ARTICLE

pubs.acs.org/JPCA

Computational Study of the Reaction Mechanism of the

Methylperoxy Self-Reaction
Yan-Ni Liang,† Jun Li,‡ Quan-De Wang,† Fan Wang,*,† and Xiang-Yuan Li‡
†
College of Chemistry and ‡College of Chemical Engineering, Sichuan University, Chengdu 610065, People’s Republic of China
bS Supporting Information
ABSTRACT: To provide insight on the reaction mechanism of the methyperoxy (CH3O2•)
self-reaction, stationary points on both the spin-singlet and the spin-triplet potential energy
surfaces of 2(CH3O2•) have been searched at the B3LYP/6-311++G(2df,2p) level. The relative
energies, enthalpies, and free energies of these stationary points are calculated using CCSD(T)/
cc-pVTZ. Our theoretical results indicate that reactions on a spin-triplet potential energy surface
are kinetically unfavorable due to high free energy barriers, while they are more complicated on the spin-singlet surface.
CH3OOCH3 + O21 can be produced directly from 2(CH3O2•), while in other channels, three spin-singlet chain-structure
intermediates are first formed and subsequently dissociated to produce different products. Besides the dominant channels producing
2CH3O• + O2 and CH3OH + CH2O + O2 as determined before, the channels leading to CH3OOOH + CH2O and CH3O• + CH2O +
HO2• are also energetically favorable in the self-reaction of CH3O2• especially at low temperature according to our results.

1. INTRODUCTION channels have been suggested for this reaction:11,12

Peroxyl radicals (RO2•, R = H or any organic group) have 2CH3 O2 • f 2CH3 O• þ O2 ð7aÞ
been the subject of intensive research during the last decades,
both experimentally and theoretically,1
3 since RO2• are
important trace tropospheric constituents that are involved 2CH3 O2 • f CH2 O þ CH3 OH þ O2 ð7bÞ
in the oxidation and production of many other atmospheric
species.4 For example, RO2• are responsible for ozone pro- 2CH3 O2 • f CH3 OOCH3 þ O2 ð7cÞ
duction in the troposphere as well as for the formation of
hydrogen peroxide (H2O2) and hydroperoxides (ROOH).5 In 2CH3 O2 • f CH3 OOH þ •CH2 O2 • ð7dÞ
addition, they are key reactive intermediates in the oxidation
and combustion of hydrocarbons,6,7 especially at low tem- Some experimental studies have been conducted to investigate
perature and moderate pressure conditions, which is initiated the kinetics of the self-reaction of CH3O2• previously. Niki and
by the following addition of molecular oxygen to the major co-workers13 studied the self-reaction of CH3O2• by means of
alkyl radicals:8 Fourier transform infrared spectroscopy (FTIR) at 297 K. Besides
the above primary reactions 7a
7d, the following secondary reac-
R• þ O2 f RO2 • ð1Þ tions were also proposed:

RO2 • þ HO2 • f ROOH þ O2 ð2Þ CH3 O• þ O2 f CH2 O þ HO2 • ð8Þ

RO2 • þ H2 O2 f ROOH þ HO2 • ð3Þ CH3 O2 • þ HO2 • f CH3 OOH þ O2 ð9Þ

RO2 • þ CH3 O2 • f RO• þ CH3 O• þ O2
RO2 • þ R 0 O2 • f R 0 O• þ RO• þ O2
ROOH f RO• þ •OH
The methylperoxy radical (CH3O2•) is an important peroxy
radical in the oxidation of hydrocarbons and other organic
compounds.9,10 The self-reaction of CH3O2• becomes sig-
nificant in the atmospheric oxidation when nitrogen oxides
are present in low concentration6 or in low temperature com-
bustion of methane as shown in reaction 4.8 Four different
ð4Þ The relative rate constants for reactions 7a
7c were determined
to be 0.32:0.60:0.08, respectively, which indicated that path 7c
was not important among all reactions. A similar conclusion has
ð5Þ also been drawn for the self-reaction of C2H5O2 radicals.14
Lightfoot et al.7 performed flash photolysis studies of the
ð6Þ CH3O2• self-reaction from 248 to 573 K. Their results showed
paths 7a and 7b to be highly competitive with a strong tempera-
ture dependence in their branching ratios, and the contribution
from pathway 7c was insignificant at all temperatures. According
Received: May 24, 2011
Revised: October 15, 2011
Published: October 17, 2011

