Direct measurement of Criegee intermediate (CH2OO) reactions with acetone, acetaldehyde, and hexafluoroacetone

Article abstract of DOI:10.1039/c2cp40294g

Criegee biradicals, i.e., carbonyl oxides, are critical intermediates in ozonolysis and have been implicated in autoignition chemistry and other hydrocarbon oxidation systems, but until recently the direct measurement of their gas-phase kinetics has not been feasible. Indirect determinations of Criegee intermediate kinetics often rely on the introduction of a scavenger molecule into an ozonolysis system and analysis of the effects of the scavenger on yields of products associated with Criegee intermediate reactions. Carbonyl species, in particular hexafluoroacetone (CF₃COCF₃), have often been used as scavengers. In this work, the reactions of the simplest Criegee intermediate, CH₂OO (formaldehyde oxide), with three carbonyl species have been measured by laser photolysis/tunable synchrotron photoionization mass spectrometry. Diiodomethane photolysis produces CH ₂I radicals, which react with O₂ to yield CH₂OO + I. The formaldehyde oxide is reacted with a large excess of a carbonyl reactant and both the disappearance of CH₂OO and the formation of reaction products are monitored. The rate coefficient for CH₂OO + hexafluoroacetone is k₁ = (3.0 ± 0.3) × 10^-11 cm³ molecule^-1 s^-1, supporting the use of hexafluoroacetone as a Criegee-intermediate scavenger. The reactions with acetaldehyde, k₂ = (9.5 ± 0.7) × 10^-13 cm³ molecule^-1 s^-1, and with acetone, k ₃ = (2.3 ± 0.3) × 10^-13 cm³ molecule^-1 s^-1, are substantially slower. Secondary ozonides and products of ozonide isomerization are observed from the reactions of CH₂OO with acetone and hexafluoroacetone. Their photoionization spectra are interpreted with the aid of quantum-chemical and Franck-Condon-factor calculations. No secondary ozonide was observable in the reaction of CH₂OO with acetaldehyde, but acetic acid was identified as a product under the conditions used (4 Torr and 293 K).

Full text of DOI:10.1039/c2cp40294g

View Online / Journal Homepage / Table of Contents for this issue
Physical Chemistry Chemical Physics
www.rsc.org/pccp
Volume 14 | Number 30 | 14 August 2012 | Pages 10357–10816
Includes a collection of articles on new insights into organic chemistry from forefront physical measurements
ISSN 1463-9076
COVER ARTICLE
Taatjes et al.
Direct measurement of Criegee intermediate (CH2OO) reactions with acetone,
acetaldehyde, and hexafluoroacetone
1463-9076(2012)14:30;1-T

Dynamic Article Links
PCCP
View Online
Cite this: Phys. Chem. Chem. Phys., 2012, 14, 10391–10400
www.rsc.org/pccp
PAPER
Direct measurement of Criegee intermediate (CH2OO) reactions with
acetone, acetaldehyde, and hexafluoroacetone
Craig A. Taatjes,*a Oliver Welz,a Arkke J. Eskola,a John D. Savee,a
David L. Osborn,a Edmond P. F. Lee,bc John M. Dyke,b Daniel W. K. Mok,c
Dudley E. Shallcrossd and Carl J. Percival*e
Received 31st January 2012, Accepted 9th March 2012
DOI: 10.1039/c2cp40294g
Criegee biradicals, i.e., carbonyl oxides, are critical intermediates in ozonolysis and have been
implicated in autoignition chemistry and other hydrocarbon oxidation systems, but until recently
the direct measurement of their gas-phase kinetics has not been feasible. Indirect determinations
of Criegee intermediate kinetics often rely on the introduction of a scavenger molecule into an
ozonolysis system and analysis of the effects of the scavenger on yields of products associated
with Criegee intermediate reactions. Carbonyl species, in particular hexafluoroacetone
(CF3COCF3), have often been used as scavengers. In this work, the reactions of the simplest
Criegee intermediate, CH2OO (formaldehyde oxide), with three carbonyl species have been
measured by laser photolysis/tunable synchrotron photoionization mass spectrometry.
Diiodomethane photolysis produces CH2I radicals, which react with O2 to yield CH2OO + I.
The formaldehyde oxide is reacted with a large excess of a carbonyl reactant and both the
disappearance of CH2OO and the formation of reaction products are monitored. The rate
coefficient for CH2OO + hexafluoroacetone is k1 = (3.0 ꢀ 0.3) ꢁ 10ꢂ11 cm3 moleculeꢂ1 sꢂ1
,
supporting the use of hexafluoroacetone as a Criegee-intermediate scavenger. The reactions
with acetaldehyde, k2 = (9.5 ꢀ 0.7) ꢁ 10ꢂ13 cm3 moleculeꢂ1
s
ꢂ1, and with acetone,
k3 = (2.3 ꢀ 0.3) ꢁ 10ꢂ13 cm3 moleculeꢂ1
s
ꢂ1, are substantially slower. Secondary ozonides
and products of ozonide isomerization are observed from the reactions of CH2OO with acetone
and hexafluoroacetone. Their photoionization spectra are interpreted with the aid of quantum-
chemical and Franck–Condon-factor calculations. No secondary ozonide was observable in the
reaction of CH2OO with acetaldehyde, but acetic acid was identified as a product under the
conditions used (4 Torr and 293 K).
reactivity of Criegee intermediates in gas-phase ozonolysis or
Introduction
in the troposphere have relied on indirect determinations.5,6
One particularly useful approach has been to introduce
‘‘Criegee scavenger’’ molecules into ozone/hydrocarbon reac-
tions and investigate the effect on products thought to origi-
nate from Criegee intermediates. One type of scavenger is a
carbonyl compound; as in the liquid-phase Criegee ozonolysis
mechanism,7 carbonyl oxide intermediates undergo cycloaddition
with carbonyl compounds to form secondary ozonides. These
secondary ozonide products can be detected, for example by
infrared spectroscopy8,9 or gas chromatography10 and have
been confirmed as products from ozonolysis in gas-phase11 as
well as in solution.7 Hexafluoroacetone (HFA) is a prominent
representative of the carbonyl class of Criegee scavengers;
Horie et al.8 deduced that HFA reacts rapidly with Criegee
intermediates based on Fourier-transform infrared (FTIR)
