
–
2091
1159
1142
–
1057
1020
981
972
848
2181
2161
2092
1160
1143
1135
1059
1021
980
Having established a method which clearly al-
lows formation and trapping of secondary
ozonides from alkenes we turned our attention to
isoprene and a range of monoterpenes. The
monoterpenes selected for study were a-pinene
(II), b-pinene (III), 3-carene (IV), limonene (V)
and b-myrcene (VI). The reactions of these com-
pounds with ozone are of much interest in an
atmospheric chemistry context [26–33] but they
had not previously been studied by matrix isola-
tion. In the current experiments the ozonides
were generated by the gas-phase reaction between
the alkene (1% in Ar) and ozone (5% in O2).
These ozonides were formed by allowing the two
gaseous samples to mix for a period of one hour,
then the resulting mixture was deposited onto
967
846
972
851
for 1 h then deposited via a needle valve. In these
experiments we again saw bands arising from
photolabile species and again these bands corre-
late well (Table 4) with bands seen in previous
experiments [10,15] where the ozonides were gen-

E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
2611
Table 5
Positions of infrared bands assigned to secondary ozonides of
isoprene and the monoterpenes a-pinene, b-pinene, 3-carene,
limonene and b-myrcene
Alkene
Ozonide band positions (cm−1
)
Isoprene
a-Pinene
b-Pinene
1234(w), 1223(w), 1124(s), 998(w) and 930(m)
1025(sh)
1337(w), 1283(w), 1234(w), 1170(m), 1125(m),
1020(m) and 1001(m)
1126(m), 1020(w) and 532(m)
1181(m), 1126(m), 1112(m), 1021(m), 1001(w)
994(w) and 930(m)
Limonene likewise has two double bonds, but
here the double bonds are quite different in char-
acter. One forms part of the ring, the other is
exocyclic. Gas phase studies [17] have shown that
the double bond within the ring reacts ꢀ10–
100× faster under similar conditions. It there-
fore seems likely that this ozonide (IX) would be
formed preferentially over the alternative (X).
3-Carene
Limonene
b-Myrcene 1234(m), 1125(m) and 1021(m)
a CsI window held at 12 K through a Pyrex tube
via a needle valve. In each case infrared bands
were seen in the condensate which decayed upon
broad-band UV-vis photolysis and which occur
in spectral regions where bands of the known
secondary ozonides of ethene and of cis- and
trans-but-2-ene appear. The positions of these
bands are listed in Table 5; bands assigned to the
secondary ozonide of isoprene are shown in Fig.
2.
b-Myrcene has three double bonds so three sec-
ondary ozonides (XI–XIII) are possible. Of these
the formation of XIII at the most substituted
double bond would be most likely as this double
bond would be expected to react ꢀ100× faster
[17] than the less-substituted bonds.
Of course, these assignments to secondary
ozonides must remain somewhat tentative, but
the fact that these bands occur in spectral regions
where secondary ozonides are known to absorb
and that the species giving rise to the bands are
clearly photolabile must point towards their for-
mation. In many cases fewer bands are seen for
these more complex alkenes which may be as-
signed to ozonides. This observation is accounted
for by the complexity of the spectra of the start-
ing materials which mean that large sections of
the spectrum are obscured. With some of the
compounds studied here there is more than one
double bond present in the alkene and therefore
more than one possible secondary ozonide
product. An example is provided by isoprene (I),
where there are two double bonds and therefore
two possible secondary ozonide products (VII
and VIII).
Seven bands, which may be assigned to sec-
ondary ozonides are seen for limonene and three
bands for b-myrcene. However, it is not possible
to state with certainty, on the basis of the exper-
imental results alone, whether one or more
ozonide has been formed.
3.2. Identification of other reaction products
The second aim of our experiments was to
identify the other reaction products formed in

2612
E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
1496(m), 1283(m) and 1174(m) cm−1 [13–15,34,
35]. The other most abundant carbonyl product
observed in these experiments is acetaldehyde,
identified by bands at 2730(vw), 1725(m), 1395(w),
1350(m), 1120(m) and 750(w) cm−1 [13–15,34,35].