r 2011 American Chemical Society 13534 dx.doi.org/10.1021/jp2048508 | J. Phys. Chem. A 2011, 115, 13534–13541
The Journal of Physical Chemistry A
Scheme 1
to their results, the existence of reaction 7d up to 573 K is not
supported by their data, and the radical-producing channel 7a is
by far the most important, which would account for nearly 90% of
all reactions at 700 K.7 They found that although reaction 7a acts
as a termination reaction under atmospheric conditions, it largely
serves to convert CH3O2• to HO2• under combustion condi-
tions via reaction 8. Tyndall et al.15 determined the products of
the self-reaction of CH3O2• at 296 K using continuous ultraviolet
(UV) photolysis with FTIR detection. Their experiments were
carried out through the photolysis of AZM (azomethane)-O2
or Cl2
CH4
O2 mixtures to determine the product yields
from the CH3O2• self-reaction. The major products observed
were CH2O, CH3OH, and CH3OOH with minor amounts
of CO and HCOOH. Wallington and co-workers16
18 have
shown that CH3OOH is the major product of reaction 9, while
CO and HCOOH are formed from secondary reactions involv-
ing Cl atoms.19 There was no evidence for the production of
CH3OOCH3 (yield <6%) according to their results. The yield of
CH3OOH was lower with AZM photolysis than that with the
photolysis of Cl2
CH4, which was also found by Niki and co-
workers13 and Kan et al.20 (who only used AZM). The reason for
the low CH3OOH yields in the AZM experiments remains
unknown, although they may be interpreted by a small loss of
HO2• or that the infrared absorption of CH3OOH is overlapped
by those of CH3OH and azomethane.13,15,20,21 The branching
fraction for the reaction channel 7a is found to be 41 ( 4% at 296 K,
which is in good agreement with two earlier studies. In a later
work, Tyndall et al.22 obtained the branching ratio of methoxy
radical (CH3O•) production to be 0.10 at 223 K and 0.82 at 573 K,
which is also consistent with previous studies. In general, FTIR
can not detect radical products in these experiments due to their
low concentrations, and UV absorption spectroscopy is not
sensitive enough to detect small amounts of radical products
from self-reaction.23
The possible mechanism for the CH3O2• self-reaction
was first proposed by Russel24 via a tetroxide intermediate
(CH3O2)2 and a cyclic transition state (TS) for reaction 7b. Later
on, Lightfoot et al.7 also suggested possible TSs for reactions 7a
and 7c based on the Russel mechanism, as illustrated in Scheme 1.
However, the proposed mechanism is not fully compatible with
the experimental results and available experimental data could
not provide definitive mechanism information.7 To assess the
relative importance of these processes and gain insight into the
reaction mechanism of the CH3O2• self-reaction, a systematic
theoretical investigation is needed. In addition, reliable mechanism
is critical to characterize the kinetic behavior of the self-reaction
13535
ARTICLE
of CH3O2• in combustion chemistry as well as to develop reliable
chemical models involving peroxyl radicals.
Various theoretical calculations have been performed pre-
viously on structures, energies, and thermochemical data of
CH3O2• and corresponding self-reaction products based on
Hartree
Fock (HF),25
27 second-order Moller
Plesset per-
turbation theory (MP2),28 density functional theory (DFT)29
using the B3LYP functional (Becke’s three parameter nonlocal
exchange functional combining with the nonlocal correlation
functional of Yang et al.),30,31 and some composite methods,
e.g., the G3 method,32 respectively.33,34 Rather reliable energies
for CH3O2• and its self-reaction products were achieved. How-
ever, the detailed mechanism for the CH3O2• self-reaction re-
mains controversial. The mechanism governing reaction 7a in gas
phase was determined at the MP2/6-311G(2d,2p) and PMP4-
(SDTQ)/6-311G(3df,2p)//MP2/6-311G(2d,2p) levels by
Feria et al. previously.35 Their results indicated that the reaction
proceeds through a nonconcerted mechanism involving two
tetroxides to the products.35 Jalbout et al.36 explored both spin-
singlet and spin-triplet potential energy surfaces (PESs) of the
2(CH3O2•) system at the B3LYP/6-311G(d,p) level of theory
and calculated single-point energies of the involving species based
on the MP2 and B3LYP methods. However, the relative energies
of the tetroxide intermediates with respect to 2(CH3O2•) are
significantly overestimated based on B3LYP and underestimated
with MP2. A similar problem has also been observed for those
in the CH3CH2O2• self-reaction.37 Ghigo et al.6 investigated
the mechanism of the CH3O2• self-reaction based on the CAS-
(16,12)-PT2/6-311G(2df,p)//CAS(8,8)-MCSCF/6-311G(d,p)
approach. Their results exclude pathways leading directly from
the tetroxide intermediate to the most important products
2CH3 O• + O2 . Instead, they proposed an indirect path-
way through a weakly bound complex of [2(CH3O•)-O2] to
products. Fairly good qualitative agreement on temperature
dependence of the relevant branching ratio with experimental
results was achieved. However, they did not investigate the
reaction leading to CH3OOCH3 , and their mechanism seems
to allow for easy formation of CH3OOCH3 from the weakly
bound complex [2(CH3O•)-O2],38 which is not consistent with
experimental results. Furthermore, internal rotations along the
O
O bonds in tetroxide would result in several stable chain-
structure isomers, while Ghigo et al. only considered one of them,
which is not the most stable one either.6,34
The purpose of this work is to provide insight on the reac-
tion mechanism through exploration of the PES of 2(CH3O2•)
at a high computation level. We do not attempt to evaluate the
reaction rates due to the possible crossover of spin-singlet and
spin-triplet PESs.6 In our investigations, stationary points are
optimized based on the B3LYP functional. Single-point en-
ergies are calculated at the coupled-cluster approach with
single and double excitations augmented by a perturbative
treatment of triple excitations (CCSD(T)) level,39,40 which is
currently the most accurate single reference method for energy
