Investigation of the radical product channel of the CH3OCH 2O2 + HO2 Reaction in the Gas Phase

Article abstract of DOI:10.1021/jp908158w

The reaction of CH₃OCH₂O₂ with HO ₂ has been investigated at 296 K and 700 Torr using long path FTIR spectroscopy, during photolysis of Cl₂/CH₃OCH ₃/CH₃OH/air mixtures. The branching ratio for the reaction channel forming CH₃OCH₂O, OH, and O₂ has been determined from experiments in which OH radicals were scavenged by addition of benzene to the system, with subsequent formation of phenol used as the primary diagnostic for OH radical formation. The dependence of the phenol yield on the initial peroxy radical precursor reagent concentration ratio, [CH ₃OH]₀/[CH₃OCH₃]₀, is consistent with prompt OH formation resulting mainly from the reaction of CH₃OCH₂O₂ with HO₂, such that the inferred prompt yield of OH is well-correlated with that of CH ₃OCH₂OOH, a well-established product of the CH ₃OCH₂O₂ + HO₂ reaction. The system was fully characterized by simulation, using a detailed chemical mechanism which included other established sources of OH in the system. This allowed a branching ratio of k_2c/k₂ = 0.19 ± 0.08 to be determined. The results therefore provide strong indirect evidence for the participation of the radical-forming channel of the title reaction.

Full text of DOI:10.1021/jp908158w

408
J. Phys. Chem. A 2010, 114, 408–416
Investigation of the Radical Product Channel of the CH3OCH2O2 + HO2 Reaction in the
Gas Phase
M. E. Jenkin,*,† M. D. Hurley,‡ and T. J. Wallington‡
Atmospheric Chemistry SerVices, Okehampton, DeVon, EX20 1FB, U.K., and Research and AdVanced Engineering,
Ford Motor Company, SRL-3083, PO Box 2053, Dearborn, Michigan 48121-2053
ReceiVed: August 24, 2009; ReVised Manuscript ReceiVed: October 16, 2009
The reaction of CH3OCH2O2 with HO2 has been investigated at 296 K and 700 Torr using long path FTIR
spectroscopy, during photolysis of Cl2/CH3OCH3/CH3OH/air mixtures. The branching ratio for the reaction
channel forming CH3OCH2O, OH, and O2 has been determined from experiments in which OH radicals were
scavenged by addition of benzene to the system, with subsequent formation of phenol used as the primary
diagnostic for OH radical formation. The dependence of the phenol yield on the initial peroxy radical precursor
reagent concentration ratio, [CH3OH]0/[CH3OCH3]0, is consistent with prompt OH formation resulting mainly
from the reaction of CH3OCH2O2 with HO2, such that the inferred prompt yield of OH is well-correlated
with that of CH3OCH2OOH, a well-established product of the CH3OCH2O2 + HO2 reaction. The system was
fully characterized by simulation, using a detailed chemical mechanism which included other established
sources of OH in the system. This allowed a branching ratio of k2c/k2 ) 0.19 ( 0.08 to be determined. The
results therefore provide strong indirect evidence for the participation of the radical-forming channel of the
title reaction.
1. Introduction
reactions of HO2 with a number of oxygenated RO2 radicals
directly using pulsed laser photolysis with laser-induced fluo-
rescence detection of OH. They were also able to establish that
the radical-propagating channel was insignificant for a number
of other oxygenated RO2 radicals, including several ꢀ-hydroxy
peroxy radicals formed significantly in the atmosphere from the
reactions of OH with alkenes and alcohols.
The studies to date have therefore established the existence
of radical-propagating channels for the reactions of HO2 with
oxygenated peroxy radicals containing acyl, R-carbonyl, and
R-hydroxy functionalities. However, the importance of such
channels has not been considered for peroxy radicals containing
an alkoxy group adjacent to the peroxy radical center. The
simplest R-alkoxy alkyl peroxy radical (which are derived from
the oxidation of ethers in general) is the methoxymethyl peroxy
radical (CH3OCH3O2). It is formed in the atmosphere from the
OH radical initiated oxidation of dimethyl ether, and is also
predicted to be generated from the breakdown of a number of
complex oxygenated solvents, in particular glycol ethers such
as 2-methoxyethanol and 1-methoxypropan-2-ol.17,18
The reactions of organic peroxy radicals (RO2) with the
hydroperoxy radical (HO2) have long been considered as chain
terminating reactions that make an important contribution to
controlling the concentrations of free radicals under atmospheric
conditions, particularly when the availability of nitrogen oxides
(NOX) is limited.1-4 Whereas radical termination via the near-
exclusive formation of organic hydroperoxide products (ROOH)
and O2 is well-established for simple alkyl peroxy radicals such
as CH3O2 and C2H5O2,5-10 a number of more recent studies have
demonstrated that selected oxygenated RO2 radicals possess
significant radical-propagating channels for their reactions with
HO2, thereby lessening their perceived impact as chain terminat-
ing processes:8,11-13
RO2 + HO2 f RO + OH + O2
(1)
This may therefore be a contributory factor in systematic under-
simulation of HOx radical concentrations observed in the
troposphere, as reported in a number of studies.14-16
The products of the reaction of CH3OCH2O2 with HO2 at
295 K and 700 Torr have previously been studied by the present
authors (Wallington et al.19), using long path FTIR spectroscopy.
Methoxymethyl hydroperoxide (CH3OCH2OOH) and methyl
formate (CH3OCHO) were observed as major primary products,
the formation of which was attributed to the following radical-
terminating reaction channels:
Laboratory studies have shown that channel 1 is particularly
significant for the reaction of the acetyl peroxy radical
(CH3C(O)O2) with HO2, accounting for about 40-50% of the
overall reaction,8,11,13 with more minor contributions in the region
of 15-20% also being reported for the reactions of HO2 with
HOCH2O2,11 CH3C(O)CH2O2,12,13 and C6H5C(O)O2;13 and for
the reaction of HO2 with the suite of three isomeric RO2 radicals
formed from the reaction of Cl with C2H5C(O)CH3.13 Of these
studies, the most direct evidence has been reported by Dillon
and Crowley,13 who measured the formation of OH from the
CH3OCH2O2 + HO2 f CH3OCH2OOH + O2
f CH3OCHO + H2O + O2
(2a)
(2b)
* Author to whom correspondence should be addressed. E-mail:
atmos.chem@btinternet.com.
