Branching ratios for the reaction of selected carbonyl-containing peroxy radicals with hydroperoxy radicals

Article abstract of DOI:10.1021/jp211799c

An important chemical sink for organic peroxy radicals (RO₂) in the troposphere is reaction with hydroperoxy radicals (HO₂). Although this reaction is typically assumed to form hydroperoxides as the major products (R1a), acetyl peroxy radicals and acetonyl peroxy radicals have been shown to undergo other reactions (R1b) and (R1c) with substantial branching ratios: RO₂ + HO₂ → ROOH + O₂ (R1a), RO ₂ + HO₂ → ROH + O₃ (R1b), RO₂ + HO₂ → RO + OH + O₂ (R1c). Theoretical work suggests that reactions (R1b) and (R1c) may be a general feature of acyl peroxy and α-carbonyl peroxy radicals. In this work, branching ratios for R1a-R1c were derived for six carbonyl-containing peroxy radicals: C₂H ₅C(O)O₂, C₃H₇C(O)O₂, CH₃C(O)CH₂O₂, CH₃C(O)CH(O ₂)CH₃, CH₂ClCH(O₂)C(O)CH ₃, and CH₂ClC(CH₃)(O₂)CHO. Branching ratios for reactions of Cl-atoms with butanal, butanone, methacrolein, and methyl vinyl ketone were also measured as a part of this work. Product yields were determined using a combination of long path Fourier transform infrared spectroscopy, high performance liquid chromatography with fluorescence detection, gas chromatography with flame ionization detection, and gas chromatography-mass spectrometry. The following branching ratios were determined: C₂H₅C(O)O₂, Y_R1a = 0.35 ± 0.1, Y_R1b = 0.25 ± 0.1, and Y_R1c = 0.4 ± 0.1; C₃H₇C(O)O₂, Y_R1a = 0.24 ± 0.15, Y_R1b = 0.29 ± 0.1, and Y_R1c = 0.47 ± 0.15; CH₃C(O)CH₂O₂, Y _R1a = 0.75 ± 0.13, Y_R1b = 0, and Y_R1c = 0.25 ± 0.13; CH₃C(O)CH(O₂)CH₃, Y _R1a = 0.42 ± 0.1, Y_R1b = 0, and Y_R1c = 0.58 ± 0.1; CH₂ClC(CH₃)(O₂)CHO, Y _R1a = 0.2 ± 0.2, Y_R1b = 0, and Y_R1c = 0.8 ± 0.2; and CH₂ClCH(O₂)C(O)CH₃, Y_R1a = 0.2 ± 0.1, Y_R1b = 0, and Y_R1c = 0.8 ± 0.2. The results give insights into possible mechanisms for cycling of OH radicals in the atmosphere.