calculations with practical applications. Meanwhile, the enthalpy
and free energy differences for the related reactions have also been
calculated for rationalizing product branching ratios.
2. COMPUTATIONAL DETAILS
On the basis of previous computational studies for reactions of
HO2• with CH3CH2O2•,41 HO2• with CH3O2•,42 and the self-
reaction of HO2•,43 we proposed the following additional
dx.doi.org/10.1021/jp2048508 |J. Phys. Chem. A 2011, 115, 13534–13541
The Journal of Physical Chemistry A
possible reaction channels besides reactions 7a
7d:
2CH3 O2 • f CH3 O• þ CH2 O þ HO2 • ð10aÞ
2CH3 O2 • f CH3 OCH3 þ O3 ð10bÞ
2CH3 O2 • f CH3 OOOH þ CH2 O ð10cÞ
2CH3 O2 • f CH4 þ CH2 O þ O3 ð10dÞ
The geometry of each stationary point along the PES has been
optimized using the B3LYP functional and the 6-311++G-
(2df,2p) basis set.44 It has been amply demonstrated that the
B3LYP functional is able to provide reliable geometries, zero-
point energies (ZPE),45 and frequencies46 for stationary points
on PES for organic reactions with a relatively high computational
efficiency. In addition, according to Bozzelli et al.’s results,34
geometries of the chain-structure tetroxide (CH 3O 2)2 and
related radicals based on B3LYP agree well with those from
QCISD calculations, with a difference of typically less or equal to
0.02 Å. Harmonic vibrational frequencies of each stationary point
were also calculated at the same level and served to compute the
ZPE without scaling. Spin-unrestricted calculations are per-
formed for spin-singlet states since wave functions for some
spin-singlet species based on the spin-restricted method are
found to be unstable. To achieve stable wave functions, using
the spin-unrestricted method for spin-singlet states will necessa-
rily introduce spin-contamination. Nonetheless, this is a com-
mon practice in the single reference method to describe bond
broken processes, and reliable geometries may still be obtained if
spin-contamination is not serious, which is the case in the present
study. However, spin-unrestricted calculations for the transition
states of reactions 7b and 7c on the spin-singlet PES in which
spin-singlet oxygen molecule is produced could result in serious
spin-contaminated results since the spin-triplet oxygen molecule
is much more stable than the spin-singlet oxygen molecule. Spin-
restricted calculations are thus adopted for the transition states of
reactions 7b and 7c on the spin-singlet PES. All stationary points
were identified to be either local minima without imaginary
frequencies or transition states with only one imaginary fre-
quency. Furthermore, the intrinsic reaction coordinate (IRC)47
calculations were carried out to validate the connection between
transition states and the corresponding reactants or products. To
achieve reliable relative energies of stationary points along the
PES, the spin-unrestricted CCSD(T) method with the cc-pVTZ
basis set48 was employed to calculate the single-point energies at
the B3LYP/6-311++G(2df,2p) optimized geometries.
The geometry optimizations and vibrational analysis, i.e.,
the DFT calculations, were carried out using the Gaussian 03
program package49 and single-point CCSD(T) energy calcula-
tions were performed with the CFOUR program package.50
However, the Thermo program51 is adopted to calculate the
entropies and heat capacities of the involved species at different
temperatures.
3. RESULTS AND DISCUSSION
The reaction channels found in the present work on spin-singlet
and spin-triplet PESs are described in Schemes 2 and 3. Figure 1
shows the optimized geometries of important reactants, pro-
ducts, and intermediates as well as transition states involved in
the self-reaction of CH3O2• calculated with B3LYP/6-311++G-
(2df,2p). The potential energy diagram is illustrated in Figure 2.
13536
ARTICLE
Scheme 2
Scheme 3
In addition, C
O and O
O bond lengths for the most stable
isomer of (CH3O2)2, CH3O2• and CH3O•, obtained in the
present work are listed in Table 1 together with those based on
B3LYP/cc-pVTZ+d and QCISD/6-311G(d,p) from ref 34 for
comparison. It can be seen from this table that the calculated
bond lengths for these species generally agree better with those
calculated with QCISD/6-311G(d,p) than those from B3LYP/
cc-pVTZ+d. Entropies at 298 K (S298) and heat capacities
(Cp(T)) in the temperature range from 298 to 1500 K for the
most stable isomer of (CH3O2)2, CH3O2• and CH3O•, using
B3LYP/6-311++G(2df,2p) geometries are presented in Table 2.
Our results again agree well with those with B3LYP/cc-pVTZ+d
by Bozzelli,34 and the difference may stem from the different
basis set and from contributions of internal rotors, which are
treated with harmonic vibrations in the present work. The
calculated relative energies at 0 K (designated by ΔE0), enthal-
pies (designated by ΔH), and free energies (designated by ΔG)
at different temperatures, for (2CH3O• + O23) and (CH3OH +
CH2O + O2) as well as two isomers of (CH3O2)2 are listed in
Table 3, together with previous results based on CASPT2, CBS-
Q , G2, CBS-APNO, G3B3, B3LYP, and MP2 calculations6,34,36
for comparison. One can see from this table that the present
results are generally in good agreement with those calculated
with CAS-PT2(16,12)/6-311G(2df,p)//CAS-MCSCF(8,8)/
6-311G(d,p)6 and those with CBS-Q.6,38 Furthermore, the relative
energy, entropy, and free energy of the most stable isomer of
(CH3O2)2 in this work are consistent with the average of CBS-
APNO and G3B3 results using isodesmic reactions.34 These
results demonstrate the reliability of methods employed in the
present work to study the reaction mechanism of the CH3O2•
self-reaction. It is also worth noting from Table 3 that the relative
energies calculated at the B3LYP and MP2 levels are unreliable.
The energies related to the reactants 2(CH3O2•) at 0 K, the rela-
tive enthalpies, the free energies at 298 and 900 K of the products
of these channels, and the intermediates as well as the transi-
tion states are listed in Table 4. The lowest spin-singlet state of
O2 is the 1Δg state and lies 22.6 kcal mol
1 above the ground
3