† Atmospheric Chemistry Services.
Within experimental uncertainties, the yields of CH3OCH2OOH
and CH3OCHO provided closure and were used to assign
‡ Ford Motor Company.
10.1021/jp908158w  2010 American Chemical Society
Published on Web 11/04/2009

Radical Product Channel of CH3OCH2O2 + HO2
J. Phys. Chem. A, Vol. 114, No. 1, 2010 409
branching ratios of k2a/k2 ) 0.53 ( 0.08 and k2b/k2 ) 0.40 (
0.04. However, formation of CH3OCHO could also result from
a contribution from a radical-propagating channel (2c),
CH3OCH2O2 + HO2 f CH3OCH2O + OH + O2
(2c)
followed by the established fates of reaction with O2 or thermal
decomposition for CH3OCH2O:20
CH3OCH2O + O2 f CH3OCHO + HO2
CH3OCH2O (+M) f CH3OCHO + H (+M)
H + O2 (+M) f HO2 (+M)
(3)
(4)
(5)
Figure 1. Loss of CH3OCH3 versus CH3OH following exposure to Cl
atoms in 700 Torr of N2 (squares) and air (circles). [CH3OH]0 and
[CH3OCH3]0 are the initial concentrations; [CH3OH]t and [CH3OCH3]t
are concentrations at time ) t. [CH3OH]0 ≈ 300 mTorr, [CH3OCH3]0
≈ 20 mTorr, and [Cl2]0 ≈ 100 mTorr. Each data set includes combined
data from two experiments with the same conditions.
Based on the reported yield of CH3OCHO,19 channel 2c could
therefore potentially account for up to about 40% of the reaction.
In the present study, the products of reaction 2 have been
reinvestigated at 296 K and 700 Torr using long path FTIR
spectroscopy, during photolysis of Cl2/CH3OCH3/CH3OH/air
mixtures, the same chemical system employed previously.19
Similarly to our recent investigations of the reactions of
CH3C(O)O2, HOCH2O2, and CH3C(O)CH2O2 with HO2,11,12 the
formation of OH radicals in the system has been investigated
by addition of benzene to scavenge variable proportions of OH,
with the subsequent formation of phenol used as a marker for
OH radical formation. Although this method provides an indirect
measurement of OH formation, it is noted that our yields
molecule-1 s-1 23). The formation of phenol from the OH-
initiated oxidation of benzene (reactions 11 and 12) was used
as the primary diagnostic for OH radical formation, the reported
yield of phenol being 53.1 ( 6.6%:24
OH + C6H6 (+M) f HOC6H6 (+M)
HOC6H6 + O2 f C6H5OH + HO2
(11)
(12)
reported for the reactions of CH3C(O)O2 and CH3C(O)CH2O2
11,12
with HO2
are in good agreement with those derived from
the direct OH measurements of Dillon and Crowley.13
2. Experimental Section
Infrared spectra of reagents and products were derived from 32
coadded interferograms with a spectral resolution of 0.25 cm-1,
and an analytical path length of 27.1 m. All experiments were
carried out at 296 ((2) K and 700 Torr total pressure of air.
The organic reagents, CH3OCH3 (g99.9%), CH3OH (g99.8%),
and benzene (g99.9%) were obtained from Aldrich Chemical
Co. Cl2 and air were obtained from Michigan Airgas at research
grade purity. Reference spectra were obtained by expanding
calibrated volumes into the chamber. Unless otherwise specified,
all quoted errors are two standard deviations.
All experiments were performed in the Ford 140 L Pyrex
chamber, interfaced with a Mattson Sirius 100 FTIR spectrom-
eter, which is described in detail elsewhere.21 The chamber is
equipped with 22 fluorescent blacklamps (GE F40BLB), emit-
ting near UV radiation in the range 300-450 nm. Radical
generation was initiated by the photolysis of Cl2, with
CH3OCH2O2 and HO2 radicals formed from the subsequent
reaction of Cl atoms with CH3OCH3 and CH3OH, respectively,
by the following well-established mechanisms:
3. Results and Discussion
Cl2 + hν f Cl + Cl
(6)
(7)
3.1. Rate Coefficient for the Reaction of Cl Atoms with
CH3OCH3. The relative formation rates of CH3OCH2O2 and
HO2 in the experiments described below is determined by the
relative rates of reactions 7 and 9. Whereas the value of k9 is
very well determined at temperatures close to 298 K (k9 ) 5.5
× 10-11 cm3 molecule-1 s-1),23 the six determinations of k7 show
some scatter, covering a range from 1.3 × 10-10 to 2.05 × 10-10
cm3 molecule-1 s-1 (see Notario et al.25 and references therein).
Initial experiments were therefore performed in which the loss
of CH3OCH3 was measured relative to that of CH3OH during
UV irradiation of Cl2/CH3OCH3/CH3OH/N2 mixtures at 700
Torr (see Figure 1). Linear regression analysis of the data gives
a value of k7/k9 ) 3.28 ( 0.07, this being the key parameter
which quantifies the relative formation of CH3OCH2O2 and HO2
in the experiments below. Using the IUPAC recommended value
of k9 given above leads to an absolute value of k7 ) (1.80 (
0.22) × 10-10 cm3 molecule-1 s-1, where the quoted error limits
Cl + CH3OCH3 f CH3OCH2 + HCl
CH3OCH2 + O2 (+M) f CH3OCH2O2 (+M) (8a)
Cl + CH3OH f CH2OH + HCl
CH2OH + O2 f HCHO + HO2
(9)
(10)
Benzene was added to the reaction mixtures to scavenge variable
proportions of OH radicals generated in the system. Benzene
was selected because, unlike CH3OCH3 and CH3OH, it reacts
slowly with Cl (k ) 1.3 × 10-16 cm3 molecule-1 s-1 22), but
comparatively rapidly with OH (k11 ) 1.2 × 10-12 cm3

410 J. Phys. Chem. A, Vol. 114, No. 1, 2010
Jenkin et al.
include an additional 10% to account for uncertainty associated
with the reference rate coefficient.