measurements of product formation from ozonolysis of small
alkenes in the presence of added HFA. Horie et al. measured a
product from adding HFA to the ethene ozonolysis system
Carbonyl oxides, known as ‘‘Criegee intermediates’’ after
Rudolf Criegee, who proposed their participation in ozonolysis,1
are important species in tropospheric chemistry. Most carbonyl
oxides in the troposphere are produced by ozonolysis, but
other tropospheric reactions can also produce Criegee inter-
mediates.2,3 However, until recently2,4 no Criegee intermediate
had been observed in the gas phase, and information about the
a Combustion Research Facility, Sandia National Laboratories,
7011 East Ave., MS 9055, Livermore, California 94551, USA.
E-mail: cataatj@sandia.gov
b School of Chemistry, University of Southampton, Highfield,
Southampton SO17 1BJ, UK
c Department of Applied Biology and Chemical Technology,
The Hong Kong Polytechnic University, Hung Hom, Hong Kong
d School of Chemistry, University of Bristol, Bristol BS8 1TS, UK
e School of Earth, Atmospheric and Environmental Sciences,
The University of Manchester, Williamson Building, Oxford Road,
Manchester M13 9PL, UK. E-mail: carl.percival@manchester.ac.uk
c
This journal is the Owner Societies 2012
Phys. Chem. Chem. Phys., 2012, 14, 10391–10400 10391

View Online
and assigned it to the secondary ozonide, 3,3-di(trifluoro-
methyl)-1,2,4-trioxolane, resulting from CH2OO addition to
HFA, as shown in Scheme 1.
radiation and are mass-analyzed in an orthogonal-acceleration
time-of-flight mass spectrometer. Pulsed extraction of the
continuous ion beam occurs at a 50 kHz repetition rate, so a
full mass spectrum is obtained every 20 microseconds. Some
experiments requiring analysis of larger-mass species were carried
out with a 40 kHz repetition rate and a corresponding 25 ms
kinetic time step. Because the ionization photon energy can be
readily tuned, the kinetic data is generally acquired as a three-
dimensional dataset of ion signal as a function of mass, kinetic
time, and photon energy.12 The measured signals are back-
ground-corrected by subtracting the signal levels before the firing
of the photolysis laser, and are normalized to the ionizing photon
flux (measured with a photodiode). The data taken at a fixed
photon energy or integrated over photon energy gives a time-
resolved mass spectrum of the reaction, as shown in Fig. 1.
Integration over a range of kinetic time yields photoionization
spectra of observed mass channels. Isomeric species can be
identified based on their photoionization spectra.4,12,14 The mass
axis was calibrated using mass peaks from a flow of a gas mixture
containg H2, ethene, propene, 1-butene, Kr, and Xe in an Ar
buffer, and the photon energy was calibrated using autoionizing
resonances in Xe. Additional kinetics measurements were carried
out at the Combustion Research Facility, using a hydrogen
discharge lamp (generating principally Lyman-a radiation at
10.20 eV) as a photoionization source.
Scheme 1
The assignment was made by analogy to determinations by
Neeb et al.9 of secondary ozonide formation in other gas-phase
ozonolysis systems. Horie et al.8 described this assignment as
tentative, but the analogy with other systems is compelling. Also,
they carried out a relative-rate measurement to deduce that HFA
reacts B13 times faster than acetaldehyde with CH2OO.
Recently it has become feasible to directly detect at least the
simplest carbonyl oxide, CH2OO, by tunable synchrotron
photoionization mass spectrometry,4 and it has been shown
that the reaction of CH2I with O2 is a convenient means for
preparing CH2OO in the laboratory and measuring its
kinetics.2 Additionally, photoionization mass spectrometry with
a synchrotron light source is a powerful tool for analyzing
isomeric products of chemical reactions.12 As a result we can
explicitly confirm the reactivity of carbonyl compounds with
CH2OO in the gas phase and investigate reaction products.
Here we present absolute measurements of the rate coefficients
for CH2OO reacting with HFA, acetaldehyde, and acetone:
Formaldehyde oxide, CH2OO, is produced from the reaction
of CH2I with O2.2 Diiodomethane (99%) was introduced into
the cell ([CH2I2] = 5.7 ꢁ 1012 molecule cmꢂ3) by a He flow
through a thermostatted bubbler and photolyzed at 248 nm or
351 nm in the presence of a large excess (B1016 molecule cmꢂ3
)
of O2. In some experiments 13CH2I2 (99 atom% 13C) was
employed to produce 13CH2OO. The absorption cross-sections
of acetaldehyde and acetone are significantly higher than that
CH2OO + CF3COCF3 - products
CH2OO + CH3CHO - products
CH2OO + CH3COCH3 - products
(1)
(2)
(3)
Secondary ozonides (trioxolanes) are observed as products in
the reactions with acetone and HFA, and are identified by
comparison of measured photoionization spectra with results
of quantum chemical calculations of the adiabatic ionization
energies (AIE) and Franck–Condon factors (FCF) for ionization.
Evidence for isomerization of the secondary ozonide is also
observed. In contrast with the 730 Torr measurements of
Horie et al.,8 no secondary ozonide is detected in the reaction
of CH2OO with acetaldehyde under the low-pressure condi-
tions of the present experiments, but formation of acetic acid is
identified as a product channel.
Methods
Fig. 1 Section of time-resolved mass spectra for the reaction of
13CH2OO with HFA, integrated over photon energies from
10.5 eV–11.3 eV. Note the break in the x-axis scale between m/z =
51 and m/z = 125. The 13CH2OO is observed at m/z = 47, and the
prominent peak at m/z = 127 is the I atom produced in the CH2I2
photolysis. The secondary ozonide appears at m/z = 213 and
dissociative ionization of (trifluoromethoxy)methyl trifluoroacetate
appears at m/z = 144. Other peaks (HI at m/z = 128, IO at m/z =
143, ICl at m/z = 162 and 164, IBr at m/z = 206 and 208) arise from
side reactions of the I atom, including possibly with the halocarbon
wax coating.
Synchrotron photoionization mass spectrometry
The kinetic measurements were carried out in the Sandia
Multiplexed Chemical Kinetics Reactor,13 operating at the
Chemical Dynamics Beamline (9.0.2) of the Advanced Light
Source (ALS) at Lawrence Berkeley National Laboratory.
Reactants flow slowly through a quartz photolysis reactor
coated with halocarbon wax. The contents of the reactor are
continuously sampled through a small orifice in the sidewall.
