Infrared Matrix Isolation and Theoretical Studies of Reactions of Ozone with Bicyclic Alkenes: α-Pinene, Norbornene, and Norbornadiene

Article abstract of DOI:10.1021/jp510883k

The reactions of ozone with three bicyclic alkenes, α-pinene, norbornene, and norbornadiene, were studied by low-temperature (14 K), argon matrix isolation infrared spectroscopy including ¹⁸O isotope-labeling studies. Theoretical calculations of some of the proposed reaction intermediates and products were carried out using the Gaussian 09 suite of programs, applying density functional theory (DFT), the B3LYP functional, and the 6-311G++(d,2p) basis set. In the α-pinene/ozone system, the thermal reaction between α-pinene and ozone was too slow to observe under the twin-jet or merged-jet deposition conditions of these experiments. However, red light (λ ≥ 600 nm) irradiation of the argon matrixes containing α-pinene and ozone caused new infrared peaks to appear that could be readily assigned to reaction products of α-pinene with O(³P) resulting from ozone photolysis: α-pinene oxide (with an epoxide ring) and two isomeric ketones. Norbornene and norbornadiene were both found to react with ozone in the gas phase during twin-jet or merged-jet deposition of these mixtures with argon. New peaks observed in the infrared spectra were assigned to the primary ozonides, Criegee intermediates, and secondary ozonides of norbornene and norbornadiene, indicating that the bulk of these reactions proceeded via the "classic" Criegee mechanism for ozonolysis of alkenes. Calculated infrared frequencies and molecular energies support these conclusions. Ultraviolet irradiation of these mixtures resulted in complete decomposition of the early intermediates and the formation of acids, aldehydes, alcohols, carbon dioxide, and carbon monoxide. In any case, no evidence for "unusual" chemistry, prompted by the bicyclic nature of the reactants, was observed. (Figure Presented).

Full text of DOI:10.1021/jp510883k

Article
pubs.acs.org/JPCA
Infrared Matrix Isolation and Theoretical Studies of Reactions of
Ozone with Bicyclic Alkenes: α‑Pinene, Norbornene, and
Norbornadiene
Roger W. Kugel and Bruce S. Ault*
Department of Chemistry, University of Cincinnati, Cincinnati, Ohio 45221, United States
S
* Supporting Information
ABSTRACT: The reactions of ozone with three bicyclic alkenes, α-pinene,
norbornene, and norbornadiene, were studied by low-temperature (14 K),
argon matrix isolation infrared spectroscopy including 18O isotope-labeling
studies. Theoretical calculations of some of the proposed reaction
intermediates and products were carried out using the Gaussian 09 suite of
programs, applying density functional theory (DFT), the B3LYP functional,
and the 6-311G++(d,2p) basis set. In the α-pinene/ozone system, the thermal
reaction between α-pinene and ozone was too slow to observe under the twin-
jet or merged-jet deposition conditions of these experiments. However, red
light (λ ≥ 600 nm) irradiation of the argon matrixes containing α-pinene and
ozone caused new infrared peaks to appear that could be readily assigned to
reaction products of α-pinene with O(3P) resulting from ozone photolysis: α-
pinene oxide (with an epoxide ring) and two isomeric ketones. Norbornene
and norbornadiene were both found to react with ozone in the gas phase
during twin-jet or merged-jet deposition of these mixtures with argon. New peaks observed in the infrared spectra were assigned
to the primary ozonides, Criegee intermediates, and secondary ozonides of norbornene and norbornadiene, indicating that the
bulk of these reactions proceeded via the “classic” Criegee mechanism for ozonolysis of alkenes. Calculated infrared frequencies
and molecular energies support these conclusions. Ultraviolet irradiation of these mixtures resulted in complete decomposition of
the early intermediates and the formation of acids, aldehydes, alcohols, carbon dioxide, and carbon monoxide. In any case, no
evidence for “unusual” chemistry, prompted by the bicyclic nature of the reactants, was observed.
oxide CI has eluded, until recently, direct detection in alkene
ozonolysis reactions.7 We report herein further direct infrared
spectral evidence for an argon-matrix-isolated carbonyl oxide
compound formed during the reactions of ozone with
norbornene and with norbornadiene.
INTRODUCTION
■
Ozone reacts readily with the double bonds in unsaturated
hydrocarbons to ultimately cleave the double bond and
produce aldehydes and/or ketones as well as organic acids,
peroxides, and other fragmentation products including CO and
CO2. The mechanism of the ozonolysis reaction was originally
discovered by Criegee.1 The three-step mechanism that Criegee
articulated in 19532 and summarized in a later review3 involves
first a 1,3-dipolar cyclo-addition of ozone to the CC double
bond, forming a 1,2,3-trioxolane ring, also called a primary
ozonide (POZ). Second, the POZ can undergo ring opening to
form a carbonyl compound and a carbonyl oxide compound,
also now referred to as the Criegee intermediate (CI). Third,
the CI can then recombine with the carbonyl compound to
form the more stable 1,2,4-trioxolane ring, also known as the
secondary ozonide (SOZ). The general reaction sequence is
shown in Scheme 1. Since Criegee’s seminal work, a number of
ozonolysis reactions of alkenes have been shown to follow this
simple three-step mechanism.4,5
In this work, we studied, using argon matrix isolation
techniques, the reactions of ozone with α-pinene (2,6,6-
trimethylbicyclo[3.1.1]hept-2-ene) (3a), norbornene
(bicyclo[2.2.1]hept-2-ene) (5a), and norbornadiene
(bicyclo[2.2.1]hepta-2,5-diene) (8a), shown in Figure 1. The
study was undertaken in an attempt (1) to see if these highly
ring-strained bicyclic compounds reacted by a normal Criegee
mechanism, (2) to see if reaction occurred to isolate early
intermediates in an argon matrix, and (3) to structurally
characterize the early intermediates using infrared spectroscopy,
18O isotope labeling, and ab initio quantum calculations using
density functional theory (DFT), applying the B3LYP func-
tional and the 6-311G++(d,2p) basis set. The results indicated
that (1) the thermal reaction of ozone with α-pinene was too
The chemical nature of the carbonyl oxide CI has been the
subject of some debate. It has been formulated as a zwitterion
in Criegee’s original work1 and in many solution mechanisms4
or as a singlet biradical or closed-shell dioxirane in gas-phase
mechanisms.6 In any case, because of its reactivity, the carbonyl
Received: October 30, 2014
Revised: December 12, 2014
© XXXX American Chemical Society
A
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
a
cycles. Molecular sieves were used to remove water from the α-
pinene, norbornene, and norbornadiene samples. Molecular
sieves were not used to dry the α-pinene oxide samples because
it was observed that the sieves caused significant isomerization
and polymerization of the α-pinene oxide.