The relative loss of CH3OCH3 and CH3OH was also
determined during UV irradiation of Cl2/CH3OCH3/CH3OH/air
mixtures at 700 Torr, as also shown in Figure 1. The measured
relative loss, 3.33 ( 0.08, is indistinguishable from that
determined in N2 diluent, indicating that any secondary chem-
istry leading to removal of the reagents when O2 is present (i.e.,
the formation of OH radicals as described below) does not distort
their relative loss significantly. This is as expected, because the
relative rate coefficient for the reactions of CH3OCH3 and
CH3OH with OH radicals, is very similar in magnitude to the
relative rate coefficient for the Cl atom reactions determined
above, based on the IUPAC recommended values of k13 ) 2.8
× 10-12 cm3 molecule-1 s-1 and k14 ) 9.0 × 10-13 cm3
molecule-1 s-1 23
.
Figure 2. Loss of phenol versus CH3OCH3 following exposure to Cl
atoms. [phenol]0 and [CH3OCH3]0 are the initial concentrations;
[phenol]t and [CH3OCH3]t are concentrations at time ) t. Triangles
are Cl2/CH3OCH3/phenol experiment in 700 Torr N2 ([phenol]0 ) 2.9
mTorr; [CH3OCH3]0 ) 20.0 mTorr). Squares are Cl2/CH3OCH3/phenol
experiment in 700 Torr air ([phenol]0 ) 2.9 mTorr; [CH3OCH3]0 )
20.0 mTorr). Circles are Cl2/CH3OCH3/CH3OH/phenol experiment in
700 Torr air ([phenol]0 ) 3.1 mTorr; [CH3OCH3]0 ) 20.6 mTorr;
[CH3OH]0 ) 307 mTorr). For clarity, the first two data sets been offset
vertically by 0.4 and 0.2, respectively.
OH + CH3OCH3 f CH2OCH2 + H2O
OH + CH3OH f CH2OH + H2O
f CH3O + H2O
(13)
(14a)
(14b)
For completeness, the relative loss of CH3OCH3 and CH3OH
was also analyzed for the set of Cl2/CH3OCH3/CH3OH/benzene/
air mixtures used for the analysis described in section 3.3 below.
No systematic dependence of the relative loss coefficient was
apparent for variation of the initial reagent concentration ratio
[CH3OH]0/[CH3OCH3]0 over the range 2-15, or with the
benzene concentration varied over the range from zero up to
about 1 Torr (corresponding to up to about 90% scavenging of
OH radicals).
3.2. Rate Coefficient for the Reaction of Cl Atoms with
Phenol. Phenol is known to react rapidly with Cl atoms to
produce phenoxy radicals (reaction 15), and it is important to
quantify the associated removal of phenol under the experi-
mental conditions employed. In our previous studies,11 an
apparent lowering of k15 was observed in experiments performed
in the presence of O2, and this was tentatively attributed to the
partial regeneration of phenol from the reaction of phenoxy
radicals with species containing labile hydrogen atoms, for
example HO2 (reaction 16):
ized and described previously11). The results of the Cl2/phenol/
CH3OCH3 experiment in N2 diluent provide a measure of the
rate constant ratio k15/k7 ) 1.07 ( 0.05. Using the value of k7
determined above, this leads to an absolute value of k15 ) (1.93
( 0.25) × 10-10 cm3 molecule-1 s-1, in excellent agreement
with the values of k15 ) (1.92 ( 0.17) × 10-10 cm3 molecule-1
s-1 and k15 ) (1.93 ( 0.36) × 10-10 cm3 molecule-1 s-1 reported
previously.11,26 The results of a corresponding experiment
performed in air diluent indicate a lowering of the measured
rate coefficient ratio to a value of 0.814 ( 0.018 (see Figure
2), consistent with an apparent rate coefficient k15obs ) (1.47 (
0.18) × 10-10 cm3 molecule-1 s-1. As indicated above, the
apparent decrease in the reactivity of phenol possibly reflects
its partial re-formation by reactions with hydrogen containing
species (e.g., HO2 radicals) formed during the oxidation
reactions that occur in air diluent. The abundance of HO2 was
therefore increased by addition of CH3OH to the reaction
mixture such that [CH3OH]0/[CH3OCH3]0 was typical of the high
end of the range employed in the experiments reported below.
This led to a further lowering of the measured rate coefficient
ratio to a value of 0.701 ( 0.024 (see Figure 2), consistent with
an apparent rate coefficient k15obs ) (1.26 ( 0.16) × 10-10 cm3
molecule-1 s-1, providing some support for phenol regeneration
via reaction 16. These values of k15obs, obtained in air with
concentrations of CH3OCH3 and CH3OH characteristic of those
used in the subsequent experiments, were therefore used in the
studies described below, either to correct the phenol yields for
its net loss via reaction 15 or to calculate the reaction rate in
simulations.
Cl + C6H5OH f HCl + C6H5O
C6H5O + HO2 f C6H5OH + O2
(15)
(16)
To characterize this for the current system, a series of experi-
ments was performed in which the loss of phenol was measured
relative to that of CH3OCH3 during UV irradiation of Cl2/phenol/
CH3OCH3 and Cl2/phenol/CH3OCH3/CH3OH mixtures (see
Figure 2). The employed initial concentrations of Cl2, CH3OCH3,
and CH3OH were typical of those in the experiments described
below (section 3.3), and the initial concentration of phenol was
comparable with the final concentrations generated in the those
experiments when about 1 Torr of benzene was present. The
experiments were performed in 700 Torr of either N2 or air.