Acetaldehyde has been observed as a secondary
product of this reaction in previous gas phase and
matrix isolation studies. In Fig. 3 are shown the
regions 1800–1500 cm−1 spectra of matrices
formed by condensing gaseous mixtures of ethene
(1%)/argon and ozone (5%)/oxygen which have
been allowed to mix for (a) 1 h and (b) 3 h prior
to deposition. It is clear that the relative yield of
the secondary product (acetaldehyde) increases
over the primary product (formaldehyde) as the
mixing time is lengthened. Our experiments
utilising 13C2H4 and C2D4 help to confirm the
identities of formaldehyde and acetaldehyde as
reaction products [35]. In Table 6 are listed the
observed bands of the various isotopomers of
these compounds.
these experiments. In particular we wished to
study the relative abundances of primary and
secondary products. In this way we hoped to learn
something of the reaction mechanism and, in the
cases where more than one secondary ozonide
product is possible, to decide which ozonide had
been formed preferentially.
The simplest reaction studied was that of ethene
with O3. Here the primary carbonyl product is
expected to be formaldehyde. Indeed bands of
formaldehyde can clearly be seen in the spectrum
of the condensate at 2859(vw), 2794(vw), 1747(m),
Other secondary products to be identified
more-or-less certainly are: H2O [13–15,34], CO2
[13–15,35], CO [13–15], the CO…CO2 complex
[36], HCOOH [13–15], (HCO)2O [13–15],
CH3COOH [15,35] and ethylene oxide [35]. In
Table 6
Bands assigned to the 13C2 and D4 isotopomers of acetalde-
hyde [35] and of the 13C and D2 isotopomers of formaldehyde
generated by ozonolysis of 13C2H4 or of C2D4
CH2O
13CH2O
CD2O
n
12C/n13C n1H/nD
2859(vw)
2794(vw)
1747(m)
1496(m)
1283(m)
1174(m)
2789(w)
–
1700(m)
1492(m)
–
2068(w)
–
1730(m)
1283(w)
955(m)
–
0.976
–
0.973
0.997
–
0.723
–
0.990
0.858
0.744
–
1151(m)
0.980
Fig. 2. The region 1800–800 cm−1 of the transmission in-
frared spectra of (a) a matrix formed by condensing (onto a
CsI window at 12 K) the gaseous mixture formed by allowing
isoprene (1%) in argon and ozone (5%) in oxygen to mix for 1
h in the gas phase, (b) of isoprene (1%) isolated in an argon
matrix at 12 K, (c) of methyl vinyl ketone (1%) isolated in an
argon matrix at 12 K and (d) of methacrolein (1%) isolated in
an argon matrix at 12 K. Legend for spectrum (a): a=
CH3COOCHO, b=CH3COOH, c=(HCO)2O, d=CH2O,
e=CH3CHO, f=H2O, g=O2, h=secondary ozonide, i=
O3.
CH3CHO 13CH313CHO CD3CDO
n
12C/n13C n1H/nD
2730(vw)
1725(m)
1395(w)
1350(m)
1120(m)
750(w)
–
–
–
–
1690(sh)
1390(m)
1340(w)
1100(sh)a
–
1700(m)
–
1025(w, sh) 0.993
950(w, sh) 0.982
0.980
0.996
0.986
–
0.759
0.848
–
–
–
a Partially obscured by secondary ozonide band.

E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
2613
Fig. 3. The region 1800–1550 cm−1 of the transmission infrared spectra of matrices formed by condensing (onto a CsI window held
at 12 K) gaseous mixtures formed by allowing mixtures of ethene (1%) in argon and ozone (5%) in oxygen to react in the gas phase
(a) for 1 h, (b) for 3 h.
general the following ‘rules’ were applied in as-
signing infrared bands.
similar results to the mixed O2/Ar matrices were
obtained with only some minor shifts (typically
much less than 1%) in band positions but a very
similar product distribution. Thus, it may be con-
cluded that the presence of O2 in our reaction
mixtures does not have a significant influence on
the chemical route of the reaction. Second, a
mixture of ethene (1%)/argon and O3 (5%)/O2 was
deposited without mixing and the condensate was
subject to broad-band UV–vis photolysis. The
findings from this experiment reproduce those of a
similar experiment previously carried out by
Hawkins and Andrews using a pure argon matrix
[35]. The main difference here is that in the photol-
ysis experiment no secondary ozonides were seen,
but higher concentrations of other products were
obtained when compared with the gas-phase mix-
ing experiments. A particular point is a much
increased yield of ethylene oxide, which is likely to
be formed by the reaction of O (3P) atoms with
ethene [35], the atoms being generated by photoly-
sis of O3. Other products seen in this photolysis
1. There should be a good match between ob-
served spectra and literature spectra for authentic
samples (where possible for matrix-isolated sam-
ples) taking into account small shifts due to
changes in phase or matrix material.