The emerging molecules are ionized by tunable synchrotron
c
10392 Phys. Chem. Chem. Phys., 2012, 14, 10391–10400
This journal is the Owner Societies 2012

View Online
of HFA at 248 nm, so 351 nm photolysis of CH2I2 was employed
for the CH2OO + CH3CHO and CH2OO + CH3COCH3
reaction measurements in order to avoid photolysis of CH3CHO
and CH3COCH3 reactants. Known dilutions of the carbonyl
reagent in He were prepared in 4-liter vessels and the mixture
was flowed through the reactor at 293 K. Flows of the reagents
and He diluent are controlled with calibrated mass flow
controllers. The total pressure is maintained at 4 Torr by
active control of the pumping speed, and the total flow is
sufficient to replenish the reactor between laser pulses (repeti-
tion rates of either 4 Hz or 10 Hz were used). The reaction of
CH2I with O2 is rapid15 and CH2OO formation is observed
almost immediately following the photolysis laser pulse
(k(CH2I + O2)ꢃ[O2] E 14 000 sꢂ1). The CH2OO can be
distinguished from other isomers such as formic acid because
of its different ionization energy.4 The decay of CH2OO in the
absence of added reagent is attributed to heterogeneous reaction
in the present experiments, as the decay constant is dependent
on the coating and history of the reactor tube. The ‘‘zero-reagent’’
decay was monitored at several points during every set of
kinetic measurements to control for possible changes in the
heterogeneous removal rate during the course of an experi-
ment. The slowest first-order decay rate coefficient observed is
B75 sꢂ1, considerably slower than previous observations,2
placing an experimental lower limit of B13 ms on the lifetime
of CH2OO at room temperature (293 K). However, the
removal of CH2OO in the present experiments may still be
dominated by heterogeneous or bimolecular reactions even at
the lowest observed decay rates, so this lower limit may be far
below the true unimolecular lifetime. Quantum chemical and
Rice-Ramsperger-Kassel-Marcus (RRKM) calculations by
Olzmann et al.16 predict a high-pressure limit of the CH2OO
lifetime that is more than a factor of 100 larger than the
present lower limit.
Fig. 2 Measured decay of 13CH2OO in the presence of varying
concentrations of HFA. The signals are fit to exponential decays,
convolved with the instrument response function.
Adding carbonyl compounds to the reaction mixture accel-
erated the decay of CH2OO, as shown in Fig. 2 for the reaction
of CH2OO with HFA. The decay of CH2OO was fit to an
exponential, corrected for the instrument response function
where appropriate (i.e. for the fastest decays).2 The inverse
time constant of the decay (ꢄ the pseudo-first order rate
coefficient), plotted as a function of added reagent concentration,
yields a straight line, as shown in Fig. 3 for the reaction of
CH2OO with HFA. The slope of this line gives the second-order
rate coefficient for the reaction. In addition, photoionization
spectra were acquired for all three reactions at a high concen-
tration of the carbonyl reactant. Products are identified by
comparison of the obtained photoionization spectra with measured
spectra of authentic samples or by comparison with calculated
photoionization spectra (see below).
Fig. 3 Plot of the pseudo-first order rate coefficient (decay constant)
of CH2OO removal as a function of HFA concentration. Two inde-
pendent datasets are shown; error bars are 95% precision of the fit to
the pseudo-first order decay. The intercept reflects other loss processes
for CH2OO, for example, heterogenous reaction on the walls.
calculations employing MOLPRO19 with various basis
sets20–22 for 3,3-dimethyl-1,2,4-trioxolane and its cation. The
purpose of carrying out RCCSD(T) and UCCSD(T)-F12x
calculations is to obtain theoretical benchmarks for the adiabatic
ionization energy (AIE) of 3,3-dimethyl-1,2,4-trioxolane. With
the RCCSD(T) results, extrapolation to the complete basis set
(CBS) limit was carried out using the 1/X3 formula.23,24 The
RCCSD(T)/CBS approach using correlation-consistent basis
sets has been considered as the ‘‘gold standard’’ of quantum
chemistry.25–27 With the recently reported, explicitly corre-
lated, UCCSD(T)-F12x (x = a or b) methods, which correct
Quantum chemistry and Franck–Condon calculations
Geometry optimization and harmonic vibrational frequency
calculations were performed at the B3LYP/6-311G** level
using Gaussian09.17 Higher-level single-energy calculations were
carried out for more reliable relative electronic energies
at computed B3LYP/6-311G** geometries. They include
CBS-QB318 calculations using Gaussian0917 for all species con-
sidered, and RCCSD(T) and explicitly correlated UCCSD(T)-F12x
c
This journal is the Owner Societies 2012
Phys. Chem. Chem. Phys., 2012, 14, 10391–10400 10393

View Online
the lack of derivative discontinuity (cusp) in standard wave
functions,28,29 the scaled perturbative triples obtained by a
HFA to the appearance of propene ozonide (methyl-1,2,4-
trioxolane, a product of the CH2OO + CH3CHO reaction)
when HFA and acetaldehyde were added to an ethene ozono-
lysis reaction. In their experiments the ethene ozonolysis
produced CH2OO biradical, which then reacted with both
HFA and acetaldehyde. Horie et al.8 deduced k1/k2 = 13,
approximately half the ratio of the rate constants in the
present experiment. It is possible that this difference (which
is well outside the stated uncertainties for the relative rate
measurement) simply reflects different falloff behavior of the
two reactions; the measurements of Horie et al.8 were carried
out at 730 Torr of synthetic air, and the present measurements
at 4 Torr in helium. We do not observe formation of methyl-
1,2,4-trioxolane in the reaction of CH2OO with acetaldehyde
(see below), further suggesting that the stabilization channel
observed by Horie et al.8 may be pressure dependent.
MP2
simple scaling factor, DE(Tsc) = DE(T) ꢁ EcorrMP2–F12/Ecorr
(i.e. the ratio between the computed correlation energies
obtained at the MP2 and MP2-F12 levels) have been
employed.30 It has been shown that explicitly correlated methods,
such as UCCSD(T)-F12, achieve a dramatic improvement of
basis set convergence of correlation energies when compared
with conventional correlation methods, such as RCCSD(T).31
In order to assist assignments of the observed photoioniza-
tion mass spectra, Franck–Condon factor (FCF) calculations
were carried out for the secondary ozonide products using the
ezSpectrum code.32 All vibrational modes were considered in
the FCF calculations, which include Duschinsky rotation33
but are within the harmonic oscillator model. Because of the
size of the molecular systems considered and the large number
of combination bands excited upon ionization (see below),
anharmonicity, which is expected to be significant particularly
for vibrational modes involving CH stretches, has to be
ignored. This approximation may result in an overestimate
of the intensity of the simulated vibrational structure in the
higher ionization energy (IE) region. In addition, the maximum
The Criegee intermediates have a polar biradical electronic
configuration and a singlet ground state.35–38 The reactions of
carbonyl oxides with aldehydes and ketones are typically
described as 1,3-dipolar cycloadditions,7,39 and the ordering
of the rate coefficients, k1 > k2 > k3, is in accordance with the
expected relative susceptibility of the carbonyl species to
dipolar attack. For example, in liquid phase ozonolysis the
formation of ozonides from carbonyl oxide addition to ketones
is less efficient than for addition to aldehydes,7,39,40 but the
ketone reactivity can be increased by electron-withdrawing
substituents.41 The substitution of the methyl groups in acetone
by more electron withdrawing groups (H, CF3) will tend to
stabilize the transition state for attack of the electron-rich O end
of the CH2OO molecule to the carbonyl carbon.
number of vibrational excitations in the cationic state (the
P
maximum value of vi, where vi is the quantum number of
vibrational mode i; i.e. the maximum total number of excitations
in all normal modes combined for a certain vibrational
component in the cationic state) has to be restricted to 7
(i.e., 7 layers), and ‘‘hot’’ bands have been ignored, because of
limitations on the total number of FCFs that ezSpectrum can
manage. Nevertheless, comparison between the overall vibra-
tional structures obtained using 6 and 7 layers in the FCF
calculations suggests that using 7 layers should be adequate.