Figure 2 shows a plot of the decay of phenol versus CH3OCH3
observed in the photolysis experiments (N.B., corrections of
up to about 20% have been applied to the phenol data to account
for its thermal heterogeneous loss in the chamber, as character-
3.3. Investigation of the Cl2/CH3OCH3/CH3OH/Benzene/
Air System. 3.3.1. Analysis and Interpretation of Phenol
Formation. The reaction of CH3OCH2O2 with HO2 (reaction
2) was investigated in a series of Cl2/CH3OCH3/CH3OH/
benzene/air UV photolysis experiments in which the ratio of
the peroxy radical precursor concentration, [CH3OH]0/
[CH3OCH3]0 was varied from a value of zero up to about 15.
This leads to corresponding relative primary production rates

Radical Product Channel of CH3OCH2O2 + HO2
J. Phys. Chem. A, Vol. 114, No. 1, 2010 411
Figure 4. Yields of CH3OCHO, CH3OCH2OOH, and phenol relative
to CH3OCH3 lost during the photolysis of Cl2/CH3OCH3/benzene/air
mixtures. Displayed data are from the repeat experiments C (closed
symbols) and M (open symbols) in the absence of CH3OH and with
about 1 Torr of benzene present. Product yields have been corrected
for losses in the system (e0.3%, e14%, and e30% for CH3OCHO,
CH3OCH2OOH, and phenol, respectively). Lines are linear regressions
of the combined data, consistent with respective yields of 0.62, 0.32,
and 0.063 relative to CH3OCH3 lost for CH3OCHO, CH3OCH2OOH,
and phenol.
Figure 3. Yields of CH3OCHO, CH3OCH2OOH, and phenol relative
to CH3OCH3 lost during the photolysis of Cl2/CH3OCH3/CH3OH/
benzene/air mixtures. (For clarity CH3OCH2OOH data have been offset
vertically by 1 mTorr.) Displayed data are from repeat experiments A
(closed symbols) and L (open symbols) with [CH3OH]0/[CH3OCH3]0
≈ 15 and about 1 Torr of benzene present. Product yields have been
corrected for losses in the system (e0.2%, e8%, and e20% for
CH3OCHO, CH3OCH2OOH, and phenol, respectively). Lines for
CH3OCHO and CH3OCH2OOH are linear regressions of the combined
data, consistent with respective yields of 0.42 and 0.49 relative to
CH3OCH3 lost. The line for phenol is a third-order polynomial fit to
the combined data: y ) 0.0787x + 0.0228x2 - 0.0010x3, indicative of
primary and delayed formation in the system.
TABLE 1: Summary of Experimental Conditions in
Cl2/CH3OCH3/CH3OH/Benzene/Air Experiments (Section
3.3)
of HO2 and CH3OCH2O2, which vary over the approximate
range 0-5, and an associated variation in the relative steady
state concentrations of the peroxy radicals. At the high end of
the [CH3OH]0/[CH3OCH3]0 range, CH3OCH2O2 is lost mainly
through its reaction with HO2, with removal via its self-reaction
making an increasing contribution toward the low end of the
[CH3OH]0/[CH3OCH3]0 range, as described further below.
As in our previous study,19 the formation of CH3OCHO and
CH3OCH2OOH was observed in experiments over the entire
[CH3OH]0/[CH3OCH3]0 range. The upper panels of Figures 3
and 4 show examples of the formation of these products vs
CH3OCH3 removed, for respective [CH3OH]0/[CH3OCH3]0
ratios of about 15 and zero. The plots demonstrate that both
CH3OCHO and CH3OCH2OOH are primary products, with the
slopes of the plots allowing quantification of the yields. The
yields obtained in all the experiments listed in Table 1 are shown
in the upper panel of Figure 5, as a function of [CH3OH]0/
[CH3OCH3]0. The magnitudes of these yields, and their depen-
dence on [CH3OH]0/[CH3OCH3]0, are fully consistent with those
reported in our previous study.19 At the high end of the
[Cl2]0 [CH3OCH3]0 [CH3OH]0 [benzene]0 [CH3OH]0/
run (mTorr)
(mTorr)
(mTorr)
(mTorr) [CH3OCH3]0 fOH
A
B
C
D
E
F
G
H
I
100
100
100
101
100
199
100
102
103
100
101
101
100
21.1
20.9
20.3
21.1
21.1
41.8
21.1
20.7
20.9
20.9
21.1
20.9
20.7
298
158
0
1017
1017
1034
1034
1051
1017
98.3
0
50.9
225
14.1
7.6
0
0.79
0.86
0.96
0.93
0.82
0.64
0.26
0
0.15
0.44
0.66
0.78
0.95
44.9
234
634
317
312
317
322
317
322
0
2.1
11.1
15.2
15.0
15.1
15.2
15.4
15.0
15.4
0
J
K
L
M
559
1034
1017
[CH3OH]0/[CH3OCH3]0 range, yields of 0.42 ( 0.06 and 0.51
( 0.10 were obtained for CH3OCHO and CH3OCH2OOH,
respectively, on the basis of all the experiments with [CH3OH]0/
[CH3OCH3]0 ≈ 15, with the error limits including 5% and 10%
uncertainties associated with the quantification of the two
products.19 In the absence of CH3OH (i.e., [CH3OH]0/

412 J. Phys. Chem. A, Vol. 114, No. 1, 2010
Jenkin et al.
Figure 6. Observed yield of phenol vs the calculated fraction of OH
scavenged by reaction with benzene (fOH), in experiments A and F-L
with [CH3OH]0/[CH3OCH3]0 ≈ 15 (see Table 1). Determination of fOH
based on k11 ) 1.2 × 10-12 cm3 molecule-1 s-1, k13 ) 2.8 × 10-12 cm3
molecule-1 s-1, and k14 ) 9.0 × 10-13 cm3 molecule-1 s-1 from the
current IUPAC evaluation.23 Phenol yields determined at 5 mTorr
consumption of CH3OCH3.
reaction of CH3OCH2O2 with HO2) are still generated signifi-
cantly in the absence of CH3OH.