Σg state,52 while it is calculated to be 30.3 kcal mol
1 according
to the present CCSD(T) method. For the lowest spin-singlet
state of O2, a multireference approach must be adopted to
achieve reasonable excitation energy, and the single reference
CCSD(T) method should not be used. An experimental value is
dx.doi.org/10.1021/jp2048508 |J. Phys. Chem. A 2011, 115, 13534–13541
The Journal of Physical Chemistry A ARTICLE

Figure 1. Optimized geometries of the reactant (CH3O2•), intermediates (IM1
IM6), transition states (TS1
TS8), and products (•CH2O2•,
CH3OCH3, CH3O•, CH3OOCH3, and CH3OOOH) involved in the self-reaction of CH3O2• at the B3LYP/6-311++G (2df, 2p) level. The bond
lengths are given in angstroms, and the angles are given in degrees.

either on the spin-singlet PES or on the spin-triplet PES. The

Figure 2. Schematic energy diagram of the CH3O2• + CH3O2•
system computed at the CCSD(T)/cc-pVTZ//B3LYP/6-311++G-
(2df, 2p) level.
thus used to correct the energy of the 1Δg state of oxygen. As
shown in Figure 2, the CH3O2• + CH3O2• reaction can occur
13537
detailed mechanisms for these processes are discussed in the
following sections.
3.1. Stable Isomers of (CH3O2)2. The methylperoxy radicals
react readily in a barrierless step to form chain-structure or ring-
structure tetroxides. Because of the internal rotations in the
tetroxides (CH3O2)2 along the CH3OO
OOCH3 bond or the
CH 3 O-OOOCH 3 bond, 34 three chemically bonded chain-
structure intermediates with similar energies have been iden-
tified according to our calculations, i.e. IM1, IM2, and IM3.
IM1 and IM2 both have C2 symmetry and the dihedral angles
D(C9
O1
O2
O3) and D(C5
O2
O3
O4) are both equal
to 80.8° in IM1 and 119.2° in IM2, respectively. With the
O1
O2
O3
O4 dihedral angle in IM1 varying from 79.8° to

84.4°, which corresponds to internal rotation along the
CH3OO
OOCH3 bond, there exists another minimum IM2. In
addition, the bridging O2
O3 bond in IM2 is 0.055 Å shorter
than that in IM1; however, the O1
O2 and O3
O4 bonds in the
CH3O2 groups of IM2 are 0.038 Å longer than those in IM1.
Because of the rotation of the CH3O moiety in IM1 along the
CH3O
OOOCH3 bond, another stable isomer, i.e., IM3 can be
found, with the C9
O1
O2
O3 dihedral angle of
95.2°.
Compared with those of IM1, the O1
O2 and O3
O4 bond
dx.doi.org/10.1021/jp2048508 |J. Phys. Chem. A 2011, 115, 13534–13541
The Journal of Physical Chemistry A ARTICLE

Table 1. Bond Lengths in Structures of (CH3O2)2 (IM1) and lengths of the CH3O2 groups in IM3 are 0.003 Å and 0.013 Å
Its Radicals at the B3LYP/6-311++g(2df,2p), Comparing longer, respectively, and the O2
O3 bond of IM3 is 0.012 Å
with Parameters at B3LYP/cc-pVTZ+d, and QCISD/ shorter. The relative energies (see Figure 2) of IM1, IM2, and
6-311G(d,p) Theoretical Levels from ref 34 IM3 are
13.0,
10.9, and
12.3 kcal mol
1 at the CCSD(T)/
cc-pVTZ level, respectively. The energy differences among the
bond length (Å) three minima are less than 2.5 kcal mol
1 and agree well with the
results of Bozzelli.34 The fact that the most stable chain-structure
B3LYP/6-311 B3LYP/ QCISD/ intermediate is IM1 may be due to its relatively smaller steric and
parameter ++g(2df,2p) cc-pVTZ+d 6-311G(d,p) electrostatic interaction between the methyl groups according to
CH3
OOOOCH3 1.4251 1.4249 1.4254 its trans-structure, as indicated in Figure 1. Besides these chain-
CH3O
OOOCH3 1.4288 1.4311 1.4156
structure isomers, there also exist two four-membered ring
minima on the spin-singlet PES, IM4, and IM5 with C2 and Ci
CH3OO
OOCH3 1.4131 1.4169 1.4082
symmetry based on our B3LYP calculations, respectively. The
CH3
OO 1.4461 1.4465 1.4409 same intermediates have been found in previous work.36 The
CH3O
O 1.3148 1.3171 1.3241 O1
O3 bond length in IM4 and O2
O4 bond length in IM5 are
CH3
O 1.3625 1.3627 1.3790 ca. 2.15 Å, which are longer than the normal O
O bond length

Table 2. Computed Entropies and Heat Capacities and Symmetry Point Groups for (CH3O2)2 (IM1) and Its Radicals from
B3LYP/6-311++g(2df,2p) and B3LYP/cc-pVTZ+d Calculations by Bozzellia
point group S298 Cp(298) Cp(400) Cp(500) Cp(600) Cp(800) Cp(1000) Cp(1500) ref

IM1 83.58 26.80 32.34 37.10 41.08 47.18 51.59 58.31 this work
C2 88.13 27.46c 32.58 36.95 40.59 46.20 50.31 56.68 Bozzellib
CH3O2• 64.32 12.55 14.79 16.92 18.77 21.69 23.84 27.16 this work
Cs 66.86 12.02c 14.06 16.11 17.91 20.78 22.90 26.20 Bozzellib
CH3O• 56.65 9.83 11.59 13.26 14.73 17.12 18.95 21.83 this work
C1 56.64 9.83c 11.57 13.24 14.71 17.12 18.95 21.84 Bozzellib
S is in cal mol
1 K
1; Cp(T) is in cal mol
1 K
1, and the value in parentheses is temperature in K. b Results from Bozzelli et al. (ref 34). c Heat
a 298

capacities at 300 K.

Table 3. Relative Energies (0 K),a Enthalpies (ΔH298), and Free Energy Differences (ΔGT) for Species Involved in the Self-
Reaction of CH3O2• at Different Levels of Theory
species ΔE0 (kcal mol
1) ΔH298 (kcal mol
1) ΔG298 (kcal mol
1) ΔG900 (kcal mol
1) ref

R(2CH3O2•) 0.0 0.0 0.0 0.0

P3(CH3OH + CH2O + O21)
58.7
58.3
66.5
84.1 this work

58.3
57.0
65.5
52.4 Ghigoa
P9(2CH3O• + O23) 0.4 0.8
9.2
30.9 this work
1.6 2.7
7.7
28.5 Ghigoa
4.4 5.5
4.9
25.3 Ghigob
6.5 7.6
2.7
23.4 Ghigoc
P10(CH3OH + CH2O + O23)
81.3
80.8
89.7
108.5 this work

82.3
81.0
90.2
106.4 Ghigoa

81.3
80.1
89.2
106.9 Ghigob

80.5
79.3
88.4
106.3 Ghigoc
IM1
13.0
13.1 0.3 27.4 this work

13.9
14.0
0.4 Bozzellid

5.4 Jalboute

20.5 Jalboutf
IM2
10.9
10.9 2.2 28.7 this work

8.7
9.2 3.3 27.6 Ghigoa

12.0
12.8
0.1 24.2 Ghigob

14.3
15.0
2.5 21.8 Ghigoc

3.4 Jalboute

18.1 Jalboutf

10.9 Dibbleg
a
At the CAS-PT2(16,12)/6-311G(2df, p)//MCSCF(8,8)/6-311G(d,p) level (ref 6). b At CBS-Q level (ref 6). c At G2 level (ref 6). d Average of CBS-
APNO and G3B3 calculations using isodesmic reactions (ref 34). e At the B3LYP/6-311++G(3df, 2p)//B3LYP/6-311G(d, p) level (ref 36). f At the
MP2/6-311G(d, p) // B3LYP/6-311G(d, p) level (ref 36). g At the CBS-Q level (ref 38).