Scheme 1. Criegee Mechanism for Ozonolysis
The reactions between ozone (O3) and either α-pinene
(C10H16), norbornene (C7H10), or norbornadiene (C7H8) were
studied using low-temperature matrix isolation/FT-IR techni-
ques. Gas mixtures of ozone with argon and one of the bicyclic
compounds with argon (at mole ratios of approximately Ar/O3
≈ 250 and Ar/bicyclic alkene ≈ 250) were prepared using
standard manometric techniques. The argon/ozone mixtures
and the argon/bicyclic alkene mixtures were prepared in
separate stainless steel vacuum manifolds. Both gas mixtures
were deposited slowly using stainless steel needle valves onto a
cryogenically cooled CsI window held at about 14 K with a
closed-cycle helium cryostat (CTI Cryogenics). Matrix
formation took place over approximately 24 h at an overall
gas deposition rate of about 2.3 mmol/h. Gases were deposited
using one of three different deposition geometries:
(1) Merged-jet (MJ): the Ar/O3 and Ar/bicyclic alkene gas
mixtures were combined at an Ultra-Torr tee on their way to
the cold window and flowed together through a Teflon merge
tube whose length could be varied. This geometry was useful
for studying later intermediates and stable products of the
thermal gas-phase (or surface-mediated) reactions.
a
Step (1): Reaction of ozone with an alkene to form a five-membered
ring, 1,2,3-trioxolane POZ. Step (2): Ring opening of the POZ to form
a carbonyl compound and a carbonyl oxide, CI. Step (3): Re-
formation of a rearranged five-membered ring, 1,2,4-trioxolane SOZ.
(2) Concentric-jet (CJ): the Ar/O3 and Ar/bicyclic alkene
gas mixtures were combined concentrically with a 1/8 in. OD
Teflon tube flowing inside of a 1/4 in. OD Teflon tube. The
longitudinal distance between the tube ends could also be
varied. This geometry was useful for studying less stable
intermediates.
(3) Twin-jet (TJ): the Ar/O3 and Ar/bicyclic alkene gas
mixtures were fed to the cold window through separate ports in
the cold head. This geometry minimized the gas-phase mixing
time of the two gas mixtures and was useful for trapping the
most unstable early intermediates in the gas-phase reactions.
The frozen matrixes formed during deposition were studied
by FT-IR spectroscopy (PerkinElmer Spectrum One) and were
subsequently irradiated and/or annealed to see the effects of
light and/or temperature on the observed reaction intermedi-
ates and products. Light sources for photochemical reactions
included a medium-pressure mercury arc lamp (Karl Zeiss
#470, Germany), a red laser diode (λ ≈ 650 nm), and a 100 W
incandescent bulb with either glass “cut-off” filters (Corning red
glass filter #2418, transmits λ ≥ 600 nm; Schott infrared glass
filter RG1000, transmits λ ≥ 1000 nm) or a 1 mm silicon wafer
(λ ≥ 1100 nm).
Figure 1. Three bicyclic alkenes: (3a) α-pinene, (5a) norbornene, and
(8a) norbornadiene.
slow to observe under the gas-phase deposition conditions of
these experiments. However, a photochemical reaction was
observed, with red (λ ≥ 600 nm) or infrared (λ ≥ 1000 nm)
irradiation, that appears to be due to the reaction of
photochemically generated O(3P) with α-pinene. (2) The
thermal reactions of ozone with norbornene and norbornadiene
occurred in the gas phase under all deposition conditions. The
reactions were evidenced by new peaks that appeared in the
infrared spectra of these matrixes upon deposition at 14 K and
increased upon annealing to 25 K. These peaks could be
assigned to the norbornene and norbornadiene POZs, CIs, and
possibly SOZs and acid−aldehyde products.
Quantum calculations were carried out with the Gaussian 09
suite of programs8 using DFT and the B3LYP functional with
the 6-311G++(d,2p) basis set. Empirical scaling factors were
not used to correct calculated vibrational frequencies for
intrinsic errors. Calculated vibrational peak assignments were
made if the calculated and experimental frequencies agreed
within 3−4% and the 18O isotope shifts agreed within 1−3
cm−1. Gaussian 09 calculations were carried out online at the
Ohio Supercomputer Center (OSC) in Columbus, OH.
EXPERIMENTAL SECTION
■
The oxygen (O2) and argon (Ar) gases were supplied by
Wright Brothers, Inc., Cincinnati, OH. 18O isotopically
enriched oxygen (18O2, 99 atom % 18O) was supplied by
Aldrich Chemical Co., Milwaukee, WI. Ozone (O3 or 18O3) was
produced by a high-frequency Tesla coil discharge through
oxygen (O2 or 18O2) gas while being condensed with liquid
nitrogen. To minimize ozone explosion hazard, the pressure of
ozone gas at room temperature was never allowed to exceed 5.0
in. of Hg (127 Torr). α-Pinene and norbornene were obtained
from Acros Organics, Thermo-Fisher Scientific, Pittsburgh, PA,
and α-pinene oxide and norbornadiene were obtained from
Sigma-Aldrich Chemical Co., St. Louis, MO. The organic
compounds were all purified by multiple freeze−pump−thaw
RESULTS AND DISCUSSION
■
A. α-Pinene. Using a MJ deposition geometry, dilute
mixtures of α-pinene in argon (Ar/α-pinene = 200) and ozone
in argon (Ar/O3 = 200) were co-deposited in the dark onto the
CsI window chilled to 14 K. Infrared spectra of the resulting
B
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
solid argon matrixes revealed no new peaks, other than those at
701, 1032, and 2099 cm−1 attributable to ozone molecules,
presumably perturbed by their proximity to α-pinene
molecules. Figure S1 (Supporting Information) shows a partial
infrared spectrum of a dark-deposited Ar/α-pinene/O3 = 400/
1/1 matrix. The results suggest that the gas-phase reaction
between ozone and α-pinene is too slow to observe under the
conditions of these experiments. The results are also consistent
with the known rate constant for the bimolecular gas-phase
reaction between ozone and α-pinene, which is reported9 to be
k298 K = 8.66 × 10−17 cm3/molecule−s. Previous experience in
our laboratory has indicated that reactions of ozone with a
bimolecular rate constant of less than about 1 × 10−16 cm3/
molecule−s were too slow to produce significant amounts of
products in a TJ matrix experiment. On the other hand, the
ozone−α-pinene-containing argon matrixes did show evidence
of association of these molecules in the solid state by the red-
shifted infrared bands assigned to perturbed ozone molecules.
(Figure S1 (Supporting Information) shows a partial infrared
spectrum of the dark-deposited argon matrix). However,
despite the fact that this reaction is known to have a relatively
low activation energy of Ea = 1.45 kcal/mol,10 there is
apparently insufficient thermal energy available for chemical
reaction by the time molecular association occurs in the nascent
argon matrix (for reference, RT = 0.17 kcal/mol at 84 K, the
melting point of solid argon). However, these association
complexes did show evidence of significant photosensitivity.
When the argon matrixes containing α-pinene and ozone
were irradiated with either infrared light (λ ≥ 1000 nm) or red
light (λ ≥ 600 nm) filtered from a 100 W incandescent bulb,
new peaks appeared in the infrared absorption spectra, while
peaks due to ozone, perturbed ozone, carbon dioxide, and α-
pinene diminished. Most of the new peaks were consistent with
those found in a separate experiment in the infrared spectrum
of α-pinene oxide isolated in an argon matrix. The new
observed peaks were also consistent with peaks calculated for α-
pinene oxide and two isomeric ketones. A summary of the new
observed and calculated infrared peaks in these experiments is
shown in Table S1 (Supporting Information). A partial infrared
spectrum is shown in Figures 2 and S2 (Supporting
Information), and calculated structures are shown in Figure
3a−d.