The experiments were carried out with benzene present in
the reaction mixtures at pressures up to about 1 Torr. Owing to
its extremely low reactivity with Cl, benzene at these pressures
does not interfere with the production of CH3OCH2O2 and HO2
under the conditions employed (scavenging e0.002% of Cl
atoms) but is sufficiently reactive to scavenge OH radicals in
competition with other reactions of OH (i.e., primarily its
reactions with CH3OCH3 and CH3OH, reactions 13 and 14),
leading to the formation of phenol by reaction sequences 11
and 12. Accordingly, phenol formation was observed in all
experiments in which benzene was added (over the entire
[CH3OH]0/[CH3OCH3]0 range), with the structured infrared band
near 750 cm-1 allowing its sensitive and specific detection and
quantification.11
Figure 5. Yields of CH3OCHO, CH3OCH2OOH, and prompt OH,
relative to CH3OCH3 lost, as a function of [CH3OH]0/[CH3OCH3]0
during the photolysis of Cl2/CH3OCH3/benzene/air mixtures. Displayed
data for CH3OCHO and CH3OCH2OOH are from all experiments in
Table 1. The prompt OH yields were derived from the initial formation
of phenol in experiments A-E, L, and M with fOH g 0.78 (see text).
The broken lines are arbitrary functions included to aid in visual
inspection of the data trends. The lower panel inset shows a correlation
of the yields of prompt OH and CH3OCH2OOH.
The lower panels of Figures 3 and 4 show the formation of
phenol vs CH3OCH3 removed, in the example experiments
discussed above, with [CH3OH]0/[CH3OCH3]0 ratios of about
15 and 0. In these experiments, approximately 1 Torr of benzene
was present, allowing between 78% and 96% of OH to be
scavenged by reaction 11. At high [CH3OH]0/[CH3OCH3]0
(Figure 3), the curved plot demonstrates that phenol is generated
as a primary product, indicative of prompt formation of OH,
but with this formation supplemented by one or more additional
sources of OH at greater extents of reaction. In the absence of
CH3OH (Figure 4), however, the linearity of the plot suggests
that phenol is formed almost exclusively as a prompt primary
product, suggesting that secondary sources of OH are much less
important under these conditions. These observations are
consistent with the majority of the delayed OH production being
due to the chemistry initiated by the reaction of HO2 with the
major product HCHO (formed from CH3OH oxidation, primarily
by reaction sequences 9 and 10) to form HOCH2O2, which has
been fully characterized and described previously.11 OH forma-
tion associated with this chemistry was found to result from
the radical propagating channel for the reaction of HOCH2O2
with HO2, and also from the reaction of Cl atoms with the
terminating channel product, HOCH2OOH.
[CH3OCH3]0 ) 0), the yield of CH3OCHO increases to 0.62 (
0.06, and that of CH3OCH2OOH decreases to 0.32 ( 0.04, in
excellent agreement with those reported in our previous
studies.19,20 In the absence of CH3OH, there is no primary route
to HO2 formation. However, CH3OCH2O2 radicals are initially
removed by their self-reaction,
CH3OCH2O2 + CH3OCH2O2 f CH3OCH2O + CH3OCH2O + O2
(17a)
f CH3OCHO + CH3OCH2OH + O2
(17b)
with the subsequent reactions of CH3OCH2O (formed with
approximately 70% yield via channel 17a) leading to secondary
HO2 formation (via reaction 3 or reactions 4 and 5). As a result,
reaction with HO2 radicals makes a reduced, but still substantial,
contribution to CH3OCH2O2 removal under steady state pho-
tolysis conditions, even when CH3OH is absent. Consequently,
products such as CH3OCH2OOH (formed exclusively from the
The initial yields of phenol (denoted [phenol]initial), derived
either from the linear component of the polynomial fits or from

Radical Product Channel of CH3OCH2O2 + HO2
J. Phys. Chem. A, Vol. 114, No. 1, 2010 413
TABLE 2: Supplementary Reaction Mechanism and Associated Parameters Used To Simulate the Cl/CH3OCH3/CH3OH/
Benzene/Air Photolysis Experimentsa
branching ratio
rate coefficientc
comment
Initiation Reactions
Cl + CH3OCH3 (+O2)
OH + CH3OCH3 (+O2)
f CH3OCH2O2 + HCl
(7) + (8a)
0.986
0.014
1.8 × 10-10
5.7 × 10-12 exp(-215/T)
d, e
f, e
f HCHO + HCHO + OH + HCl
f CH3OCH2O2 + H2O
(7) + (8b)
(13) + (8a) 0.986
(13) + (8b) 0.014
f HCHO + HCHO + OH + H2O
RO2 + HO2 Reactions
CH3OCH2O2 + HO2
f CH3OCH2OOH + O2
f CH3OCHO + H2O + O2
f CH3OCH2O + OH + O2
(2a)
(2b)
(2c)
varied (see text) 1.0 × 10-11
g
RO2 + RO2 Reactions
2 CH3OCH2O2
f 2 CH3OCH2O + O2
(17a)
0.67
0.33
0.774
0.113
0.113
0.5
2.1 × 10-12
3.46 × 10-12
h
i
f CH3OCHO + CH3OCH2OH + O2 (17b)
f CH3OCH2O + HOCH2O + O2
f CH3OCHO + HOCH2OH + O2
f CH3OCH2OH + HCOOH + O2
f CH3OCH2O + Φ-O + O2
CH3OCH2O2 + HOCH2O2
CH3OCH2O2 + Φ-O2
CH3OCH2O + O2
2.0 × 10-12
j
f CH3OCHO + Φ-OH + O2
0.5
RO Reactions
f CH3OCHO + HO2
(3)
assumed instantaneous
k
Cl + Product Reactions
Cl + CH3OCHO
f products
1.3 × 10-12
5.0 × 10-11
6.1 × 10-11
l, m
n
o
Cl + CH3OCH2OH (+O2)
Cl + CH3OCH2OOH
f CH3OCHO + HO2 + HCl
f CH3OCHO + OH + HCl
(18) + (19)
OH + Product Reactions
OH + CH3OCHO
f products
9.39 × 10-13 exp(-461/T)
4.56 × 10-12
p, m
q
q
OH + CH3OCH2OH (+O2) f CH3OCHO + HO2 + H2O
OH + CH3OCH2OOH
f CH3OCH2O2 + H2O
f CH3OCHO + OH + H2O
0.286
0.714
1.26 × 10-11
a These reactions were used in conjunction with relevant chemistry in Table 3 of Jenkin et al.11 (see note b). The reaction number in the text
is shown in brackets where applicable. b The Cl/CH3OH/benzene/air system was previously characterized within study of the reaction of
CH3C(O)O2 with HO2.11 Relevant reactions were used in the present simulations, with the adjustments to k15obs described in section 3.2. c Units
cm3 molecule-1 s-1
.