13538 dx.doi.org/10.1021/jp2048508 |J. Phys. Chem. A 2011, 115, 13534–13541

The Journal of Physical Chemistry A ARTICLE

Table 4. Relative Energies (0 K),a Enthalpies, and Free Energies at 298 and 900 K, Calculated at the CCSD(T)/cc-pVTZ Level
Using B3LYP/6-311++G(2df,2p) Optimized Geometries for the Self-Reaction of CH3O2•
species ΔE0 (kcal mol
1) ΔH298 (kcal mol
1) ΔH900 (kcal mol
1) ΔG298 (kcal mol
1) ΔG900 (kcal mol
1)

R(2CH3O2•) 0.0 0.0 0.0 0.0 0.0

P1(CH3OOCH3 + O21)
13.3b
13.4
13.0
11.0
6.5
P2 (CH3O• + CH2O + HO2•)
25.9
25.2
27.1
35.6
57.7
P3 (CH3OH + CH2O + O21)
58.7b
58.3
60.4
66.5
84.1
P4(CH3OCH3 + O3)
12.5
12.8
13.2
10.3
4.8
P5(CH3OOOH + CH2O)
52.0
51.9
51.5
50.1
46.8
P6(CH4 + CH2O + O3)
14.1
13.5
16.0
21.0
36.7
P7(CH3OOCH3 + O23)
36.0
35.9
35.5
34.2
30.9
P8(CH3OOH + •CH2O2•) 19.5 19.8 21.2 19.5 17.7
P9(2CH3O• + O23) 0.4 0.8
0.5
9.2
30.9
IM1
13.0
13.1
9.9 0.3 27.4
IM2
10.9
10.9
7.7 2.2 28.7
IM3
12.3
12.4
9.2 0.4 26.4
TS1(R f P1) 27.1 27.0 30.0 40.3 67.1
TS2(IM1 f P2) 10.5 10.0 12.6 24.0 52.8
TS3(IM1 f P3) 20.3 20.2 23.1 33.0 59.0
TS4(IM2 f P4) 77.9 78.3 81.5 90.1 113.8
TS5(IM3 f P5) 7.8 7.9 10.9 20.1 44.9
TS6(IM3 f P6) 71.9 71.7 75.2 85.1 111.9
TS7(R f P7) 58.6 59.3 62.4 68.8 87.9
TS8(R f P8) 24.4 24.9 28.4 34.5 53.8
a
The zero point energy correction is included. b The energy for the1Δg state of O2 is corrected based on an experimental value.52

1.48 Å according to our calculations. These intermediates are in
fact unstable since their energies are even higher than that of
2(CH3O2•) calculated with B3LYP and CCSD(T). Besides the
spin-singlet intermediates indicated above, the two CH3O2
radicals can also form a triplet planar eight-membered ring
intermediate IM6 with two hydrogen bonds. It is
5.2 kcal
mol
1 related to 2(CH3O2•) based on CCSD(T)/cc-pVTZ.
According to Figure 1, the distance between the terminal oxygen
atoms of CH3O2 is 3.972 Å, and the O
H bond length is 2.489 Å
calculated with B3LYP. We do not find evidence for these three
loose ring-structure tetroxides to lead to any product in present
work. Therefore, we ignore them in the following calculations.
The saddle points that connect different chain-structure
intermediates are also found. However, the imaginary frequen-
cies are very small and correspond to modes involving the
rotation of the methyl group around the C
O or O
O bond,
as shown previously in the self-reaction of HO2•.34,43 These
transition states could not have a significant impact on the self-
reactions of CH3O2• and are thus not considered further in the
present study. The geometries in the Z-matrix, rotational con-
stants, harmonic frequencies, and average of S2 of all the stable
species as well as the transition states involved in the self-reaction
of CH3O2• are provided in the Supporting Information.
3.2. Product Formation. 3.2.1. Spin-Singlet Product Chan-
nels. CH3OOCH3 + O21. As shown in Figure 2, CH3OOCH3 +
O21 can be produced through TS1 with an imaginary frequency
of 537 cm
1. From the structure of TS1, it can be seen that the O
atom in one of the CH3O2• (see Figure 1) attacks the C atom of
the other CH3O2• with C5
O4 bond stretching to ca. 2 Å and
O3
O4
C5
O2 dihedral angles of ca.
120° and C9
O1