As an example to show the size of the FCF calculations, for
3,3-di(trifluoromethyl)-1,2,4-trioxolane, with 7 layers, 53 524 680
vibrational states in the cationic state were considered, and the
FCFs of 2 140 987 200 combination bands were computed.
The present experiments consider only the reaction of CH2OO.
For singly alkyl-subsituted Criegee biradicals, (at least) two
conformers exist: syn- (O–O pointing towards the alkyl group)
Results and discussion
The rate coefficients for reaction of CH2OO with the three
carbonyl compounds considered here are derived from the
data in Fig. 3–5:
k1 (CH2OO + CF3COCF3)
= (3.0 ꢀ 0.3) ꢁ 10ꢂ11 cm3 moleculeꢂ1 sꢂ1
,
,
.
k2 (CH2OO + CH3CHO)
= (9.4 ꢀ 0.7) ꢁ 10ꢂ13 cm3 moleculeꢂ1 sꢂ1
k3 (CH2OO + CH3COCH3)
= (2.3 ꢀ 0.3) ꢁ 10ꢂ13 cm3 moleculeꢂ1 sꢂ1
The uncertainty limits are 95% precision of the fit. The reaction
with hexafluoroacetone is markedly faster than the reaction
with acetone or with acetaldehyde, corroborating earlier
observations8 and tending to confirm the utility of HFA as a
scavenger of Criegee intermediates.6,34 Horie et al.8 measured
the rate constant of the reaction of CH2OO with HFA relative
to that with acetaldehyde by comparing the consumption of
Fig. 4 Plot of the pseudo-first order rate coefficient (decay constant)
of CH2OO removal as a function of acetaldehyde concentration. Error
bars are 95% precision of the fit to the pseudo-first order decay. The
intercept reflects other loss processes for CH2OO, for example,
heterogenous reaction on the walls.
c
10394 Phys. Chem. Chem. Phys., 2012, 14, 10391–10400
This journal is the Owner Societies 2012

View Online
Fig. 6 for the reaction of 13CH2OO with HFA. The use of
13CH2OO moves the peak for the ozonide minus CF3 away
from the IO signal at m/z = 143. (Signal at m/z = 144 from
HOI formation is negligible compared to the observed product
signal.) The time behavior at the ozonide mass is identical to
that of m/z = 144. Similar behavior is observed for the
reaction of CH2OO with acetone, where the main product
signals are at the mass of the ozonide and the ozonide minus
CH3. Although these observations at first sight suggest that the
signals corresponding to CH3/CF3 loss arise from dissociative
ionization of the ozonides, certain aspects of the photoionization
spectra, discussed later, argue against such an assignment.
No previous measurements of the photoionization spectra
of secondary ozonides exist. Martinez et al.43 observed signals
at the mass of the secondary ozonides in photoionization mass
spectrometry measurements of the ozonolysis of alkenes, but
did not have the capability to measure photoionization spectra
and could not definitively determine the isomeric nature of the
observed molecules. In order to identify the signals at the
ozonide masses we compare the measured photoionization
spectra to calculated spectra based on quantum chemistry and
Franck–Condon simulations. Such calculated spectra assume
that the ionization is dominated by direct continuum ionization
to the parent cation.
Fig. 5 Plot of the pseudo-first order rate coefficient (decay constant)
of CH2OO removal as a function of acetone concentration. Error bars
are 95% precision of the fit to the pseudo-first order decay. The
intercept reflects other loss processes for CH2OO, for example,
heterogenous reaction on the walls.
Table 1 reports the computed electronic energy difference,
DEion, for the ionization of 3,3-dimethyl-1,2,4-trioxolane, the
secondary ozonide expected from reaction of CH2OO with
acetone, obtained at different levels of calculation. When
corrected for zero-point energies, this energy difference gives
the AIE, which is defined as the energy from the ground
and anti- (O–O pointing away from the alkyl group. Steric or
electronic considerations in substituted Criegee intermediates
may affect the reactivity of the radicals and produce different
reactivity for the two conformers.6,38 Murray et al.40 noted
that aldehydic carbonyl oxides such as CH3CHOO reacted
more readily with acetone than did their ketonic counterparts
such as (CH3)2COO. One might imagine that measurements of
CH2OO kinetics may be representative of anti-conformers
because there is no methyl group for the O–O group to point
towards. By comparison, Fenske et al.42 deduced an absolute
rate coefficient of 1 ꢁ 10ꢂ12 cm3 moleculeꢂ1 sꢂ1 (with a factor
of 6 uncertainty) for the reaction of CH3CHOO with
acetaldehyde, a similar value to the present measurement of
k2. The unimolecular decomposition of the syn-CH3CHOO,
through vinyl hydroperoxide, should be more rapid than
decomposition of the anti-conformer.5,6 If as a result the
observed secondary ozonide in the experiments of Fenske
et al.42 principally arose from reaction of the anti-CH3CHOO,
the rough congruence of the present measurement with that of
Fenske et al.42 might indeed suggest that the CH2OO rate
coefficients with carbonyl compounds are similar to those of
the anti- conformer of larger Criegee intermediates. Explicit
measurement of reactions of syn- and anti-carbonyl oxides is
clearly desirable.