d Rate coefficient (k7) determined in the present study, relative to a value of 5.5 × 10-11 cm3 molecule-1 s-1 for k9.
e Branching ratio for the consecutive reaction of CH3OCH2 + O2, based on Sehested et al.27 f Rate coefficient based on current IUPAC
recommendation.23 g Rate coefficient estimated by Jenkin et al.;20 branching ratios were varied on the basis of the results of the present study
and Wallington et al.19 h Rate coefficient and branching ratios from Jenkin et al.20 i Rate coefficient based on geometric mean of self-reaction
rate coefficients for CH3OCH2O2 and HOCH2O2; branching ratios based on arithmetic mean of branching ratios for self-reactions of
CH3OCH2O2 and HOCH2O2. j Rate coefficients and branching ratios estimated on the basis of known reactions of peroxy radicals. “Φ-O2”
represents the complex organic peroxy radical, formed from OH + benzene (see Jenkin et al.11). k RO reactions occur on e30 µs time scale.
Reaction with O2 estimated to dominate over decomposition by H atom ejection, based on Jenkin et al.,20 although ultimate products under
experimental conditions are indistinguishable. l Rate coefficient from Wallington et al.28 m Products not declared, because e0.3% removal of
CH3OCHO occurred under experimental conditions. n Kinetic parameters estimated on the basis of known reactivity of Cl with species
containing-OH groups. o Rate coefficient from Wallington et al.;19 products defined by analogy with chemistry of CH3OOH and CH2OOH,
Vaghjiani and Ravishankara.29 p Rate coefficient based on Wallington et al.30 and Le Calve´ et al.,31 as described by Jenkin and Hayman.32
q Kinetic parameters estimated on the basis of known reactivity of OH with oxygen-bridged species and species containing -OH and -OOH
groups.
a linear regression of the data where appropriate, were used to
determine the prompt yield of OH, as follows:
[CH3OCH3]0 range, a limiting OH yield of about 0.18 was
determined, with the yield falling to about 0.11 in the absence
of CH3OH (i.e., [CH3OH]0/[CH3OCH3]0 ) 0). The relative
dependence of the prompt OH yield on [CH3OH]0/[CH3OCH3]0
is strikingly similar to that observed for CH3OCH2OOH, as
confirmed by the correlation in Figure 5 (lower panel inset).
This provides strong support for the prompt OH formation
resulting mainly from the reaction of CH3OCH2O2 with HO2,
via channel 2c. The OH yield at high [CH3OH]0/[CH3OCH3]0
(i.e., when CH3OCH2O2 reacts mainly with HO2) thus provides
an initial estimate of the branching ratio k2c/k2 ≈ 0.18.
prompt OH yield )
([phenol]initial/∆[CH3OCH3])/(0.531fOH
) (i)
fOH ) k11[benzene]0/(k11[benzene]0 + k13[CH3OCH3]0 +
k14[CH3OH]0) (ii)
Here, fOH is the calculated fraction of OH which reacts with
benzene (reaction 11) under the experimental conditions, and
the factor 0.531 in equation (i) is the yield of phenol from this
reaction (i.e., via reaction sequence 11 and 12). The resultant
OH yields are plotted in the lower panel of Figure 5, as a
function of [CH3OH]0/[CH3OCH3]0, for all experiments with
values of fOH g 0.78. Toward the high end of the [CH3OH]0/
As in our previous studies,11,12 a series of experiments with
a high precursor concentration ratio ([CH3OH]0/[CH3OCH3]0 ≈
15) was also carried out with a series of benzene concentrations
over the range from zero up to about 1 Torr (see Table 1). This
was to confirm that phenol formation was due to the OH-
initiated oxidation of benzene, and not via some alternative
oxidation mechanism not requiring OH formation. The resultant

414 J. Phys. Chem. A, Vol. 114, No. 1, 2010
Jenkin et al.
Figure 7. Formation of phenol vs CH3OCH3 removed in all experiments with approximately 1 Torr of benzene present. The figures in brackets
are the ratio [CH3OH]0/[CH3OCH3]0 (see Table 1). The points are the observed concentrations. The lines are simulations using the mechanism in
Table 2. The bold lines are simulations with the optimized value of k2c/k2 ) 0.19. The upper and lower lines are simulations with k2c/k2 ) 0.4 and
k2c/k2 ) 0.0, respectively (corrections of up to 5% have been applied to the observed phenol concentration data to correct for its thermal heterogeneous
loss, as described in section 3.2).
phenol yields are presented in Figure 6 as a function of fOH
Because quantification of the initial slope, [phenol]initial
.
/
mechanism, which is shown partially in Table 2. The reactions
listed in the table were used to supplement a detailed mechanism
which was optimized previously for the Cl/CH3OH/benzene/
air system, as a component part of our study of the reaction of
CH3C(O)O2 with HO2.11 The supplementary chemistry shown
in Table 2 thus treats the initial stages of CH3OCH3 degradation,
including its reactions with Cl and OH, the reactions of
CH3OCH2O2 with the suite of peroxy radicals formed in the
system (i.e., HO2, CH3OCH2O2, HOCH2O2, and Φ-O2), and the
further reactions of products specific to this system (e.g.,
CH3OCH2OOH).
∆[CH3OCH3], was unreliable in experiments with lower benzene
concentrations present, the phenol yields were based on total
phenol formation at 5 mTorr consumption of CH3OCH3 in each
experiment The results in Figure 6 confirm a linear dependence
of phenol formation on fOH, consistent with its production being
due to the reaction of OH with benzene occurring in competition
with reactions 13 and 14a.