Figure 3. Images of the calculated structures of the reactants and
products of the photochemical reactions of ozone (O3) with α-pinene
(C10H16): 3a = α-pinene (parent compound); 3b = α-pinene oxide; 3c
= ketone I; 3d = ketone II.
The first reaction in Scheme 2 is the primary photochemical
event, the red photolysis of matrix-isolated ozone. This reaction
has an overall energy of −22 kcal/mol (ozone bond cleavage, E
Scheme 2. Ozone + α-Pinene Red (λ ≥ 600 nm) Photolysis
Reaction Mechanism
Figure 2. Ozone plus α-pinene in an argon matrix; partial infrared
spectra, 400−800 cm−1: Ar/α-pinene/O3 = 400/1/1; blue trace =
ozone blank; green trace = α-pinene blank; brown trace = ozone + α-
pinene dark deposit, 14 K; black trace = ozone + α-pinene irradiated
deposit, with infrared light (λ ≥ 1000 nm) for 1/2 h and with red light
(λ ≥ 600 nm) for 1/2 h. Note the new peaks at 468, 507, 521, 622,
649, 734, and 771 cm−1.
C
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
= +25.4 kcal/mol; red photon energy, Ehν = −47.6 kcal/mol).
The O(3P) atoms produced by the red photolysis of ozone can
thus have excess thermal energy, allowing them to diffuse some
distance through the matrix to react with co-isolated α-pinene
molecules. The O(3P) atoms can react with the double bond in
α-pinene (3a) at either carbon atom and form one of four
possible triplet α-pinene oxide molecules. Molecules such as
these, which readily form upon reaction of O(3P) atoms with
alkenes, are believed to be stable 1,3-biradicals.11,12 Relative to
α-pinene and O(3P), the calculated energies of the triplet 1,3-
biradicals are −30 and −29 kcal/mol, respectively. We believe
that the triplet biradicals can cross over to the singlet manifold,
either through an intersystem crossing mechanism or perhaps
through spin−lattice interactions with the argon matrix. Once
in the singlet state, the molecules can readily rearrange to one
of the more stable epoxide or ketone structures. The calculated
energy of the α-pinene oxide (epoxide) (3b) structure relative
to α-pinene (3a) and O(3P) atoms is −86 kcal/mol, and those
of the ketone I (3c) and ketone II (3d) structures are −110 and
−111 kcal/mol, respectively. It would seem that the epoxide
(3b) structure is more likely to form in this reaction scheme
because it occurs without significant molecular rearrangement.
In contrast, the ketone I (3c) structure requires hydrogen atom
migration, and the ketone II (3d) structure requires methyl
radical migration. Nevertheless, the observation of an
absorption band at 1723 cm−1 with a possible shoulder at
1737 cm−1 and the absence of an aldehyde C−H stretch in the
2700−2900 cm−1 region are strong indications of the
occurrence of some ketone formation.
characteristic peak intensities in the ketones’ spectra. Finally,
the matrix photolysis experiments did not show any evidence of
the appearance of an aldehyde C−H stretch, which would have
appeared in the 2700−2900 cm−1 region, had any isomeric
aldehydes been produced.
B. Norbornene. When mixtures of argon with norbornene
(Ar/C7H10 = 200/1) and argon with ozone (Ar/O3 = 200/1)
were co-deposited on the CsI window chilled to 14 K, the solid
argon matrixes formed showed infrared spectral evidence for a
gas-phase thermal reaction between norbornene and ozone.
The apparent fast thermal reaction between ozone and
norbornene is consistent with the gas-phase bimolecular rate
constant of (1.55 0.05) × 10−15 cm3/molecule−s observed
by Greene and Atkinson at 296 K.15 The matrixes, formed in
the dark, had similar infrared spectra regardless of the
deposition geometry, MJ, CJ, or TJ. The peaks in all of these
experiments increased substantially when the matrixes were
annealed to 30 K, indicating that the thermal reaction between
norbornene and ozone had very low activation energy, on the
order of Ea ≈ 1−2 kcal/mol. Similar argon matrix annealing
experiments carried out in our laboratory16 showed that ozone
and cis-2-butene, which have a known activation barrier of 1.95
kcal/mol, reacted significantly under cryogenic annealing
conditions at 35 K. A summary of the new observed and
calculated infrared peaks in these experiments is shown in
Table 1. Partial infrared spectra are shown in Figures 4a,b and
S3 (Supporting Information), and calculated structures are
shown in Figure 5a−g.
It is worth pointing out that the epoxide isomer of α-pinene
oxide indicated in Scheme 2 (or its mirror image), with all three
pendent methyl groups on the same side of the six-membered
ring, is the only one observed in the products of the second
reaction. The other isomer, with the methyl group at position 2
on the opposite side of the six-membered ring as the two
methyl groups at position 6, was also calculated. Although this
second isomer of α-pinene oxide is only 4 kcal/mol less stable
than the isomer indicated in Scheme 2, its calculated infrared
absorptions did not match those from the experiment. The
apparent exclusion of the second isomer of α-pinene oxide from
the product mix is understandable in terms of a greater
accessibility of the double bond on the methylene rather than
the dimethyl side of the six-membered ring. In other words, an
ozone molecule or the newly photoproduced oxygen atom has
more room to approach the double bond and form the triplet
1,3-biradical on the methylene side of the six-membered ring
rather than the dimethyl side due to the steric obstruction
caused by the latter.
It is also worth noting that the assignment of the peaks at
1723 and 1737 cm−1 to the two isomeric ketones shown in
Figure 3c and d is based on these peaks alone and is therefore
somewhat tenuous. This is especially true because numerous
isomers of α-pinene oxide are observed in zeolite-catalyzed
isomerization reactions run at elevated temperatures.13
However, these ketones (I and II) (3c) and (3d) were
observed as the principal products, after α-pinene oxide, in
previous investigations of the gas-phase reactions of O(3P)
atoms with α-pinene.14 It is reasonable that these ketones are
also produced in the present matrix experiments. It is also
reasonable that if these ketones are minor products, the
carbonyl peaks are the only observable peaks for these
molecules because the carbonyl stretching intensity is calculated
to be nearly 2 orders of magnitude greater than any other
Figure 4. (a) Partial infrared spectra (800−1200 cm−1) of norbornene
(C7H10) and ozone (O3) in an argon matrix at 14 K: (Ar/O3 = 150,
Ar/C7H10 = 180); TJ deposition; blue trace = ozone blank; green trace
= norbornene blank; red trace = dark deposit (O3/C7H10 = 1.2); black
trace = anneal to 25 K. Note new peaks at 803, 823, 854, 857, 880,
893, 919, 951, 984, 989, 1015, and 1147 cm−1. (b) Partial infrared
spectra (1600−2000 cm−1) of norbornene (C7H10) and ozone (O3) in
an argon matrix at 14 K: (Ar/O3 = 150, Ar/C7H10 = 180); TJ
deposition; blue trace = ozone blank; green trace = norbornene blank;