The present experiments detect the appearance of product
species as well as the disappearance of reactants, as shown in
Fig. 1. The major product of both liquid- and gas-phase
reactions of Criegee biradicals with carbonyl compounds is
generally thought to be the secondary ozonides.7,11 The dominant
product signals in the reaction of 13CH2OO with HFA are at
the mass of the ozonide (m/z = 213) and that of the ozonide
minus CF3 (m/z = 144). The time traces of the rising product
signals and the decaying Criegee intermediate are shown in
Fig. 6 Time behavior of the 13CH2OO and the product signals for the
reaction with HFA. The slight long-time increase in products, and the
slow secondary decay of 13CH2OO on the same timescale, is the result
of small contributions of secondary reactions with CH2I2 that regenerate
CH2I and hence CH2OO.
c
This journal is the Owner Societies 2012
Phys. Chem. Chem. Phys., 2012, 14, 10391–10400 10395

View Online
Table 1 Computed difference in electronic energy (eV) between the ground neutral and cationic states (DEion) of 3,3-dimethyl-1,2,4-trioxolane at
different levels of calculations
Methods
DEion
Nbasis
RMP2/cc-pVDZ
RCCSD/cc-pVDZ
RCCSD(T)/cc-pVDZ
RMP2/cc-pVTZ
RCCSD/cc-pVTZ
RCCSD(T)/cc-pVTZ
RMP2/aug-cc-pVDZ
RCCSD/aug-cc-pVDZ
9.0274
8.8960
8.8232
9.3455
9.1611
9.1169
9.2981
9.1112
9.0747
9.4530
9.2465
9.3017
9.5251
9.2708
9.2450
9.2337
9.2078
9.2727
9.2285
9.3035
9.2670
9.2406
9.245 ꢀ 0.022
9.21 ꢀ 0.03
138
138
138
322
322
322
233
233
RCCSD(T)/aug-cc-pVDZ
RMP2/aug-cc-pVTZ
RCCSD/aug-cc-pVTZ
233
506
506
282, 638, 688a
RMP2/cc-pVDZ-F12
RMP2-F12/cc-pVDZ-F12
UCCSD-F12a/cc-pVDZ-F12
UCCSD-F12b/cc-pVDZ-F12
UCCSD(T)-F12a/cc-pVDZ-F12
UCCSD(T)-F12b/cc-pVDZ-F12
RCCSD/CBS(1/X3:VTZ;VDZ)
RCCSD/CBS(1/X3:VTZ,VDZ) + D(T)/VTZb
RCCSD/CBS(1/X3:AVTZ;AVDZ)
RCCSD/CBS(1/X3:AVTZ;AVDZ) + D(T)/AVDZb
RCCSD(T)/CBS(1/X3:VTZ;VDZ)
c
Best estimated DEion
Best estimated AIEd
a
b
AO, RI and DF basis sets: VDZ-F12, cc-pVDZ-F12/OPTRI and aug-cc-pVDZ/MPFIT. D(T) is the correction due to contributions of triple
excitations obtained at the basis set given. The best estimate is obtained as the average of the RCCSD/CBS(1/X3:VTZ,VDZ) + D(T)/VTZ,
c
RCCSD/CBS(1/X3:AVTZ;AVDZ) + D(T)/AVDZ and RCCSD(T)/CBS(1/X3:VTZ;VDZ) values. The estimated uncertainty is the difference
between the best estimate and the RCCSD/CBS(1/X3:AVTZ;AVDZ) + D(T)/AVDZ value. Zero-point energy corrections from computed
B3LYP/6-311G** vibrational frequencies.
d
vibrational and electronic state of the neutral to the ground
vibrational and electronic state of the cation. From the results
culminating in RCCSD(T) calculations—RMP2, RCCSD and
RCCSD(T)—with different basis sets shown in Table 1, it is
clear that basis-size effects increase the computed DEion values,
while electron-correlation effects decrease the computed DEion
values. Similar conclusions can be drawn from the results
obtained from calculations leading to UCCSD(T)-F12x/
cc-pVDZ-F12: RMP2, RMP2-F12, UCCSD-F12x and
UCCSD(T)-F12x with the cc-pVDZ-F12 basis set as shown
in Table 1. The best theoretical estimate of the DEion value
obtained based on various RCCSD(T)/CBS values is 9.245 ꢀ
0.022 eV (see footnotes b and c of Table 1). It is pleasing that
this best estimate agrees very well with the explicitly correlated
UCCSD(T)-F12a value of 9.234 eV and the UCCSD(T)-F12b
value of 9.208 eV. The consistency of these high-level results
confirms their reliability. Including corrections for zero-point
energies (see footnote d of Table 1), the best theoretical AIE
value is 9.21 ꢀ 0.03 eV, whereas CBS-QB3 predicts a value of
9.29 eV. Two conformers exist for the ground state cation of
3,3-dimethyl-1,2,4-trioxolane. They are mirror images of each
other and as such have the same energy. However the computed
FCFs to these two cationic conformers are very different,
because the geometry changes to one conformer are consider-
ably larger than those to the other, giving considerably smaller
computed overlap integrals. In this connection, only the
conformer with the more favorable FCF has been included
in spectral simulations (a similar situation applies to the
3,3-di(trifluoromethyl)-1,2,4-trioxolane).
Fig. 7 and 8 show the photoionization spectra of the
observed products of the reactions of CH2OO with acetone
and with HFA, compared with the calculated photoionization
spectra of the secondary ozonides. The calculated spectrum for
the 3,3-dimethyl-1,2,4-trioxolane product arising from the
reaction of CH2OO with acetone (Fig. 7) uses the best-estimate
AIE and is scaled vertically to match the amplitude of the
experimental signal. The calculated 3,3-di(trifluoromethyl)-
1,2,4-trioxolane spectrum in Fig. 8 employs a lower-accuracy
CBS-QB3 estimate of the adiabatic ionization energy and is
additionally shifted to higher energy by 55 meV (5.3 kJ molꢂ1).
Such a discrepancy is not unreasonable for the calculation of a
relatively heavy fluorinated molecule by CBS-QB3. The good
agreement of the calculated and experimental spectra identifies
the carriers as the expected secondary ozonide products. Calcu-
lated AIEs (Table 2) for other possible isomers (Scheme 2) are in
poor agreement with the observed photoionization spectra.
Scheme 2 C4H8O3 isomers.
Fig. 7 also shows the calculated (CBS-QB3) asymptotic
product energy for formation of CH3 and C3H5O3+ by simple
c
10396 Phys. Chem. Chem. Phys., 2012, 14, 10391–10400
This journal is the Owner Societies 2012

View Online
Fig. 7 Photoionization spectra of the m/z = 104 product from the
reaction of CH2OO with acetone, compared with the calculated
spectrum of the secondary ozonide. The simulated spectrum uses the
best high-level theoretical estimate of the AIE and Franck–Condon
calculations, as described in the text. The AIE for the methoxymethyl
acetate isomer and the calculated energetic threshold for dissociative
ionization of the secondary ozonide isomer (‘‘DI threshold for ozo-
nide’’) are also shown for reference. The m/z = 89 signal is attributed
to dissociative ionization of methoxymethyl acetate.