3.3.2. Numerical Simulation of the System. The system was
also characterized by simulation using a detailed explicit

Radical Product Channel of CH3OCH2O2 + HO2
J. Phys. Chem. A, Vol. 114, No. 1, 2010 415
Simulations of all the experiments indicated that they were
well described using a value of k2c/k2 ) 0.19 ( 0.03, with these
uncertainty limits representing statistical errors alone. This was
based on a global fit to the experiments illustrated in Figure 7,
which are those in which high benzene concentrations (≈1 Torr)
were used. The results clearly indicate that the observed phenol
formation, particularly at low reagent conversions, cannot be
accounted for without the participation of the radical-propagating
channel, reaction 2c. The consistent performance of the mech-
anism across the complete range of [CH3OH]0/[CH3OCH3]0 also
confirms that the competition between reactions 2 and 17 for
CH3OCH2O2 is adequately described by the parameters applied.
This provides support for the relative production rates simulated
for CH3OCH2O2 and HO2, and for the rate coefficients applied
to reactions 2 and 17. Whereas k17 was measured directly in
our previous study,20 the applied value of k2 (1 × 10-11 cm3
molecule-1 s-1) is based on that inferred previously from the
relative steady state concentrations of CH3OCH2O2 and HO2
observed in molecular modulation experiments.20 On the basis
of this analysis, we report a value of k2c/k2 ) 0.19 ( 0.08, where
the error limits include a 25% contribution from possible
systematic errors in the above analysis. This figure is based on
an appraisal of the sensitivity of the system to varying the key
rate coefficients in Table 2 over realistic uncertainty ranges,
and includes uncertainties associated with the yield of phenol
from the OH-initiated oxidation of benzene.
Figure 7 also shows the results of simulations in which k2c/
k2 was set to values of 0.4 and zero. The latter case demonstrates
the contribution from other sources of OH in the system. These
are most notable at high [CH3OH]0/[CH3OCH3]0, where sub-
stantial secondary OH production results from the chemistry
of HOCH2O2 formed from the reaction of HO2 with HCHO, as
indicated above and characterized previously.11 In the absence
of CH3OH (i.e., [CH3OH]0/[CH3OCH3]0 ) 0), significant
formation of HCHO does not occur and the observed formation
of phenol is dominated throughout by the production of OH
via reaction 2c, which is simulated to account for 73% of the
integrated OH produced for a CH3OCH3 depletion of 5 mTorr
(i.e., about 20% depletion). The residual production of OH under
these conditions is partially accounted for by the secondary
removal of CH3OCH2OOH, as follows,
[CH3OCH3]0. The formation of CH3OCH2OOH and CH3OCHO
actually observed in the present experiments with [CH3OH]0/
[CH3OCH3]0 ≈ 15 was also used to provide optimized values
of k2a/k2 and (k2b + k2c)/k2. For this procedure, k2c/k2 was fixed
at a value of 0.19 on the basis of the results of the phenol
analysis above, and compensating changes were made to the
branching ratio not being optimized (i.e., k2b/k2 when optimizing
k2a/k2, and vice versa) to account for the balance of the reaction.
This provided optimized values of k2a/k2 ) 0.55 ( 0.05 and
(k2b + k2c)/k2 ) 0.39 ( 0.04. This analysis also confirmed that
the modest variations applied to k2a/k2 and k2b/k2 during this
procedure had a negligible impact on simulated formation of
phenol. After inclusion of 5% and 10% uncertainties associated
with the quantification of CH3OCHO and CH3OCH2OOH,19 this
leads to values of k2a/k2 ) 0.55 ( 0.08 and (k2b + k2c)/k2 )
0.39 ( 0.05. These branching ratios are therefore consistent
with channels 2a-2c accounting for the entire reaction, within
experimental uncertainties.
4. Conclusions and Atmospheric Implications
The results presented above provide further evidence that the
reactions of HO2 with selected oxygenated RO2 radicals proceed
partially via radical-propagating channels, thereby lessening their
perceived impact as chain terminating processes in the atmo-
sphere. The existence of such channels has previously been
reported for the reactions of HO2 with examples of acyl
8,11,13
(CH3C(O)O2
CH2O2
and C6H5C(O)O213), R-carbonyl (CH3C(O)-
and RO2 radicals produced from C2H5C(O)CH313),
12,13
and R-hydroxy (HOCH2O211) peroxy radicals. The present work
adds the simplest R-alkoxy peroxy radical, CH3OCH2O2, to this
list. It therefore appears that a channel producing OH radicals
exists for the reactions of HO2 with peroxy radicals containing
a number of oxygenated substitutions of atmospheric signifi-
cance. As indicated in section 1, it has been recognized recently
that such reactions probably make a contribution to the
unexpectedly high HOx concentrations observed in locations
where reaction with HO2 radicals is believed to be an important
fate of organic peroxy radicals.14-16 A precise estimation of this
contribution awaits a more detailed understanding of the factors
which influence organic oxidation chemistry under low NOx
conditions. Further experimental and theoretical work in this
area is needed.
Cl + CH3OCH2OOH f CH3OCHOOH + HCl
(18)
References and Notes
(1) Wallington, T. J.; Dagaut, P.; Kurylo, M. J. Chem. ReV. 1992, 92,
667.
CH3OCHOOH f CH3OCHO + OH
(19)
(2) Lightfoot, P. D.; Cox, R. A.; Crowley, J. N.; Destriau, M.; Hayman,
G. D.; Jenkin, M. E.; Moortgat, G. K.; Zabel, F. Atmos. EnViron. 1992,
26A, 1805.
(3) 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. D 2001, 106, 12157.
which is simulated to account for about 15% of the integrated
OH produced. The remaining 12% is due to direct formation
from the minor channel of the reaction of CH3OCH2 with O2:
(4) Jenkin, M. E.; Clemitshaw, K. C. Atmos. EnViron. 2000, 34, 2499.
(5) Wallington, T. J. J. Chem. Soc., Faraday Trans. 1991, 87, 2379.
(6) Spittler, M.; Barnes, I.; Becker, K. H.; Wallington, T. J. Chem.
Phys. Lett. 2000, 321, 57.