O2
C5 of ca. 96°. This is a bimolecular nucleophilic substitution
reaction with the attacking nucleophile CH3O2• delivered from
13539
the same side as the leaving group oxygen. IRC analysis shows
that TS1 directly connects the products CH3OOCH3 + O21 with
the reactants 2(CH3O2•). This is different from the mechanism
proposed by Lightfoot, where the tetroxide intermediate is
involved as indicated in Scheme 1. The calculated barrier of this
process lies 27.1 kcal mol
1 above the reactants. This channel is
kinetically unimportant due to its high relative free energies at
298 K (ΔG298 = 40.3 kcal mol
1), which is consistent with
previous findings.14
CH3O• + CH2O + HO2•. The intermediate IM1 decomposes to
form CH3O• + CH2O + HO2• via a bicycle transition state TS2,
with an imaginary frequency of 659 cm
1. From the structure of
TS2 shown in Figure 1, the hydrogen (H7) is migrating from C5
to terminal O2 to produce HO2• while breaking the remaining
O3
O4 (2.05 Å) and O1
O2 (1.84 Å) bonds at the same time,
with a H7
O2
O3 angle of 101.2°. The energy barrier for this
channel is predicted to be 23.5 kcal mol
1, and the relative free
energy at 298 K is 23.7 kcal mol
1, with respect to IM1. CH3O•
and CH2O are major products in the experiment, but HO2• was
proposed to be produced by the secondary reaction 8 previously,
which can remove an additional CH3O2• to produce CH3OOH.7,15
According to the present study, it is possible to be produced
directly from the self-reaction of CH3O2•.
CH3OH + CH2O + O21. The other path from IM1 leads to
formaldehyde, methanol, and the spin-singlet oxygen through
the six-membered ring transition state TS3, which is just the
Russel TS as indicated in Scheme 1, with an imaginary frequency
of 493 cm
1. According to the structure of TS3, one of the
terminal hydrogen atom (H6) shifts to the terminal oxygen atom
(O1) of the other methoxyl radical with the O3
O4 and O1
O2
bonds broken simultaneously to produce a spin-singlet oxygen
molecule. The relative energy at 0 K of TS3 is 9.8 kcal mol
1
dx.doi.org/10.1021/jp2048508 |J. Phys. Chem. A 2011, 115, 13534–13541
The Journal of Physical Chemistry A
higher than that of TS2 and the relative free energy at 298 K is 9
kcal mol
1 higher than that of TS2. These two channels are thus
both significant at low temperature, which agrees with the experi-
mental detected major products.
CH3OCH3 + O3. CH3OCH3 + O3 can be produced from the
intermediate IM2 through TS4 with an imaginary frequency of
772 cm
1. The relative energy of TS4 is 77.9 kcal mol
1, and
ΔG298 is 90.1 kcal mol
1 above the initial reactants, which
implies that this channel is unimportant among all channels.
CH3OOOH + CH2O. The intermediate IM3 in the entrance
channel along the reaction path can lead to the products
CH3OOOH + CH 2O through TS5 with an imaginary vibra-
tional frequency of 662 cm
1. As shown in Figure 1, the
C5
O4
O3
O2 dihedral angle changes from 87.3° in IM3 to
173.2° in TS5, due to the rotation of one CH3O fragment. At the
same time, one of the hydrogens (H7) is migrating from C5 to the
nearest O3 to produce CH2O, while the O3
O4 bond of CH3O2
is stretched to be broken. The energy barrier for TS5 is 20.1 kcal
mol
1, the relative enthalpy barrier is 20.0 kcal mol
1, and the
free energy barrier is 19.7 kcal mol
1 above IM3 at 298 K. The
energy barrier to produce CH3OOOH and CH2O is one of
the lowest among all channels according to the present study.
A similar channel with a relatively low energy barrier to form
CH2O + HOOOH in the reaction of HO2• + CH3O2• was also
found.42 However, this channel is not even presented in the
previous experimental study. This may be due to the fact that
CH3OOOH is an active species especially in the presence of
water and may further decompose to produce CH3OH + O2.53,54
This channel thus merits further experimental investigation.
CH4 + CH2O + O3. The other reaction from IM3 occurs via the
seven-center transition state TS6, with an imaginary frequency of
949 cm
1. The relative energy of TS6, i.e., 84.1 kcal mol
1 above
IM3, is much higher than that of TS5, which makes this reaction
pathway significantly less activated than other channels.
2CH3O• + O23. The spin-singlet chain-structure tetroxides can
also evolve by complete dissociation to produce 2CH3O• and
O23. IM1 and IM2 with C2 symmetry are possible reactants from
which CH3O• and O2 are produced since this channel involves
two concerted O
O cleavages synchronously. However, Ghigo
et al.6 proposed that the intermediate first breaks up into the
weakly bound complex [2(CH3O•)-O2]1 before leading to
2CH3O• + O23. However, the distances between O2 and the
remaining atoms in this weakly bound complex are larger than
3.5 Å.6 The structure of such complex could be highly sensitive to
the adopted method. In the present work, we propose this
channel to be barrierless through a concerted dissociation from
IM1 or IM2, which is consistent with some other previous
studies.35,36 The two spin-doublet CH3O will give rise to a spin-
triplet state and then couple with the spin-triplet ground state
dioxygen so that the total spin of the products is zero (spin-
singlet). The computed energy change for this reaction is 0.4
kcal mol
1, whereas the free energy change is favorable by
9.2
kcal mol
1 at 298 K and
30.9 kcal mol
1 at 900 K, respec-
tively, with respect to the two initial methyl peroxyl radicals.
The fact that the free energy change for this channel decreases
as the temperature rises (ΔG298 =
9.5 kcal mol
1 and ΔG900 =