Fig. 8 Photoionization spectra of the m/z = 213 product from the
reaction of 13CH2OO with HFA, compared with a calculated spectrum
of the secondary ozonide. The simulated spectrum uses a calculated
Franck–Condon envelope and is shifted up by 55 meV from the
CBS-QB3 estimate of the AIE. The signal at m/z = 144 is tentatively
assigned to dissociative ionization of the (trifluoromethyl)methoxy
trifluoroacetate isomer.
C–C bond fission after photoexcitation of the neutral 3,3-
dimethyl-1,2,4-trioxolane. The apparent experimental threshold
of the methyl loss channel, B9.9 eV, is B20 kJ molꢂ1 below
the CBS-QB3 asymptotic energy. Such a discrepancy is
unusually large for CBS-QB3 calculations; for comparison,
the CBS-QB3 value of 9.29 eV for the adiabatic ionization
energy of 3,3-dimethyl-1,2,4-trioxolane (Table 2) is only
80 meV (7.7 kJ molꢂ1) higher than the best estimate of
(9.21 ꢀ 0.03) eV. It is possible that a lower-energy pathway
to a more stable C3H5O3+ cation in the dissociative ionization
explains the energetic discrepancy. However, the unexpectedly
large intensity of the methyl-loss peak compared with the
Table 2 Adiabatic ionization energies (AIE) for the secondary ozo-
nides formed from the reaction of CH2OO with acetaldehyde, acetone,
and HFA. AIEs and relative reaction enthalpies DH0,rel at 0 K for
potential products from isomerization of 3,3-dimethyl-1,2,4-trioxolane
and 3,3-di(trifluoromethyl)-1,2,4-trioxolane are also shown. All
energies are obtained from CBS-QB3 calculations
DH0,rel
Species
AIE (eV) (kJ molꢂ1
)
3-Methyl-1,2,4-trioxolane
9.44
—
3,3-Dimethyl-1,2,4-trioxolane
Methoxymethyl acetate
(Dimethyl)hydroxymethyl formate
9.29a
9.79
0
ꢂ235
+
parent C4H8O3 ion suggests that the methyl loss channel is
10.20
ꢂ307
not from dissociation of the 3,3-dimethyl-1,2,4-trioxolane
cation. Nevertheless, the kinetic time traces of the daughter
peaks (Fig. 6) are those of a stable product, not of a reactive
radical, so the methyl- or CF3- loss must occur after ionization.
Some alternatives may be ionization to a dissociative excited
state of the cation, or autoionization arising from excitation to
an excited electronic state of neutral 3,3-dimethyl-1,2,4-trioxo-
lane that then autoionizes to the daughter ion and CH3. In this
case the unusually large intensity may reflect a near-resonance
excitation to the excited state of interest.
3,3-Di(trifluoromethyl)-1,2,4-trioxolane
(Trifluoromethoxy)methyl trifluoroacetate
Di(trifluoromethyl)hydroxymethyl formate 11.28
10.59
—
0
b
ꢂ297
ꢂ304
a
Best-estimate value from higher-level calculations is 9.21 ꢀ 0.03 eV
b
(Table 1). The cation ground state is calculated to be unbound.
ionization pathway exists (Scheme 3) where energies (in kJ molꢂ1
)
are at 0 K relative to the methoxymethyl acetate cation (from
CBS-QB3 and G3MP2B345 calculations). The energies given
below the reaction arrows in Scheme 3 are for the transition
states. Both CBS-QB3 and G3MP2B3 predict that the transition
state of the C–C bond fission is the highest-energy configuration
in the pathway shown in Scheme 3 leading to methyl loss.
The energy of this transition state relative to the cation
A more likely alternative is ionization from a different neutral
C4H8O3 isomer, for which dissociative ionization dominates.
The methoxymethyl acetate isomer (calculated AIE 9.79 eV)
can be formed by a 1,4-methyl shift following O–O bond
scission of the secondary ozonide.10,44 A low-lying dissociative
c
This journal is the Owner Societies 2012
Phys. Chem. Chem. Phys., 2012, 14, 10391–10400 10397

View Online
Scheme 3
of methoxymethyl acetate differs between the two methods
(CBS-QB3: 20 kJ molꢂ1; G3MP2B3: 2 kJ molꢂ1). The calcula-
tions on this transition state show substantial spin conta-
mination (hS2i B0.90–0.96 in the different single-point
calculations), which will result in an increased uncertainty in
the calculated energies. Nevertheless, based on the AIE of
methoxymethyl acetate obtained from CBS-QB3, the appear-
ance energy for the methyl loss channel is 9.81 eV (G3MP2B3)
and 10.00 eV (CBS-QB3). Both values are in reasonable agree-
ment with the observed threshold for m/z = 89 in Fig. 7.
Because the dissociation channel has only a small barrier,
significant ion signal at the parent mass of methoxymethyl
acetate is not expected, in accordance with the experiment.
Based on these observations we attribute the m/z = 89 signal
to dissociative ionization of the methoxymethyl acetate
species. The ionization energy of the other isomerization
product of the secondary ozonide, hydroxyisopropyl formate,10
is calculated as 10.20 eV. No significant onset in the photo-
ionization spectrum at m/z = 104 around this energy is
observed, and if methyl loss were a significant dissociative
ionization channel in its cation, the appearance energy of
Fig. 9 Photoionization spectrum of the mass-60 product of the
CH2OO reaction with acetaldehyde, compared to a calibration spectrum
of acetic acid.
The secondary ozonides and their isomerization products are not
the only products observed. In the reaction of CH2OO with HFA a
contribution from formic acid was observed, although it was not
quantified. (Studies did not go to high enough photon energy to
investigate formic acid production in the reaction of CH2OO with
acetone.) Moreover, in the reaction of 12CH2OO with acetalde-
hyde, no product signal is observed at either the ozonide mass
(m/z = 90) or the ozonide minus CH3 at m/z = 75. The product
channels observed are formaldehyde at m/z = 30, and a second
product at m/z = 60. Evidence of formic acid is visible in the
photoionization spectrum of the reaction of 12CH2OO with acet-
aldehyde, but the large background from doubly-13C substituted
acetaldehyde nearly obscures it. The photoionization spectrum
of the mass-60 product is in reasonable agreement with the
photoionization spectrum of acetic acid (Fig. 9).
+
C3H5O3 would be at least 10.20 eV, too high to explain the
experimental onset.