2. The strongest bands of a product should be
visible unless they would obviously be obscured by
bands of another ingredient of the matrix.
3. Ideally several bands, certainly more than one,
should be seen for positive identification of a
product.
The last four compounds listed above were
obtained only in relatively low concentrations,
however, and the bands assigned to these species
were weak in intensity, so here the assignments are
tentative.
Two further experiments were carried out. First
a pure oxygen matrix was used to isolate the
products. This was done in order to assess the
influence of O2 on the reaction. In this case very

2614
E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
crolein (XV) (formed from the secondary ozonide
VIII).
Unfortunately we obtained no spectroscopic evi-
dence for the existence of either of these products
in our experiments. The spectra reproduced in
Fig. 2, illustrating the reaction products of iso-
prene and O3, show no correspondence to the
spectra of authentic matrix-isolated samples of
methyl vinyl ketone and methacrolein which are
also reproduced. It is likely that this finding
reflects the high reactivity of methyl vinyl ketone
and methacrolein (both of which contain a CꢀC
bond) towards O3. In fact several secondary prod-
ucts are identified. These are CH3COOCHO,
CH3COOH, (HCO)2O, CH3CHO, H2O, O2, CO
and CO2. The primary product H2CO is also seen.
Bands at 1234, 1223, 1124, 998 and 930 cm−1 are
assigned to a secondary ozonide of isoprene on
the basis of their disappearance upon photolysis.
In the ozonolysis of b-pinene formaldehyde is
expected as a primary product and this compound
is seen. The other primary aldehyde product is
expected to be nopinone (XVI).
Fig. 4. The region 1800–800 cm−1 of the transmission in-
frared spectra of (a) a matrix formed by condensing (onto a
CsI window at 12 K) the gaseous mixture formed by allowing
b-pinene (1%) in argon and ozone (5%) in oxygen to react in
the gas phase for 1 h, (b) of b-pinene (1%) isolated in an argon
matrix at 12 K and (c) of nopinone (1%) isolated in an argon
matrix at 12 K. Legend for spectrum (a): a=b-pinene, b=O3,
c=(HCO)2O, d=CH2O, e=CH3CHO, f=H2O, g=O2, h=
secondary ozonide.
experiment are: H2O [13–15,34], CO [13–15],
CO2, [13–15,35] CH2O [13–15,34,35], CH3CHO
[13,15,34,35], HCOOH [13–15] and (HCO)2O
[13–15].
Cis- and trans-but-2-ene both give a large num-
ber of products upon photolysis. Formaldehyde is
clearly observed as a primary product. Amongst
other secondary products the following are posi-
tively identified on the basis of their infrared
spectra and applying the ‘rules’ outlined above.
These are H2O, CO, CO2, CH3CHO, CH3COOH,
CH3CO3H, (HCO)2O and CH4. All of these com-
pounds have been observed in previous ozonolysis
studies of these alkenes [10–15].
With isoprene two primary products alongside
formaldehyde (which is observed) are possible.
These are methyl vinyl ketone (XIV) (formed
from the secondary ozonide VII) and metha-
The infrared spectrum of the reaction products
of b-pinene and O3 are reproduced in Fig. 4 where
they are compared with spectra of authentic ma-
trix-isolated samples of b-pinene and nopinone. It
may be seen that while some unreacted b-pinene
remains there is no evidence for the presence of
nopinone. Secondary products which are posi-
tively identified are (HCO)2O, CH3CHO, H2O,
O2, CO and CO2. Bands at 1337, 1283, 1234,
1170, 1125, 1067, 1020 and 1001 cm−1 are as-
signed to a secondary ozonide on the basis of
their disappearance upon photolysis.

E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
2615
Table 7
complexity of the spectra make a definite assign-
ment difficult. However, the evidence obtained
suggests it is possible that, under the conditions
of our experiments, reaction at all three double
bonds takes place. The secondary products
positively identified from each of these reactions
are listed in Table 7.