CH3OCH2 + O2 f HCHO + HCHO + OH (8b)
(7) Elrod, M. J.; Ranschaert, D. L.; Schneider, N. J. Int. J. Chem. Kinet.
2001, 33, 363.
(8) Hasson, A. S.; Tyndall, G. S.; Orlando, J. J. J. Phys. Chem. A 2004,
108, 5979.
(9) Ravento´s-Duran, M. T.; Percival, C. J.; McGillen, M. R.; Hamer,
P. D.; Shallcross, D. E. Phys. Chem. Chem. Phys. 2007, 9, 4338.
(10) Ravento´s-Duran, M. T.; McGillen, M. R.; Percival, C. J.; Hamer,
P. D.; Shallcross, D. E. Int. J. Chem. Kinet. 2007, 39, 571.
(11) Jenkin, M. E.; Hurley, M. D.; Wallington, T. J. Phys. Chem. Chem.
Phys. 2007, 9, 3149.
This accounts for 1.4% of this reaction at 700 Torr, based on
the parameters derived by Sehested et al.27 Support for this level
of contribution was provided by the observed formation of trace
amounts of HCHO, which agreed well with those simulated.
For the above analysis, the balance of reaction 2 was divided
between channels 2a and 2b, with branching ratios k2a/k2 ) 0.6
and k2b/k2 ) (0.4 - k2c/k2). These ratios correspond to respective
yields of CH3OCH2OOH and CH3OCHO of 0.6 and 0.4 from
reaction 2, which are broadly consistent with those reported
above, and previously,19 for experiments at high [CH3OH]0/
(12) Jenkin, M. E.; Hurley, M. D.; Wallington, T. J. Phys. Chem. Chem.
Phys. 2008, 10, 4274.
(13) Dillon, T. J.; Crowley, J. N. Atmos. Chem. Phys. 2008, 8, 4877.

416 J. Phys. Chem. A, Vol. 114, No. 1, 2010
Jenkin et al.
(14) Thornton, J. A.; Wooldridge, P. J.; Cohen, R. C.; Martinez, M.;
Harder, H.; Brune, W. H.; Williams, E. J.; Roberts, J. M.; Fehsenfeld, F. C.;
Hall, S. R.; Shetter, R. E.; Wert, B. P.; Fried, A. J. Geophys. Res. 2002,
107 (D12), 4146 (doi: 10.1029/2001JD000932).
(15) Butler, T. M.; Taraborrelli, D.; Bru¨hl, C.; Fischer, H.; Harder, H.;
Martinez, M.; Williams, J.; Lawrence, M. G.; Lelieveld, J. Atmos. Chem.
Phys. 2008, 8, 4529.
(16) Hofzumahaus, A.; Rohrer, F.; Lu, K. D.; Bohn, B.; Brauers, T.;
Chang, C. C.; Fuchs, H.; Holland, F.; Kita, K.; Kondo, Y.; Li, X.; Lou,
S. R.; Shao, M.; Zeng, L. M.; Wahner, A.; Zhang, Y. H. Science 2009,
324, 1702.
(17) Jenkin, M. E.; Saunders, S. M.; Pilling, M. J. Atmos. EnViron. 1997,
31, 81.
(18) Saunders, S. M.; Jenkin, M. E.; Derwent, R. G.; Pilling, M. J. Atmos.
Chem. Phys. 2003, 3, 161.
(19) Wallington, T. J.; Hurley, M. D.; Ball, J. C.; Jenkin, M. E. Chem.
Phys. Lett. 1993, 211, 41.
Rate coefficients k9, k13, and k14 also in: Atkinson, R.; Baulch, D. L.; Cox,
R. A.; Crowley, J. N.; Hampson, R. F.; Hynes, R. G.; Jenkin, M. E.; Rossi,
M. J.; Troe, J. Atmos. Chem. Phys. 2006, 6, 3625.
(24) Volkamer, R.; Klotz, B.; Barnes, I.; Imamura, T.; Wirtz, K.;
Washida, N.; Becker, K. H.; Platt, U. Phys. Chem. Chem. Phys. 2002, 4,
1598.
(25) Notario, A.; Mellouki, A.; Le Bras, G. Int. J. Chem. Kinet. 2000,
32, 105.
(26) Platz, J.; Nielsen, O. J.; Wallington, T. J.; Ball, J. C.; Hurley, M. D.;
Straccia, A. M.; Schneider, W. F.; Sehested, J. J. Phys. Chem. A. 1998,
102, 7964.
(27) Sehested, J.; Møgelberg, T.; Wallington, T. J.; Kaiser, E. W.;
Nielsen, O. J. J. Phys. Chem. 1996, 100, 17218.
(28) Wallington, T. J.; Hurley, M. D.; Haryanto, A. Chem. Phys. Lett.
2006, 432, 57.
(29) Vaghjiani, G. L.; Ravishankara, A. R. J. Phys. Chem. 1989, 93,
1948.
(30) Wallington, T. J.; Dagaut, P.; Liu, R.; Kurylo, M. J. Int. J. Chem.
(20) Jenkin, M. E.; Hayman, G. D.; Wallington, T. J.; Hurley, M. D.;
Ball, J. C.; Nielsen, O. J.; Ellermann, T. J. Phys. Chem. 1993, 97, 11712.
(21) Wallington, T. J.; Japar, S. M. J. Atmos. Chem. 1989, 9, 399.
(22) Sokolov, O.; Hurley, M. D.; Wallington, T. J.; Kaiser, E. W.; Platz,
J.; Nielsen, O. J.; Berho, F.; Rayez, M. T.; Lesclaux, R. J. Phys. Chem. A.
1998, 102, 10671.
Kinet. 1988, 20, 177.
(31) Le Calve´, S.; Le Bras, G.; Mellouki, A. J. Phys. Chem. A 1997,
101, 5489.
(32) Jenkin, M. E.; Hayman, G. D. Atmos. EnViron. 1999, 33, 1275.
(23) Current recommendation of the IUPAC Subcommittee for Gas
Kinetic Data Evaluation, available at http://www.iupac-kinetic.ch.cam.ac.uk/.
JP908158W