red trace = dark deposit (O3/ C7H10 = 1.2); black trace = anneal to 25
K. Note the new peak at 1734 cm−1.
D
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
structures, the presence of interfering parent peaks due to
norbornene, and the mode mixing of the O−O stretching
vibration with ring vibrations all contribute to make definitive
assignments of these bands difficult. However, the mere
presence of substantial peaks with large, variable 18O isotope
shifts in this spectral region is characteristic of the carbonyl-O-
oxide moiety and not of any of the other likely structures
calculated here. Third, the observed spectra all show intense
peaks near 1000 cm−1 (at 984, 989, and 1015 cm−1) with
substantial 18O shifts (of −14, −16, and −6 cm−1) that are
characteristic of the C−O stretch in either the norbornene
primary or SOZ. Both of these peak assignments are indicated
in the table. This dual assignment does not necessitate the
presence of the norbornene SOZ in these experiments,
especially because the more intense, calculated SOZ peak
near 925 cm−1 is not observed in any of the experimental
spectra. However, the peaks near 1300 cm−1 (at 1289, 1304,
1311, 1317, and 1328 cm−1) with small, observed 18O isotope
shifts (of −2, 0, 0, 0, and 1 cm−1, respectively) are readily
assigned to the norbornene SOZ and none of the other
calculated structures, although they are not characteristic.
Further, the inference of the presence of the norbornene acid−
aldehyde molecule (discussed below) in the matrixes almost
necessitates the presence of the norbornene SOZ as a probable
intermediate. Fourth, the appearances of an intense band in the
carbonyl region at 1734 cm−1 with a shoulder at 1728 cm−1 and
a large 18O isotope shift of −34 cm−1 and two peaks in the low
C−H stretching region at 2729 and 2820 cm−1 are strong
indicators of an aldehyde group, which, of course, has already
been implicated by the assignment of the CI. The presence of
two aldehyde C−H stretching bands separated by nearly 100
cm−1 is interesting. The calculations show that the aldehyde
C−H stretch is quite sensitive to the local environment of the
aldehyde group. In particular, the interaction of the aldehyde
group with the carbonyl-O-oxide on the same molecule can
blue shift the calculated aldehyde C−H stretch by 75 cm−1
when the carbonyl-O-oxide oxygen is near the aldehyde
hydrogen (2.65 Å in the interacting rotamer versus 6.67 Å in
the noninteracting rotamer). Interestingly, a similar effect is
seen in the calculated norbornene acid−aldehyde structure
when the acid hydrogen is in proximity to the aldehyde oxygen.
Finally, the presence of the norbornene acid−aldehyde
molecule is inferred by the observation of a broad peak at
1387 cm−1 with a 6 cm−1 18O isotope shift and the coincidence
of broad peaks at 1734, 2729, and 2820 cm−1 with the CI.
These broad peaks are all enhanced upon UV (λ ≥ 200 nm)
irradiation, while the sharper norbornene Criegee aldehyde C−
H peaks are diminished.
Figure 5. Calculated structures for the reaction of ozone (O3) with
norbornene (C7H10): Gaussian 09 suite of programs, applying DFT
using the B3LYP functional and the 6-311 G++ (d,2p) basis set (5a =
norbornene (parent compound); 5b = norbornene POZ; 5c =
norbornene CI-I; 5d = norbornene CI-II; 5e = norbornene SOZ; 5f =
norbornene acid−aldehyde I; 5g = norbornene acid−aldehyde II).
The observed infrared peaks listed in Table 1 include a
number of characteristic features. First, there are three peaks in
the 600−700 cm−1 region (at 646, 680, and 682 cm−1) with
large 18O isotope shifts that are characteristic of POZs.7 These
peaks were assigned to three of the four possible spatial isomers
of the norbornene POZ. These isomers include two
conformations each of two different stereoisomers determined
by whether the ozone approached the double bond of
norbornene from the methylene side or from the ethylene
side. All of these isomers have relatively intense absorptions in
this region due to the antisymmetric stretch of the O−O bonds
in the five-membered POZ ring. Second, the experimental
spectra all have bands in the 800−900 cm−1 region whose
isotope shifts are substantial but difficult to definitively assign.
These peaks were assigned to the O−O stretch of the carbonyl-
O-oxide of various isomers of the CI of norbornene ozonide. At
least 18 stable, spatial isomers of the norbornene CI were
calculated, nine rotamers each of two geometric (E and Z)
isomers of the carbonyl-O-oxide. The abundance of calculated
A possible reaction scenario for the thermal reaction between
ozone and norbornene is illustrated in Scheme 3. This reaction
scenario, consistent with the experimental observations,
indicates that the ozonolysis of norbornene proceeds via the
expected, traditional Criegee mechanism. As alluded to earlier,
various spatial isomers of the POZ, the CI, the SOZ, and the
acid−aldehyde were observed in the present experiments. The
calculated potential energy diagram for the formation of the
POZ from ozone and norbornene is shown in Figure 6a, and
that for the formation of the POZ from ozone and α-pinene is
shown in Figure 6b. Significantly, the calculations indicate that
both the norbornene and α-pinene POZs should form from
their parent molecules without a noticeable activation barrier. It
should be noted that some test points were calculated using a
counterpoise correction between the ozone and α-pinene
E
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Table 1. Infrared Absorption Peaks Observed in an Argon Matrix at 14 K upon TJ Dark Deposition of Mixtures of Argon with
a
Norbornene (Ar/C7H10 ≈ 200) and Argon with Ozone (Ar/O3 ≈ 200)
experimental IR absorptions
calculated IR absorptions
(Gaussian 09 DFT-B3LYP)
experimental IR absorptions
calculated IR absorptions
(Gaussian 09 DFT-B3LYP)
(Ar/C7H10/O3 ≈ 400/1/1)
assignment
(Ar/C7H10/O3 ≈ 400/1/1)
assignment
molecule
freq (int)
18O shift
freq (int)
18O shift
freq (int)
18O shift
freq (int)
18O shift
[cm−1 (au)]
[cm−1
]
[cm−1(km/mol)]
[cm−1
]
molecule
[cm−1 (au)]
[cm−1
]
[cm−1(km/mol)]
[cm−1
]
b
501 (0.14)
619 (0.10)
646 (0.14)
680 (0.25)
682 (0.28)
721 (0.15)
734 (0.10)
781 (0.07)
798 (0.32)
803 (0.08)
823 (0.10)
854 (0.16)
857 (0.16)
857 (0.16)
880 (0.42)
893 (0.07)
919 (0.11)
951 (0.26)
984 (0.60)
989 (0.80)
−5
−11
−31
−34
−33
0
507 (4)
−4
POz-a
1461 (0.31)
1479 (0.16)
1734 (1.52)
−3
−1
−34
b
c
641 (23)
672 (32)
706 (26)
−32
−36
−37
POz-b
1798 (171)
1798 (275)
1803 (324)
2138
−37
−37
−33
−52
0
CI-a
b
e
POz-l
acid-a
b
e
POz-a
acid-a
h
2138 (0.18)
2729 (0.38)
−51
−8
CO
c
0
2840 (91)
2868 (119)
2915 (36)
2922 (100)
CI-I
b
e
−4
−6
−3
−14
−2
−14
−48
−2
796 (3)
−5
−7
−4
−9
−3
−13
−46
−2
−5
−2
−2
−15
−14
−12
−3
−6
−1
0
POz-a
0
acid-a
b
c
806 (5)
POz-a
2820 (0.40)
−6
0
CI-II
c
e
822 (19)
822 (8)
CI-I
0
acid-l
c
,
,
f
f
a
CI-l
CI-l
CI-a
CI-II
The observed peaks were all significantly enhanced upon annealing
the matrix to 30 K. Calculated peaks with assignments are listed for
likely products of the reaction between norbornene and ozone. POz-
a, POz-b, and POz-l are different conformers/isomers of the
norbornene POZ that have different orientations of the 1,2,3-
trioxolane ring relative to the norbornene moiety. The molecular
c
851 (11)
878 (20)
874 (76)
909 (24)
902 (35)
927 (14)
971 (31)
987 (48)
997 (41)
1016 (45)
1018 (28)
1024 (76)
1151 (5)
1266 (12)
1268 (3)
1306 (12)
1310 (1)
1314 (1)
1334 (12)
1347 (40)
b
g
g
c,f
c
f
,f
c
,
CI-l
c
−7
0
CI-b
c
c
structure of POz-a is shown in Figure 5b. CI-I, CI-II, CI-a, CI-b, and
CI-a
CI-a
CI-l are different rotamers of the norbornene CI that have different
rotational orientations of the aldehyde and carbonyl oxide groups
relative to the norbornene moiety. The molecular structures of CI-I
c
−2
−14
−16
b
d
POz-a
b
d
POz-l
SOz-n
and CI-II are shown in Figure 5c and d, respectively. SOz-d and SOz-
n are different isomers of the norbornene SOZ that have different
orientations of the 1,2,4-trioxolane ring relative to the norbornene
moiety. The molecular structure of SOz-d is shown in Figure 5e.