Similar arguments apply to the origin of the m/z = 144
product signal from the reaction of 13CH2OO with HFA. The
energy of the products from simple C–C bond fission in the
cation of the secondary ozonide, CF3 and C3H2F3O3+, relative
to the neutral secondary ozonide is 11.47 eV, which is B0.5 eV
higher than the experimental onset in the m/z = 144 product
spectrum (see Fig. 8). Di(trifluoromethyl)hydroxymethyl
formate has an AIE of 11.28 eV. Although one of the C–C
bonds in the cationic structure is already substantially elongated,
CF3 loss would still have an appearance energy of at least
the value of the AIE. No secondary threshold that could
correspond to this isomerization product is observed in the
m/z = 144 and 213 photoionization spectra. The cationic state
of methoxymethyl acetate analogue in the HFA reaction,
(trifluoromethoxy)methyl trifluoroacetate, is calculated to be
unbound and undergoes dissociative ionization. It seems likely that
CF3 loss occurs in this cation, and we assign the m/z = 144 product
signal from the reaction of 13CH2OO with HFA to the dissocia-
tive ionization of (trifluoromethoxy)methyl trifluoroacetate.
The use of photoelectron spectroscopy to study small
Criegee intermediates and the secondary ozonides produced
in their reaction with aldehydes and ketones would be extre-
mely valuable in resolving the vibrational structure in their
first photoelectron bands to establish ionic state vibrational
separations and to test the Franck–Condon simulations, as
used, for example, in obtaining the simulated photoionization
spectra shown in Fig. 7 and 8.
If the pathway for reaction of CH2OO with carbonyl
compounds is cycloaddition, as the body of ozonolysis experi-
ments indicates, decomposition and isomerization of the
secondary ozonide (Scheme 4) might contribute to the product
spectrum. The thermal decomposition of the secondary
ozonide product is very slow; using the reported Arrhenius
parameters for unimolecular decomposition of methyl-1,2,4-
trioxolane (propene ozonide)46 yields a rate coefficient of only
B4 ꢁ 10ꢂ8 sꢂ1 at 295 K. However, if isomerization and
dissociation channels exist with barriers near or below the
CH2OO + carbonyl entrance barrier, the chemically activated
secondary ozonides may isomerize and/or dissociate before
collisional stabilization can occur under the low-pressure
conditions of the present experiments.
Scheme 4
c
10398 Phys. Chem. Chem. Phys., 2012, 14, 10391–10400
This journal is the Owner Societies 2012

View Online
Isomerization and decomposition of excited secondary
ozonide products has been implied in previous experiments.
For example, the yield of secondary ozonides in gas-phase
1-alkene ozonolysis experiments was found to decrease sub-
stantially with increasing temperature,10 which was rationalized
as a decomposition of excited initial secondary ozonide species
from reactions of the CH2OO and aldehyde products of the
initial ozonolysis step. The CO2-laser excitation of 3,5-dimethyl-
1,2,4-trioxolane yielded the ester products 1-hydroxyethyl
acetate and 1-methoxyethyl formate.44 These products would
result from breaking the O–O bond in the secondary ozonide,
followed by a 1,4-H or 1,4-methyl shift, and their presence in
laser-sensitized decomposition suggests that such rearrangements
may precede decomposition of the secondary ozonide to
carboxylic acid and ketone or aldehyde products. The obser-
vation of similar products in the present direct reactions of
CH2OO with carbonyl species indicates that the transition
states for such transformations likely lie near or below the
energy of the reactants.
at Lawrence Berkeley National Laboratory. Sandia is a multi-
program laboratory operated by Sandia Corporation, a Lockheed
Martin Company, for the National Nuclear Security Administra-
tion under contract DE-AC04-94-AL85000.
References
1 R. Criegee and G. Wenner, Justus Liebigs Ann. Chem., 1949, 564, 9–15.
2 O. Welz, J. D. Savee, D. L. Osborn, S. S. Vasu, C. J. Percival,
D. E. Shallcross and C. A. Taatjes, Science, 2012, 335, 204–207.
3 R. Asatryan and J. W. Bozzelli, Phys. Chem. Chem. Phys., 2008,
10, 1769–1780.
4 C. A. Taatjes, G. Meloni, T. M. Selby, A. J. Trevitt, D. L. Osborn,
C. J. Percival and D. E. Shallcross, J. Am. Chem. Soc., 2008, 130,
11883–11885.
5 D. Johnson and G. Marston, Chem. Soc. Rev., 2008, 37, 699–716.
6 N. M. Donahue, G. T. Drozd, S. A. Epstein, A. A. Presto and
J. H. Kroll, Phys. Chem. Chem. Phys., 2011, 13, 10848–10857.
7 R. Criegee, Angew. Chem., Int. Ed. Engl., 1975, 14, 745–752.
8 O. Horie, C. Schafer and G. K. Moortgat, Int. J. Chem. Kinet.,
1999, 31, 261–269.
9 P. Neeb, O. Horie and G. K. Moortgat, Tetrahedron Lett., 1996,
37, 9297–9300.
10 R. Fajgar, J. Vıtek, Y. Haas and J. Pola, J. Chem. Soc., Perkin
Trans. 2, 1999, 239–248.
Conclusions
11 O. Horie and G. K. Moortgat, Acc. Chem. Res., 1998, 31, 387–396.
12 C. A. Taatjes, N. Hansen, D. L. Osborn, K. Kohse-Hoinghaus,
T. A. Cool and P. R. Westmoreland, Phys. Chem. Chem. Phys.,
2008, 10, 20–34.
13 D. L. Osborn, P. Zou, H. Johnsen, C. C. Hayden, C. A. Taatjes,
V. D. Knyazev, S. W. North, D. S. Peterka, M. Ahmed and
S. R. Leone, Rev. Sci. Instrum., 2008, 79, 104103.
14 C. A. Taatjes, N. Hansen, A. McIlroy, J. A. Miller, J. P. Senosiain,
S. J. Klippenstein, F. Qi, L. S. Sheng, Y. W. Zhang, T. A. Cool,
J. Wang, P. R. Westmoreland, M. E. Law, T. Kasper and
K. Kohse-Hoinghaus, Science, 2005, 308, 1887–1889.
15 A. J. Eskola, D. Wojcik-Pastuszka, E. Ratajczak and R. S. Timonen,
Phys. Chem. Chem. Phys., 2006, 8, 1416–1424.
Absolute total rate coefficients have been directly measured for
the removal of CH2OO by reaction with hexafluoroacetone
(HFA), acetone, and acetaldehyde. The reaction with HFA is
rapid, tending to validate the use of HFA as a Criegee
intermediate scavenger. Secondary ozonide products are observed
for reactions of CH2OO with acetone and HFA, but the
presence of ozonide isomerization products and the observation
of acetic acid as a product of the reaction of CH2OO with
acetaldehyde suggest that transition states for isomerization
and decomposition pathways of the secondary ozonide are
near or below the energy of the reactants.