58.3 kcal mol
1, with respect to IM1; ΔG298 =
11.5 kcal
mol
1 and ΔG900 =
59.6 kcal mol
1, with respect to IM2)
demonstrates that it is easier for the intermediates to go forward
to the products at high temperature. This is consistent with the
experimental finding that the branching ratio for this channel
increases at higher temperature.22
13540
ARTICLE
3.2.2. Spin-Triplet Product Channels. CH3OOCH3 + O23. The
transition state TS7 is found to directly connect the products
CH3OOCH3 + O23 with the reactants 2(CH3O2•). This was
verified by IRC scanning calculations. This reaction is a typical
bimolecular nucleophilic substitution reaction and involves
C9
O1 rupture (r19 = 1.9 Å) and O3
C9 recombination (r39 =
2.1 Å) on the spin-triplet surface. The spin-triplet CH3OOCH3 + O2
is energetically more stable than their spin-singlet counterpart,
but the energy barrier for the spin-triplet channel is 31.5 kcal
mol
1 higher than that of the corresponding spin-singlet channel
producing CH3OOCH3 + O2. This indicates that the production
of CH3OOCH3 and O2 will occur mainly via spin-singlet PES.
CH3OOH + •CH2O2•. A direct hydrogen abstraction path
leading to CH3OOH and the spin-triplet CH2O2 molecule from
2(CH3O2•) through TS8 is found. The bond length of the
forming O2
H7 in TS8 is 1.09 Å, while the breaking of the
C5
H7 bond is stretched to 1.48 Å, which is 0.40 Å longer than
that in CH3O2•. For the transition state to correlate with
reactants and products, the unpaired electrons in TS8 should
mainly locate on the O3 and C5 atoms, which is in accord with
Mulliken's spin density analysis (see the Supporting Information).
The free energy of TS8 is higher than that of the initial reactants
by 34.5 kcal mol
1 at 298 K and 53.8 kcal mol
1 at 900 K. Furthe-
rmore, this abstraction reaction is endothermic by 19.5 kcal mol
1
and is thus difficult to occur thermodynamically, which is con-
sistent with Lightfoot’s results7 and may be able to explain the
low CH3OOH yields.15
From our calculations, the self-reaction of CH3O2• mainly
occurs on spin-singlet PES. Among all channels investigated in
present work, channels producing CH3O• + CH2O + HO2•,
CH3OH + CH2O + O21, CH3OOOH + CH2O, and 2CH3O• +
O23 are found to be the most important. On the basis of the
change of free energies with temperature as shown in Table 4, the
importance of the first three major channels is expected to
decline, while that of the last dissociation channel will increase
as temperature rises, which is qualitatively consistent with
previous experimental results.7,13,15,22
4. CONCLUSIONS
In the present work, PESs for the self-reaction of CH3O2• have
been explored in detail at the CCSD(T)/cc-pVTZ//B3LYP/6-
311++G(2df,2p) level of theory. Reactions both on spin-triplet
and on spin-singlet potential energy surfaces are investigated.
Reaction mechanism on the spin-triplet PES is rather simple and
only involves hydrogen abstraction and bimolecular nucleophilic
substitution. Furthermore, reactions on the spin-triplet surface
are unimportant due to high free energy barriers. However, the
reaction mechanism on the spin-singlet surface is shown to be much
more complicated. On the spin-singlet surface, CH3OOCH3 +
O21 can be produced directly from 2(CH3O2•), while in other
channels three spin-singlet chain-structure intermediates are first
formed and subsequently dissociated to produce different pro-
ducts. The barrierless channel to produce 2CH3O• + O23 can
occur easily, in accord with previous studies6 or experimental
results.22 The results show channels IM2 f TS2 f CH3O• +
CH2O + HO2• and IM2 f TS3 f CH3OH + CH2O + O21
could be important at low temperature. Besides these channels,
the pathway of IM3 f TS5 f CH3OOOH + CH2O, which is
not presented in any previous experimental study, has the lowest
barrier in the overall self-reaction of CH3O2• according to our
calculations and deserves further experimental investigation.
dx.doi.org/10.1021/jp2048508 |J. Phys. Chem. A 2011, 115, 13534–13541
The Journal of Physical Chemistry A
’ ASSOCIATED CONTENT
bS Supporting Information. Calculated geometries (in the
Z-matrix), vibrational frequencies, rotational constants, and averages
of S2 for some molecules and transition states. This material is
available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: wangf44@yahoo.com.cn.
’ ACKNOWLEDGMENT
We thank the National Nature Science Foundation of China
(grant no. 91016002) for financial support.
’ REFERENCES
(1) Seinfield, J. H.; Barker, J. R. Progress and Problems in Atmospheric
Chemistry, Advanced Series in Physical Chemistry; World Scientific:
Singapore, 1995.
(2) Lightfoot, P. D.; Cox, R. A.; Crowley, J. N.; Destriau, M.;
Hayman, G. D.; Jenkin, M. E.; Moorgat, G. K.; Zabel, F. Atmos. Chem.
1964, 26A, 246.
(3) Aplincourt, P.; Ruiz-Lopez, M. F.; Assfeld, X.; Bohr., F. J. Comput.
Chem. 1999, 20, 1039.
(4) Wallington, T. J.; Dagaut, P.; Kurylo, M. J. Chem. Phys. Rev. 1992,
92, 667.
(5) (a) Finlayson-Pitta, B. J.; Pitta, J. N. Jr. Atmospheric Chemistry;
Wiley: New York, 1986. (b) Wayne, R. P. Chemistry of Atmospheres;
Oxford University Press: Oxford, U.K., 1985.
(6) Ghigo, G.; Maranzana, A.; Tonachini, G. J. Chem. Phys. 2003,
118, 10575.