b
1015 (0.20)
−6
POz-l
SOz-dc
e
b
1147 (0.33)
1261 (0.11)
1289 (0.09)
1304 (0.32)
1311 (0.22)
1317 (0.22)
1328 (0.28)
1387 (0.18)
1444 (0.25)
−3
0
POz-l
Acid-a and acid-l are different rotamers of the norbornene acid−
c
aldehyde structure that have different rotational orientations of the
acid group and the aldehyde group relative to the norbornene moiety.
The molecular structures of acid-a and acid-l are shown in Figure 5f
CI-a
d
−2
0
−3
−1
0
SOz-d
SOz-d
SOz-d
SOz-d
SOz-d
d
d
d
d
f
and g, respectively. Possible mode mixing observed in the calculated
0
18O isotopomers make assignments in the 800−900 cm−1 region
0
0
g
difficult (see the text). It is unclear whether the peak at 857 cm−1
1
−3
−4
shifts to 843 cm−1 or to 809 cm−1 upon 18O substitution. Both are
e
−6
0
acid-a
h
possible and assignable to CI rotamers. References 17 and18.
fragments at 200 and 220 pm in order to test for any basis set
superposition errors (BSSEs). These errors were found to be
+2.7 and +2.3 kcal/mol, respectively, at 200 and 220 pm
separation and thus, we believe, insignificant relative to the
energy profile itself.
These observations may be understood in terms of the initial
UV photolysis of ozone molecules to produce O(1D) atoms
with sufficient excess thermal energy to allow them to diffuse
through the matrix and react with nearby molecules. The
excited oxygen atoms can insert into a C−C or C−H bond of
the initial products of ozonolysis to produce an unstable
intermediate. The intermediate product of this insertion
reaction can then rearrange to a more stable form, possibly
ejecting CO or CO2 in the process. It is also possible that the
initial products of the ozonolysis of norbornene are themselves
photolabile in UV light, but our results can neither substantiate
nor refute this alternate reaction. The proposed photochemical
mechanism is illustrated in Scheme 4, and the observed and
calculated infrared absorption peaks are listed in Table S2
(Supporting Information).
The initial spin-allowed ozone photolysis reaction producing
O(1D) and O2(1Δ) is energetically possible at wavelengths less
than 310 nm.19 Because we irradiated the matrix with the
quartz/water-filtered light from a medium-pressure, short arc
mercury lamp, the radiation included UV light with wave-
lengths above 200 nm. The O(1D) atoms thus produced could
have as much as 50 kcal/mol of excess thermal energy, enough
to allow some limited diffusion of these atoms through the solid
When argon matrixes containing ozone, norbornene, and
their reaction intermediates and products were irradiated for 1
h with UV light from a medium-pressure mercury lamp filtered
through quartz and 10 cm of water (λ ≥ 200 nm), all of the
infrared peaks in the spectrum diminished except those
assigned to the norbornene acid−aldehyde at 1387, 1734,
2729, and 2820 cm−1, which increased slightly. In addition, a
number of new peaks appeared. These new peaks are listed in
Table S2 (Supporting Information) with proposed assignments
from calculated structures or from literature references. The
most notable features in the spectra of these UV-irradiated
matrixes include new bands assignable to carbonyl (CO)
stretches in the 1700−1800 cm−1 region, new bands assignable
to hydroxyl (O−H) stretches in the 3500−3700 cm−1 region,
and significant increases in bands assignable to carbon
monoxide (CO) and carbon dioxide (CO2). Partial infrared
difference spectra showing bands in the CO and O−H
regions are illustrated in Figure S4 (Supporting Information).
F
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Scheme 3. Norbornene Ozonolysis: Criegee Mechanism
argon matrix.20 As the O(1D) atoms encounter organic species
in the matrix, they can react by insertion reactions,21 producing
a variety of organic acids, aldehydes, and alcohols, as well as CO
and CO2.
Figure 6. Calculated potential energy diagrams of the formation of
POZs. (a) Norbornene POZ (5b) separated into norbornene (5a) and
ozone along a linear path in 20 pm steps by redundant coordinate
optimization. (b) α-Pinene POZ (3b) separated into α-pinene (3a)
and ozone along a linear path in 10 pm steps by redundant coordinate
optimization.
C. Norbornadiene. When mixtures of argon with
norbornadiene (Ar/C7H8 = 440/1) and argon with ozone
(Ar/O3 = 100/1) were co-deposited on the CsI window chilled
to 14 K, the solid argon matrixes formed showed infrared
spectral evidence for a gas-phase thermal reaction between
norbornadiene and ozone. The apparent fast thermal reaction
between ozone and norbornadiene is consistent with the gas-
phase bimolecular rate constant of (3.55 0.07) × 10−15 cm3/
molecule−s observed by Greene and Atkinson at 296 K.15 The
new infrared peaks were observed when the matrixes were
formed in the dark under TJ deposition conditions, indicating
that a fast thermal reaction was taking place between the
reactants. Matrixes formed under MJ conditions showed
significantly different infrared spectra, indicating the formation
of later reaction intermediates and/or stable products given the
longer reaction times. Because our intention was to isolate early
reaction intermediates in the ozone norbornadiene reactant ion,
we focused our attention on the TJ experimental results.