Secondary products generated by the ozonolysis of isoprene and
the monoterpenes a-pinene, b-pinene, 3-carene, limonene and
b-myrcene
Alkene
Secondary products
Isoprene
H2O, CO2, CO, CH3COOCHO, CH3COOH,
HCOOH, (HCO)2O, CH3CHO and activated
a
O2
a-Pinene
b-Pinene
3-Carene
Limonene
H2O, CO2, CO, CH3COOCHO, CH3COOH,
HCOOH, (HCO)2O, CH2O, CH3CHO and
activated O2
4. Conclusions
a
H2O, CO2, CO, CH3COOCHO, CH3COOH,
HCOOH, (HCO)2O, CH3CHO and activated
The results which we report in this paper
provide a clear indication that the secondary
ozonides of ethene and cis- and trans-but-2-ene
are generated by the gas-phase reaction of the
appropriate alkene with O3 in a gaseous mixture
of argon and oxygen, followed by condensation
at 12 K. This conclusion is reached by a com-
parison of our spectra with previously obtained
spectra where the ozonides are generated by an-
nealing of matrices containing the appropriate
alkene and ozone. The other products of these
reactions may be designated as primary or sec-
ondary products. Thus with ethene, formalde-
hyde is a primary product and acetaldehyde is a
secondary product. An increased yield of acetal-
dehyde over formaldehyde is seen as the reaction
time is increased showing the increased impor-
tance of secondary reactions at longer reaction
times. Isoprene, b-pinene, limonene and b-
myrcene show formaldehyde as a product. This
finding is entirely as expected. For isoprene and
the monoterpenes we see no evidence for the
larger primary aldehyde products presumably be-
cause these compounds are themselves highly
reactive towards O3. Several secondary products,
however, are seen and are identified more-or-less
certainly. These are listed in Table 7. Because of
the difficulty in positively identifying primary
reaction products it is difficult to assess the reac-
tivity of different double bonds within the
molecules studied towards ozone. For b-
myrcene, however, it does appear that reaction
occurs at all three double bonds. This conclusion
is based upon the observation of formaldehyde
as a product which may be formed by ozonolysis
at the less-reactive double bonds.
a
O2
H2O, CO2, CO, CH3COOCHO, CH3COOH,
HCOOH, (HCO)2O, CH2O, CH3CHO and
a
activated O2
H2O, CO2, CO, CH3COOCHO, CH3COOH,
a
(HCO)2O, CH3CHO and activated O2
b-Myrcene H2O, CO2, CO, CH3COOCHO, CH3COOH,
a
(HCO)2O, CH3CHO and activated O2
a Activated O2; O2 trapped in a matrix cage in such a way
that the O–O stretch becomes infrared active.
For a-pinene a ring-opening reaction to give a
complex dialdehyde primary product is expected.
It is not possible to say unequivocally whether
this product is present or not in the reaction
mixture because an authentic sample was not
available. Limonene gives a significant yield of
formaldehyde suggesting that there is some
reaction at the exocyclic double bond. This
double bond is expected to be the less reactive of
the two double bonds in the molecule [17]. It is
almost certain that reaction at the cyclic double
bond leading to ring opening will also occur.
3-Carene, also must ozonolyse via ring opening.
For both 3-carene and limonene infrared spectra
of the possible larger aldehyde primary products,
formed by ring opening, were not available.
b-Myrcene has three double bonds. The
less-substituted double bonds, which are also
expected to be less-reactive [17], would give
formaldehyde as a primary product. The more
substituted double bond should yield acetone.
Formaldehyde is positively identified as
product; acetone may well be present but the
a

2616
E.J. Feltham et al. / Spectrochimica Acta Part A 56 (2000) 2605–2616
[18] R. Atkinson, S.M. Aschmann, J. Arey, B. Shorees, J.
References
Geophys. Res. 97 (1992) 6065.
[19] S.E. Paulson, A.D. Sen, P. Liu, J.D. Fenske, M.J. Fox,
Geophys. Res. Lett. 24 (1997) 3193.
[20] N.M. Donahue, J.H. Kroll, J.G. Anderson, K.L. Demer-
jian, Geophys. Res. Lett. 25 (1998) 59.
[21] G. Marston, C.D. McGill, A.R. Rickard, Geophys. Res.