16 M. Olzmann, E. Kraka, D. Cremer, R. Gutbrod and S. Andersson,
J. Phys. Chem. A, 1997, 101, 9421–9429.
17 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li,
H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. Montgomery, J. A. J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin,
V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari,
A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,
N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross,
V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A.
Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels,
O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and
D. J. Fox, Gaussian 09, Revision A1.
18 J. A. Montgomery, M. J. Frisch, J. W. Ochterski and
G. A. Petersson, J. Chem. Phys., 1999, 110, 2822–2827.
19 H.-J. Werner, P. J. Knowles, F. R. Manby, M. Schutz, P. Celani,
G. Knizia, T. Korona, R. Lindh, A. Mitrushenkov, G. Rauhut,
T. B. Adler, R. D. Amos, A. Bernhardsson, A. Berning, D. L.
Cooper, M. J. O. Deegan, A. J. Dobbyn, F. Eckert, E. Goll,
C. Hampel, A. Hesselmann, G. Hetzer, T. Hrenar, G. Jansen,
C. Koppl, Y. Liu, A. W. Lloyd, R. A. Mata, A. J. May,
S. J. McNicholas, W. Meyer, M. E. Mura, A. Nicklass,
P. Palmieri, K. Pfluger, R. Pitzer, M. Reiher, T. Shiozaki,
H. Stoll, A. J. Stone, R. Tarroni, T. Thorsteinsson, M. Wang
and A. Wolf, MOLPRO, version 2010.1, a package of ab initio
programs. http://www.molpro.net.
Author contributions
The investigations were conceived by CJP, DES, and CAT.
The experimental studies were designed, carried out and
analyzed by CAT, OW, AJE, JDS, DLO, and CJP, with
assistance from DES. EPFL, JMD, OW, JDS, and DWKM
were responsible for quantum chemistry and Franck–Condon
calculations.
Acknowledgements
This work is supported by the Division of Chemical Sciences,
Geosciences, and Biosciences, the Office of Basic Energy
Sciences, the U. S. Department of Energy. DES, CJP, and
JMD thank NERC for funding. EPFL, JMD and DKWM
acknowledge support from Research Grant Council (RGC) of
the Hong Kong Special Administrative Region (HKSAR,
Grant No. PolyU 5018/09P), the National Service for Com-
putational Chemistry Software (NSCCS), EPSRC (UK), and
the IRIDIS High Performance Computing Facility at the
University of Southampton. JMD thanks the Leverhulme
Trust for an Emeritus Fellowship. The Advanced Light Source
is supported by the Director, Office of Science, Office of Basic
Energy Sciences, Materials Sciences Division, of the U. S.
Department of Energy under Contract DE-AC02-05CH11231
20 T. H. Dunning, J. Chem. Phys., 1989, 90, 1007–1023.
21 K. A. Peterson, Thomas B. Adler and H.-J. Werner, J. Chem.
Phys., 2008, 128, 084102.
c
This journal is the Owner Societies 2012
Phys. Chem. Chem. Phys., 2012, 14, 10391–10400 10399

View Online
22 J. G. Hill, S. Mazumder and K. A. Peterson, J. Chem. Phys., 2010,
132, 054108.
23 T. Helgaker, W. Klopper, H. Koch and J. Noga, J. Chem. Phys.,
1997, 106, 9639–9646.
24 A. Halkier, T. Helgaker, W. Klopper, P. Jørgensen and
A. G. Csaszar, Chem. Phys. Lett., 1999, 310, 385–389.
25 M. E. Harding, J. Vazquez, B. Ruscic, A. K. Wilson, J. Gauss and
J. F. Stanton, J. Chem. Phys., 2008, 128, 114111.
26 E. F. Valeev and T. D. Crawford, J. Chem. Phys., 2008, 128, 244113.
27 G. Czako, B. Nagy, G. Tasi, A. Somogyi, J. Simunek, J. Noga,
B. J. Braams, J. M. Bowman and A. G. Csaszar, Int. J. Quantum
Chem., 2009, 109, 2393–2409.
28 X. Huang, E. F. Valeev and T. J. Lee, J. Chem. Phys., 2010,
133, 244108.
29 F. Neese and E. F. Valeev, J. Chem. Theory Comput., 2010, 7, 33–43.
30 G. Knizia, T. B. Adler and H.-J. Werner, J. Chem. Phys., 2009,
130, 054104.
33 F. Duschinsky, Acta Physicochim. URSS, 1937, 7, 551.
34 G. T. Drozd and N. M. Donahue, J. Phys. Chem. A, 2011, 115,
4381–4387.
35 D. Cremer, T. Schmidt, W. Sander and P. Bischof, J. Org. Chem.,
1989, 54, 2515–2522.
36 M. T. Nguyen, T. L. Nguyen, V. T. Ngan and H. M. T. Nguyen,
Chem. Phys. Lett., 2007, 448, 183–188.
37 W. Sander, Angew. Chem., Int. Ed. Engl., 1990, 29, 344–354.
38 L. B. Harding and W. A. Goddard, J. Am. Chem. Soc., 1978, 100,
7180–7188.
39 W. H. Bunnelle, Chem. Rev., 1991, 91, 335–362.
40 R. W. Murray, P. R. Story and L. D. Loan, J. Am. Chem. Soc.,
1965, 87, 3025–3026.
41 R. Criegee, S. S. Bath and B. von Bornhaupt, Chem. Ber., 1960, 93,
2891–2897.
42 J. D. Fenske, A. S. Hasson, A. W. Ho and S. E. Paulson, J. Phys.
Chem. A, 2000, 104, 9921–9932.
31 T. Helgaker, W. Klopper and D. P. Tew, Mol. Phys., 2008, 106,
2107–2143.
43 R. I. Martinez, J. T. Herron and R. E. Huie, J. Am. Chem. Soc.,
1981, 103, 3807–3820.
32 V. A. Mozhayskiy and A. I. Krylov, ezSpectrum. http://iopenshell.
usc.edu/downloads This work was conducted using the resources
of the iOpenShell Center for Computational Studies of Electronic
Structure and Spectroscopy of Open-Shell and Electronically
Excited Species (http://iopenshell.usc.edu) supported by the
National Science Foundation through the CRIF:CRF program.
44 L. Khachatryan, R. Fajgar, Y. Haas and J. Pola, J. Chem. Soc.,
Perkin Trans. 2, 1996, 1981–1984.
45 A. G. Baboul, L. A. Curtiss, P. C. Redfern and K. Raghavachari,
J. Chem. Phys., 1999, 110, 7650–7657.
46 L. A. Hull, I. C. Hisatsune and J. Heicklen, J. Phys. Chem., 1972,
76, 2659–2665.
c
10400 Phys. Chem. Chem. Phys., 2012, 14, 10391–10400
This journal is the Owner Societies 2012