The infrared peaks that formed upon deposition in the TJ
experiments increased substantially when the matrixes were
annealed to 25 K, indicating that the thermal reaction between
norbornadiene and ozone, like that between norbornene and
ozone, had very low activation energy, on the order of Ea ≤ 1−
2 kcal/mol. A summary of the new observed and calculated
infrared peaks in these experiments is shown in Table 2. Partial
infrared spectra are shown in Figure 7a,b, and likely calculated
structures are shown in Figure 8a−e.
Scheme 4. UV Photolysis of Ozone and Norbornene in an
Argon Matrix
G
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Table 2. Infrared Absorption Peaks Observed in an Argon Matrix at 14 K upon TJ Dark Deposition of Mixtures of Argon with
a
Norbornadiene (Ar/C7H8 ≈ 440) and Argon with Ozone (Ar/O3 ≈ 100)
experimental IR
absorptions
experimental IR
absorptions
(Ar/C7H8 ≈ 440/1)
calculated IR absorptions
(Gaussian 09 DFT-B3LYP)
(Ar/C7H8 ≈ 440/1)
calculated IR absorptions
(Gaussian 09 DFT-B3LYP)
(Ar/O3 ≈ 100/1)
assignment
molecule
(Ar/O3 ≈ 100/1)
assignment
molecule
freq (int)
18O shift
freq (int)
18O shift
freq (int)
18O shift
freq (int)
18O shift
[cm−1 (au)]
[cm−1
]
[cm−1(km/mol)]
[cm−1
]
[cm−1 (au)]
[cm−1
]
[cm−1(km/mol)]
[cm−1
]
406 (0.12)
613 (0.10)
619 (0.10)
628 (0.10)
682 (0.30)
713 (0.06)
719 (0.56)
752 (0.32)
775 (0.10)
784 (0.24)
809 (0.24)
857 (0.28)
859 (0.56)
889 (0.16)
899 (0.28)
956 (0.22)
970 (0.10)
982 (0.50)
1003 (0.19)
1003 (0.19)
1105 (0.19)
1181 (0.12)
1239 (0.16)
1258 (0.05)
1282 (0.06)
1289 (0.05)
1303 (0.18)
1335 (0.24)
1387 (0.14)
−4
−5
−6
−7
−28
−11
−2
−1
−10
−3
−1
−28
−28
−6
0
1442 (0.24)
1463 (0.08)
1732 (2.00)
1738 (2.50)
1771 (0.25)
1785 (0.25)
2138 (0.10)
2728 (0.24)
−1
c
−37
−32
−32
−38
1793 (160)
1801 (137)
1811 (417)
1812 (341)
−37
−38
−34
−34
CI-Z-0-180
c
CI-Z-0-0
b
e
680 (18)
715 (43)
728 (56)
755 (43)
775 (48)
785 (21)
809 (23)
854 (61)
857 (61)
879 (50)
−26
−9
−1
−2
−6
−2
−3
−34
−17
−7
POz-a
acid-u
e
e
acid-t
acid-v
b
f
f
POz-e
POz-g
2138
CO
b
c
−8
−8
2861 (121)
2863 (114)
2894 (118)
2932 (107)
0
0
0
0
CI-Z-0-0
b
e
POz-c
acid-v
c
c
CI-Z-0-0
2829 (0.30)
CI-Z-0-180
c
c
e
CI-Z-0-180
CI-Z-0-180
acid-u
a
The observed peaks were all significantly enhanced upon annealing
the matrix to 30 K. Calculated peaks with assignments are listed for
likely products of the reaction between norbornadiene and ozone.
POz-a, POz-c, POz-e, and POz-g are different conformers/isomers of
the norbornadiene POZ that have different orientations of the 1,2,3-
trioxolane ring relative to the norbornadiene moiety. The molecular
structure of POz-a is shown in Figure 8b. CI-Z-0-0, CI-Z-0-180, and
CI-Z-180-180 are different rotamers of the norbornadiene CI that have
different rotational orientations of the aldehyde and carbonyl oxide
groups relative to the norbornadiene moiety. The molecular structure
of CI-Z-0-0 is shown in Figure 8c. SOz-m and SOz-n are different
isomers of the norbornadiene SOZ that have different orientations of
the 1,2,4-trioxolane ring relative to the norbornadiene moiety. The
molecular structure of SOz-m is shown in Figure 8d. Acid-t, acid-u,
and acid-v are different rotamers of the norbornadiene acid−aldehyde
structure that have different rotational orientations of the acid group
and the aldehyde group relative to the norbornadiene moiety. The
molecular structure of acid-u is shown in Figure 8e. References 17
and18.
c
CI-Z-0-0
c
CI-Z-0-0
b
c
0
963 (34)
−1
CI-Z-180-180
c
b
−13
−7
−7
980 (46)
1000 (73)
1006 (42)
−14
−8
−3
POz-a
d
SOz-n
b
POz-e
d
d
−2
−1
0
1180 (5)
1252 (5)
1263 (5)
1248 (15)
1292 (22)
−2
−1
−1
−3
−1
SOz-n
c
CI-Z-180-180
b
e
POz-a
e
−3
0
acid-u
d
SOz-m
−12
f
e
−6
1377 (60)
−3
acid-t
The observed infrared peaks listed in Table 2 include a
number of characteristic features. First, there are several peaks
in the 600−800 cm−1 region (at 682, 719, 752, and 775 cm−1)
with variable 18O isotope shifts that are characteristic of POZs.
These peaks were assigned to the four possible spatial isomers
of the norbornadiene POZ. As with norbornene, these isomers
include two conformations, each of two different stereoisomers
determined by whether the ozone approached the double bond
of norbornadiene from the methylene side or from the ethyne
side. All of these isomers have relatively intense absorptions in
this region due to the antisymmetric stretch of the O−O−O
group in the five-membered POZ ring. Second, the
experimental spectra have a doublet band at 857 and 859
cm−1 with a substantial 18O isotope shift of −28 cm−1. These
peaks were assigned to the O−O stretch of the carbonyl-O-
oxide of the CI of norbornadiene ozonide. As with norbornene,
at least 18 stable, spatial isomers of the norbornadiene CI were
calculated, nine rotamers each of two geometric (E and Z)
isomers of the carbonyl-O-oxide. The large number of
calculated structures and the mode mixing of the O−O
stretching vibration with ring vibrations contribute to make
definitive assignments of these bands difficult. However, unlike
the norbornene case, there are few interfering peaks from the
parent norbornadiene molecule in this spectral region, and the
18O isotope shifts are more easily measured. In particular,
because all three oxygen positions in the norbornadiene CI are
unique, there are eight possible 16,18O mixed isotopomers for
each spatial isomer. Figure 9a shows the partial infrared
spectrum for the pure and scrambled isotope experiments and
compares them to those calculated for statistical mixes of the
eight isotopomers of C7H8O3-jb (C7H8O3-Z-0-180), shown in
Figure 9b, or for statistical mixes of the eight isotopomers of
C7H8O3-i (C7H8O3-Z-0−0), shown in Figure 9c. The agree-
ment between the experimental and calculated spectra provides
compelling evidence for the assignment of the bands at 857 and
859 cm−1 to the carbonyl-O-oxide of the norbornadiene CI.