Lett. 25 (1998) 2177.
[22] R. Fajgar, J. Vitek, Y. Haas, J. Pola, Tetrahed. Lett. 37
(1996) 3391.
[23] O. Horie, P. Neeb, G.K. Moortgat, Int. J. Chem. Kinet.
29 (1997) 461.
[24] M.J. Almond, M. Hahne, J. Chem. Soc. Dalton Trans.
(1988) 2255.
[25] E.J. Feltham, PhD Thesis, University of Reading, 2000.
[26] S.E. Paulson, R.C. Flagan, J.H. Seinfeld, Int. J. Chem.
Kinet. 24 (1992) 103.
[27] A.R. Rickard, D. Johnson, C.D. McGill, G. Marston, J.
Phys. Chem. A 103 (1999) 7656.
[28] P. Neeb, F. Sauer, O. Horie, G.K. Moortgat, Atmos.
Environ. 31 (1997) 1417.
[29] R. Atkinson, J. Arey, S.M. Aschmann, E.C. Tuazon, Res.
Chem. Intermed. 20 (1994) 385.
[30] R.M. Kamens, M.W. Gery, H.E. Jeffries, M. Jackson, E.I.
Cole, Int. J. Chem. Kinet. 14 (1992) 955.
[31] K. Griesbaum, M. Hilss, J. Bosch, Tetrahedron 52 (1996)
14813.
[32] K. Griesbaum, V. Miclaus, I.C. Jung, Environ. Sci. Tech-
nol. 32 (1998) 647.
[33] M. Jang, R.M. Kamens, Atmos. Environ. 33 (1999) 459.
[34] B. Schrader, Raman and Infrared Atlas of Organic Com-
pounds, 2nd edn., VCH Publishers, New York, 1989.
[35] M. Hawkins, L. Andrews, J. Am. Chem. Soc. 105 (1983)
2523.
[1] P. Warneck, Chemistry of the Natural Atmosphere, In-
ternational Geophysics Series, vol. 41, Academic Press,
London, 1988.
[2] R.P. Wayne, Chemistry of Atmospheres, Oxford Univer-
sity Press, Oxford, 1985.
[3] A.M. Thomson, Science 256 (1992) 1157.
[4] R.A. Atkinson, J. Arey, Acc. Chem. Res. 31 (1998) 574.
[5] A. Calogirou, B.R. Larsen, D. Kotzias, Atmos. Environ.
33 (1999) 1423.
[6] M. Olzmann, E. Kraka, D. Kremer, R. Gutbrod, S.
Andersson, J. Phys. Chem. A 101 (1997) 9421.
[7] W.C.D. Rathman, T.A. Claxton, A.R. Rickard, G.
Marston, Phys. Chem. Chem. Phys. 1 (1999) 3981.
[8] C.K. Kohlmiller, L. Andrews, J. Am. Chem. Soc. 103
(1981) 2578.
[9] U. Samuni, R. Fraenkel, Y. Haas, R. Fajgar, J. Pola, J.
Am. Chem. Soc. 118 (1996) 3687.
[10] L. Andrews, C.K. Kohlmiller, J. Phys. Chem. 86 (1982)
4548.
[11] M. Hawkins, C.K. Kohlmiller, L. Andrews, J. Phys.
Chem. 86 (1982) 3154.
[12] H. Kuehne, H.H. Guenthard, J. Phys. Chem. 80 (1976)
1238.
[13] H. Kuehne, S. Vaccani, T.-K. Ha, A. Bauder, H.H.
Guenthard, Chem. Phys. Lett. 38 (1976) 449.
[14] H. Kuehne, S. Vaccani, A. Bauder, H.H. Guenthard,
Chem. Phys. 28 (1978) 11.
[15] H. Kuehne, M. Forster, J. Hulliger, H. Ruprecht, A.
Bauder, H.H. Guenthard, Helv. Chim. Acta 63 (1980)
1971.
[16] R. Atkinson, J. Phys. Chem. Ref. Data, Monograph 2, 1
(1994).
[17] R. Atkinson, J. Phys. Chem. Ref. Data 26 (1997) 215.
[36] V. Raducu, D. Jasmin, R. Dahoo, P. Brosset, B. Gauthier-
Roy, L. Abouaf-Marguin, J. Chem. Phys. 102 (1995) 9235.
.