(7) Lightfoot, P. D.; Lesclaux, R.; Veyret, B. J. Phys. Chem. 1990,
94, 700.
(8) Westbrook, C. K.; Pitz, W. J. Combust. Flame 2009, 156, 181.
(9) Cox, R. A.; Tyndall, G. S. Chem. Phys. Lett. 1979, 65, 357.
(10) Dagaut, P.; Wallington, T. J.; Kurylo, M. J. J. Phys. Chem. 1988,
92, 3833.
(11) Parkes, D. A. In Proceedings of the 15th International Symposium
on Combustion; The Combustion Institute: Pittsburgh, PA, 1975.
(12) Benson, S. W.; Nangia, P. S. Acc. Chem. Res. 1979, 7, 223.
(13) Niki, H.; Maker, P. D.; Savage, C. M.; Breitenbach, L. P. J. Phys.
Chem. 1981, 85, 877.
(14) Niki, H.; Maker, P. D.; Savage, C. M.; Breitenbach, L. P. J. Phys.
Chem. 1982, 86, 3825.
(15) Tyndall, G. S.; Wallington, T. J.; Ball, J. C. J. Phys. Chem. A 1998,
102, 2547.
(16) Wallington, T. J.; Japar, S. M. Chem. Phys. Lett. 1990, 167, 513.
(17) Wallington, T. J. J. Chem. Soc., Faraday Trans. 1991, 87, 2379.
(18) Wallington, T. J.; Hurley, M. D. Chem. Phys. Lett. 1992, 193, 84.
(19) Veyret, B.; Lesclaux, R.; Rayez, M. T.; Rayez, J. C.; Cox, R. A.;
Moortgat, G. K. J. Phys. Chem. 1989, 93, 2368.
(20) Kan, C. S.; Calvert, J. G.; Shaw, J. H. J. Phys. Chem. 1980,
84, 3411.
(21) Cox, R. A.; Tyndall, G. S. J. Chem. Soc., Faraday Trans. 2 1980,
76, 153.
(22) Tyndall, G. S.; Cox, R. A.; Granier, C.; Lesclaux, R.; Moortgat,
G. K.; Pilling, M. J.; Ravishankara, A. R.; Wallington, T. J. J. Geophys. Res.
2001, 106, 12157.
(23) Noell, A. C.; Alconcel, L. S.; Robichaud, D. J.; Okumura, M.;
Sander, S. P. J. Phys. Chem. A 2010, 114, 6983.
(24) Russell, G. A. J. Am. Chem. Soc. 1957, 79, 3871.
(25) Roothan, C. C. J. Rev. Mod. Phys. 1951, 23, 69.
(26) Pople, J. A.; Nesbet, R. K. J. Chem. Phys. 1954, 22, 571.
(27) McWeeny, R.; Dierksen, G. J. Chem. Phys. 1968, 49, 4852.
13541
ARTICLE
(28) Saebo, S.; Almlof, J. Chem. Phys. Lett. 1989, 154, 83.
(29) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and
Molecules; Oxford University Press: Oxford, U.K., 1989.
(30) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(31) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. 1988, B41, 785.
(32) Curtiss, L. A.; Raghavachari, K.; Redfern, P. C.; Rassolov, V.;
Pople, J. A. J. Chem. Phys. 1998, 109, 7764.
(33) Bohr, F.; Henon, E.; Garcıa, I.; Castro, M. Int. J. Quantum Chem.
1999, 75, 671.
(34) da Silva, G.; Bozzelli, J. W. J. Phys. Chem. A 2007, 111, 12026.
(35) Feria, L.; Gonzalez, C.; Castro, M. Int. J. Quantum Chem. 2004,
99, 605.
(36) Jalbout, A. F.; Zhou, Z.; Li, X. H. Chem. Phys. Lett. 2006, 420, 215.
(37) Feria, L.; Gonzalez, C.; Castro, M. Int. J. Quantum Chem. 2004,
96, 380.
(38) Dibble, T. S. Atmos. Environ. 2008, 42, 5837.
(39) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M.
Chem. Phys. Lett. 1989, 157, 479.
(40) Bartlett, R. J.; Musial, M. Rev. Mod. Phys. 2007, 79, 291.
(41) Hasson, A. S.; Kuwata, K. T.; Arroyo, M. C.; Petersen, E. B.
J. Photochem. Photobiol., A 2005, 176, 218.
(42) Hua, H.; Wang, B. S. J. Phys. Chem. A 2005, 109, 451.
(43) Zhu, R. S.; Lin, M. C. PhysChemComm 2001, 4, 106.
(44) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys.
1980, 72, 650.
(45) Baboul, A. G.; Curtiss, L. A.; Redfern, P. C.; Raghavachari, K.
J. Chem. Phys. 1999, 110, 7650.
(46) Halls, M. D.; Velkovski, J.; Schlegel, H. B. Theor. Chem. Acc.
2001, 105, 413.
(47) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154.
(48) Dunning, T. H. J. Chem. Phys. 1989, 90, 1007.
(49) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson,
G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;
Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.;
Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.;
Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski,
V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui,
Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.;
Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill,
P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.
Gaussian 03, revision B.05; Gaussian, Inc.: Wallingford, CT, 2003.
(50) CFOUR, a quantum chemical program package written by-
Stanton, J. F.; Gauss, J.; Harding, M. E.; Szalay, P. G.; with contributions
from Auer, A. A.; Bartlett, R. J.; Benedikt, U.; Berger, C.; Bernholdt,
D. E.; Bomble, Y. J.; Cheng, L.; Christiansen, O.; Heckert, M.; Heun, O.;
Huber, C.; Jagau, T.-C.; Jonsson, D.; Jus
elius, J., Klein, K.; Lauderdale,
W. J.; Matthews, D. A.; Metzroth, T.; M€uck, L. A.; O’Neill, D. P., Price,
D. R.; Prochnow, E.; Puzzarini, C.; Ruud, K.; Schiffmann, F.; Schwalbach,
W.; Stopkowicz, S.; Tajti, A.; V
azquez, J., Wang, F.; Watts, J. D.; and the
integral packages MOLECULE (Alml€of, J.; Taylor, P. R.), PROPS (Taylor,
P. R.), ABACUS (Helgaker, T.; Aa. Jensen, H. J.; Jørgensen, P.; Olsen, J. ),
and ECP routines by Mitin, A. V.; van W€ullen, C. For the current version,
see http://www.cfour.de.
(51) Truong, T. N.; Nayak, M.; Huynh, H. H.; et al. J. Chem. Inf.
Model. 2006, 46, 971–984.
(52) Huber, K. P.; Herzberg, G. Molecular Spectra and Molecular Structure.
IV. Constants of Diatomic Molecules; Van Nostrand: New York, 1979; p 498.
(53) Plesnicar, B.; Cerkovnik, J.; Tekavec, T.; Joze, K. Chem.—Eur. J.
2000, 6, 809.
(54) Plesnicar, B. Acta Chim. Slov. 2005, 52, 1.
dx.doi.org/10.1021/jp2048508 |J. Phys. Chem. A 2011, 115, 13534–13541