Third, the observed spectra all show intense peaks near 1000
cm−1 (at 982 and 1003 cm−1) with substantial 18O shifts of −13
and −7 cm−1, respectively that are characteristic of the C−O
stretch in either the norbornadiene primary or SOZ. Both of
these peak assignments are indicated in the table. Fourth, the
appearance of an intense band in the carbonyl region at 1732
and 1738 cm−1 with large 18O isotope shifts of −37 and −32
cm−1, respectively, and two peaks in the low C−H stretching
region at 2728 and 2829 cm−1 are strong indicators of the
aldehyde group of the CI. The presence of two aldehyde C−H
stretching bands separated by nearly 100 cm−1 is similar to the
results observed for norbornene. As with norbornene, the
calculations of norbornadiene CIs show that the aldehyde C−H
stretch is quite sensitive to the local environment of the
H
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
Figure 7. (a) Partial infrared spectra (400−800 cm−1) of
norbornadiene (C7H8) and ozone (O3) in an argon matrix at 14 K:
(Ar/O3 = 100, Ar/C7H8 = 440); TJ deposition; orange trace =
norbornadiene blank; green trace = dark deposit (O3/C7H8 = 3.3);
black trace = anneal to 25 K; purple trace = anneal to 30 K. Note new
peaks at 406, 519, 613, 619, 628, 682, 713, 719, 752, 775, and 784
cm−1. (b) Partial infrared spectra (800−1200 cm−1) of norbornadiene
(C7H8) and ozone (O3) in an argon matrix at 14 K: (Ar/O3 = 100, Ar/
C7H8 = 440); TJ deposition; orange trace = norbornadiene blank;
green trace = dark deposit (O3/C7H8 = 3.3); black trace = anneal to 25
K; purple trace = anneal to 30 K. Note new peaks at 809, 857, 859,
889, 899, 970, 982, 1003, 1105, and 1181 cm−1.
Figure 8. Calculated structures for the reaction of ozone (O3) with
norbornadiene (C7H8): Gaussian 09 suite of programs, applying DFT
using the B3LYP functional and the 6-311 G++ (d,2p) basis set (8a =
norbornadiene (parent compound); 8b = norbornadiene POZ; 8c =
norbornadiene CI; 8d = norbornadiene SOZ; 8e = norbornadiene
acid−aldehyde).
three of these reactions are near zero, it appears that steric
effects on the entropy of activation are much more important
than ring strain effects. A small activation energy of 1.4 kcal/
mol was measured for the reaction of α-pinene with ozone.22
The additional methyl groups in α-pinene must significantly
limit the number of successful collisions of ozone with the
double bond relative to the corresponding number of successful
collisions in norbornene and in norbornadiene.
aldehyde group. In rotamers of the norbornadiene CI and of
the norbornadiene acid−aldehyde, the aldehyde C−H stretch-
ing frequency is significantly affected by the proximity of the
carbonyl-O-oxide or the carboxyl group to the aldehyde group
on the same molecule. Finally, the presence of the
norbornadiene acid−aldehyde molecule is inferred by the
observation of a broad peak at 1387 cm−1 with a 6 cm−1 18
O
SUMMARY AND CONCLUSIONS
isotope shift and the coincidence of broad peaks at 1771, 2728,
■
and 2829 cm−1 with those of the CI.
(1) The thermal reaction between α-pinene and ozone was too
slow to observe under the TJ or MJ deposition conditions of
these experiments. Because the activation energy for this
reaction is near zero, the slower rate of the α-pinene reaction
with ozone as compared to that of norbornene or
norbornadiene is attributed primarily to the steric effects of
the three additional methyl groups in α-pinene.
D. Structural Effects on Reaction Rates. Although all
three of these bicyclic compounds (α-pinene, norbornene, and
norbornadiene) are known to react with ozone in the gas
phase,5 we did not see evidence for the α-pinene−ozone gas-
phase reaction under any of the deposition or annealing
conditions of these experiments. This failure to see the α-
pinene ozonolysis reaction can be understood if one compares
the gas-phase rate constants for these three reactions,5 shown in
Table 3.
Table 3 clearly shows that norbornene and norbornadiene
react about 20 and 40 times faster with ozone at 296 K than α-
pinene does. This rate difference can account for the inability to
observe the α-pinene gas-phase reaction with ozone in the
present work. It has been observed over the years in our
laboratory that a bimolecular rate constant, k ≥ 1 × 10−16 cm3/
molecule−s, is needed to observe gas-phase reactions in TJ
matrix experiments. The reason for these rate differences in the
reaction of ozone with these bicyclic compounds has been
ascribed to a combination of “steric and ring strain effects”.15,25
Because our calculations show that the activation energies of all
(2) Irradiation of argon matrixes containing α-pinene and
ozone (Ar/α-pinene/ozone = 400/1/1) with red (λ ≥ 600 nm)
or with infrared (λ ≥ 1000 nm) light resulted in observable
photochemical changes in the infrared spectra of these matrixes.
(3) New infrared peaks in α-pinene/ozone photolysis
experiments could be assigned to α-pinene oxide and one or
two isomeric ketones that could have resulted from initial
ozone photolysis followed by subsequent reaction of O(3P)
oxygen atoms with matrix-isolated α-pinene molecules.
(4) Thermal, gas-phase reactions of ozone with norbornene
and with norbornadiene were observed in both TJ and MJ
deposition experiments. Reaction was evidenced by the
appearance of new peaks in the infrared spectra of the mixtures
of ozone with norbornene in argon (Ar/norbornene/ozone =
I
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
deposits could be assigned to the early intermediates of
ozonolysis. Proposed products include POZs, CIs, and SOZs,
as well as acid−aldehydes of norbornene or norbornadiene,
respectively.
(6) Peak assignments were made by comparing the
experimental spectra and 18O isotope shifts with those
calculated for intermediate and product molecules using DFT
with the B3LYP functional and the 6-311G++(d,2p) basis set.
(7) In all of these bicyclic systems, no evidence for unusual
reactions prompted by the bicyclic nature of the organic
reactants was observed.
ASSOCIATED CONTENT
* Supporting Information
■
S
Full citation for ref 8, additional infrared spectra from twin-jet
deposition and red photolysis experiments involving ozone and
α-pinene, norbornene, and norbornadiene, and the energy
profile for the reaction of ozone with norbornadiene. Also,
tables of spectral data for the photochemical reactions of ozone
with α-pinene and norbornene are included. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: bruce.ault@uc.edu. Phone: 513-556-9238.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Anna Gudmundsdottir for helpful
discussions, Saint Mary’s University of Minnesota for sabbatical
funding for one of us (R.W.K.) during the 2011−2012
academic year, and the National Science Foundation for
research funding under Grant NSF CHE-0749109.
REFERENCES
■
(1) Criegee, R.; Wenner, G. Die Ozonisierung des 9,10-Oktalins.
Justus Liebigs Ann. Chem. 1949, 564, 9−15.
̈
(2) Criegee, R. Uber den Verlauf der Ozonspaltung (III. Mitteilung).
Justus Liebigs Ann. Chem. 1953, 583, 1−36.
Figure 9. (a) Experimental spectra from three isotope experiments:
blue trace, C7H8 + 16O3; red trace, C7H8 + 18O3; green trace, C7H8 +
16,18O3; all dark deposited for 20+ h and annealed to 25 K. (b)
Calculated spectra of a statistical mixture of eight isotopomers of CI-Z-
0-180; all Gaussian peaks are with the fwhm set to 7 cm−1. (c)
Calculated spectra of a statistical mixture of eight isotopomers of CI-Z-
0-0; all Gaussian peaks are with the fwhm set to 7 cm−1.
(3) Criegee, R. Mechanismus der Ozonolyse. Angew. Chem. 1975, 87,
765−771.
(4) Bailey, P. S. Ozonation in Organic Chemistry: Vol. I Olefinic
Compounds; Organic Chemistry, A Series of Monographs; Academic
Press: New York, 1978; 39-I.
(5) Calvert, J. G.; Atkinson, R.; Kerr, J. A.; Madronich, S.; Moortgat,
G. K.; Wallington, T. J.; Yarwood, G. The Mechanisms of the
Atmospheric Oxidation of the Alkenes; Oxford University Press: New
York, 2000; 172−335.
(6) Sander, W. Carbonyl Oxides: Zwitterions or Diradicals? Angew.
Chem., Int. Ed. Engl. 1990, 29, 344−354.
(7) Hoops, M. D.; Ault, B. S. Matrix-Isolation Study of the Early
Intermediates in the Ozonolysis of Cyclopentene and Cyclo-
pentadiene: Observation of Two Criegee Intermediates. J. Am.
Chem. Soc. 2009, 131, 2853−2863.
(8) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; et al. Gaussian 09, revision C.01; Gaussian, Inc.:
Wallingford, CT, 2009.
(9) Atkinson, R. Gas-Phase Tropospheric Chemistry of Volatile
Organic Compounds: 1. Alkanes and Alkenes. J. Phys. Chem. Ref. Data
1997, 26, 215−290.
(10) Atkinson, R.; Carter, W. P. L. Kinetics and Mechanisms of the
Gas-Phase Reactions of Ozone with Organic Compounds under
Atmospheric Conditions. Chem. Rev. 1984, 84, 437−470.
Table 3. Second-Order Gas-Phase Rate Ozonolysis Rate
Constants of Bicyclic Compounds Measured at 296 1 K
bicyclic compound
k (10−16 cm3/molecule−s)
a,b,c,d
α-pinene
0.85
a,e
norbornene
15.5
35.5
a,
e
norbornadiene
a
b
c
d
Reference 5. Reference 22. Reference 23. Reference 24.
e
Reference 15.
400/1/1) and of ozone with norbornadiene in argon (Ar/
norbornadiene/ozone = 400/1/1) upon deposition. Peaks
formed upon initial deposition at 14 K all increased upon
annealing to 25 K.
(5) The new peaks in the infrared spectra of argon/
norbornene/ozone or argon/norbornadiene/ozone matrix
J
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A
Article
(11) Cvetanovic, R. J. Mechanism of the Interaction of Oxygen
Atoms with Olefins. J. Chem. Phys. 1956, 25, 376−377.
(12) Havel, J. J.; Chan, K. H. Atomic Oxygen IV. Rearrangements in
the Reactions of Oxygen (3P) Atoms with Cyclic and Bicyclic
Hydrocarbons. J. Am. Chem. Soc. 1975, 97, 5800−5804.
(13) Holderich, W. F.; Roseler, J.; Heitmann, G.; Liebens, A. T. The
̈
̈
Use of Zeolites in the Synthesis of Fine and Intermediate Chemicals.
Catal. Today 1997, 37, 353−366.
(14) Alvarado, A.; Tuazon, E. C.; Aschmann, S. M.; Atkinson, R.
Products of the Gas-Phase Reactions of O(3P) Atoms and O3 with α-
Pinene and with 1,2-Dimethyl-1-cyclohexene. J. Geophys. Res. 1998,
103, 25541−25551.
(15) Greene, C. R.; Atkinson, R. Rate Constants for the Gas-Phase
Reactions of Ozone with a Series of Cyclic Alkenes and α,β-
Unsaturated Ketones at 296 2K. Int. J. Chem. Kinet. 1994, 26, 37−
44.
(16) Clay, M.; Ault, B. S. Infrared Matrix Isolation and Theoretical
Study of the Initial Intermediates in the Reaction of Ozone with cis-2-
Butene. J. Phys. Chem. A 2010, 114, 2799−2805.
(17) Abe, H.; Yamada, K. M. T. Spectroscopic Identification of the
CO-H2O 2−1 Cluster Trapped in an Argon Matrix. J. Chem. Phys.
2004, 121, 7803−7810.
(18) Nakamoto, K. Infrared Spectra of Inorganic and Coordination
Compounds, 2nd ed.; Wiley Interscience: New York, 1970; 78.
(19) Calvert, J. G.; Pitts, J. N. Photochemistry; John Wiley & Sons:
New York, 1966; p 209.
(20) Chen, M.; Wang, X.; Zhang, L.; Qin, Q.; Zheng, Q.
Photomobility of O(1D) Atom in Solid Argon and its Reaction with
CF3I. Chem. Phys. 2000, 255, 95−102.
(21) Parnis, J. M.; Hoover, L. E.; Fridgen, T. D.; Lafleur, R. D.
Stabilization of the Primary Products of Atomic Oxygen (1D)
Reactions with Carbon Monoxide, Carbon Dioxide, and Methane,
and Other Hydrocarbons in Cryogenic Matrices. J. Phys. Chem. 1993,
97, 10708−10711.
(22) Atkinson, R.; Winer, A. M.; Pitts, J. N., Jr. Rate Constants for
the Reactions of Ozone with the Natural Hydrocarbons of Isoprene
and α- and β-Pinene. Atmos. Environ. 1982, 16, 1017−1020.
(23) Nolting, F.; Behnke, W.; Zetzsch, C. A Smog Chamber for
Studies of the Reactions of Terpenes and Alkanes with Ozone and
Hydroxyl. J. Atmos. Chem. 1988, 6, 47−59.
(24) Atkinson, R.; Hasegawa, D.; Aschmann, S. M. Rate Constants
for the Gas-Phase Reactions of Ozone with a Series of Monoterpenes
and Related Compounds at 296 2 K. Int. J. Chem. Kin. 1990, 22,
871−887.
(25) Atkinson, R.; Aschmann, S. M.; Carter, W. P. L.; Pitts, J. N., Jr.
Effects of Ring Strain on Gas-Phase Rate Constants. I. Ozone
Reactions with Cyclic Alkenes. Int. J. Chem. Kin 1983, 15, 721−731.
K
DOI: 10.1021/jp510883k
J. Phys. Chem. A XXXX, XXX, XXX−XXX