Full text of DOI:10.1021/jp211799c

Article
pubs.acs.org/JPCA
Branching Ratios for the Reaction of Selected Carbonyl-Containing
Peroxy Radicals with Hydroperoxy Radicals
Alam S. Hasson,*,† Geoffrey S. Tyndall,‡ John J. Orlando,‡ Sukhdeep Singh,†,§
Samuel Q. Hernandez,†,∥ Sean Campbell,†,⊥ and Yesenia Ibarra†
†Department of Chemistry, 2555 East San Ramon Avenue M/S SB70, California State University,
Fresno, Fresno, California 93740, United States
‡Atmospheric Chemistry Division, National Center for Atmospheric Research, P.O. Box 3000, Boulder, Colorado 80307,
United States
S
* Supporting Information
ABSTRACT: An important chemical sink for organic peroxy
radicals (RO2) in the troposphere is reaction with hydroperoxy
radicals (HO2). Although this reaction is typically assumed to
form hydroperoxides as the major products (R1a), acetyl
peroxy radicals and acetonyl peroxy radicals have been shown
to undergo other reactions (R1b) and (R1c) with substantial
branching ratios: RO2 + HO2 → ROOH + O2 (R1a), RO2 +
HO2 → ROH + O3 (R1b), RO2 + HO2 → RO + OH + O2
(R1c). Theoretical work suggests that reactions (R1b) and
(R1c) may be a general feature of acyl peroxy and α-carbonyl
peroxy radicals. In this work, branching ratios for R1a−R1c
were derived for six carbonyl-containing peroxy radicals:
C2H5C(O)O2, C3H7C(O)O2, CH3C(O)CH2O2, CH3C(O)-
CH(O2)CH3, CH2ClCH(O2)C(O)CH3, and CH2ClC(CH3)-
(O2)CHO. Branching ratios for reactions of Cl-atoms with
butanal, butanone, methacrolein, and methyl vinyl ketone were
also measured as a part of this work. Product yields were determined using a combination of long path Fourier transform infrared
spectroscopy, high performance liquid chromatography with fluorescence detection, gas chromatography with flame ionization
detection, and gas chromatography−mass spectrometry. The following branching ratios were determined: C2H5C(O)O2, YR1a
=
0.35 0.1, YR1b = 0.25 0.1, and YR1c = 0.4 0.1; C3H7C(O)O2, YR1a = 0.24 0.15, YR1b = 0.29 0.1, and YR1c = 0.47 0.15;
CH3C(O)CH2O2, YR1a = 0.75 0.13, YR1b = 0, and YR1c = 0.25 0.13; CH3C(O)CH(O2)CH3, YR1a = 0.42 0.1, YR1b = 0, and
YR1c = 0.58 0.1; CH2ClC(CH3)(O2)CHO, YR1a = 0.2 0.2, YR1b = 0, and YR1c = 0.8 0.2; and CH2ClCH(O2)C(O)CH3,
Y
R1a = 0.2 0.1, YR1b = 0, and YR1c = 0.8 0.2. The results give insights into possible mechanisms for cycling of OH radicals in
the atmosphere.
where R = CH3C(O)) with relative branching ratios for
INTRODUCTION
■
YR1a:YR1b of 3:1.9−13
Organic peroxy radicals (RO2) are generated during the photo-
oxidation of the vast majority of organic pollutants in the
troposphere.1 The principal chemical sinks are reaction with
nitrogen oxides (NOx = NO + NO2) and hydroperoxy radicals
(HO2). The reaction between organic peroxy and hydroperoxy
radicals is generally considered to be a chain terminating pro-
cess, resulting in the formation of an organic hydroperoxide2−7
(ROOH, reaction R1a).
RO2 + HO2 → ROH + O3
(R1b)
More recently, convincing evidence has emerged that acetyl
peroxy and acetonyl peroxy radicals form radical products
(reaction R1c) with substantial yields.14−17
RO2 + HO2 → RO + OH + O2
(R1c)
A theoretical study provided supporting evidence for the experi-
mental results.18 Structural calculations with Master Equation
RO2 + HO2 → ROOH + O2
(R1a)
Because organic hydroperoxides are relatively stable in the
troposphere,8 reaction R1a is a radical sink.
Special Issue: A. R. Ravishankara Festschrift
Experimental and theoretical studies have demonstrated that
reactions between carbonyl-containing organic peroxy radicals
and hydroperoxy radicals are more complex. Acetyl peroxy
radicals have been shown to generate acetic acid (reaction R1b,
Received: December 7, 2011
Revised: March 8, 2012
Published: April 6, 2012
© 2012 American Chemical Society
6264
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
analysis identified a reaction mechanism that could potentially
account for the large observed branching ratios for (R1c). In
this mechanism, the HO2 and RO2 radicals form a hydro-
tetroxide intermediate that can subsequently decompose to
form the radical products. The intermediate is internally
stabilized by hydrogen bonding between the OH and carbonyl
groups, and it is this interaction that enables the RO and OH
products to be formed. This suggests that reaction R1c may be
a general feature for all acylperoxy and α-carbonyl peroxy
radicals.
A significant branching ratio for reaction R1c could
substantially affect the predicted photochemistry in atmo-
spheric models because this channel results in radical recycling
as opposed to removal, which would be the case for reaction
R1a. However, the impact of these reactions is clearly
dependent on how ubiquitous reaction R1c is for carbonyl-
containing peroxy radicals. In this work, branching ratios for
reactions R1a−R1c are reported for six carbonyl-containing
peroxy radicals. These data are used to evaluate the relationship
between the branching ratios and the radical structure.
and products were calibrated using commercially available
standards unless stated otherwise.
The NCAR reactor is a 47 L stainless steel chamber coupled
to a Bomem DA3.01 FTIR spectrometer. Photochemistry is
initiated using a filtered xenon arc lamp with radiation in the
wavelength range 235−400 nm. The Fresno State chamber is a
142 L Teflon-lined reactor coupled to a Nicolet Nexus-470
FTIR spectrometer. Internally mounted blacklight lamps,
generating radiation in the wavelength range 300−400 nm,
are used as the photolysis source. Hanst-type optics are used to
produce IR pathlengths of 32.6 and 50 m for the NCAR and
Fresno State chambers, respectively. IR spectra were collected
at 1 cm−1 resolution by averaging 200 scans following
irradiation times ranging from 30 s to 15 min, depending on
the reaction mixture. This process was typically repeated 4−5
times over the course of the experiment, resulting in total
irradiation times ranging from 2−75 min.
Ports mounted at the end of each chamber were used to
extract samples for analysis by HPLC and GC. Additional
experiments were carried out at Fresno State using 20 L Tedlar
bags surrounded by blacklight lamps within a custom-built
chamber.
EXPERIMENTAL SECTION
■
Samples were withdrawn from the chamber for analysis by
GC-FID and HPLC at a flow rate of approximately 200 mL/min
through unheated Teflon tubing. HPLC samples were bubbled
through an impinger containing 15 mL of 18 MΩ water contain-
ing 1 mM sulfuric acid and 0.1 mM Na2EDTA. The volume
of air sampled from the chamber was typically in the range
0.5−9 L. Samples were subsequently analyzed as described
previously14 using an HPLC with postcolumn derivatization
and fluorescence detection technique first described by
The majority of the experiments were performed using the
photochemical reactor located at the National Center for
Atmospheric Research (NCAR). Additional experiments with
propanal and butanal were carried out using the chamber at
California State University Fresno (Fresno State). Experiments
were carried out at a total pressure of 800 Torr and at room
temperature (295 K). Reaction mixtures containing (3.5−7) × 1014
molecules cm−3 of an organic peroxy radical precursor
(aldehyde or ketone), (0, 0.35−3.5) × 1015 molecules cm−3
hydroperoxy radical precursor (an alcohol), (1.4−3.5) × 1015
molecules cm−3 chlorine, nitrogen, and oxygen were introduced
into the chamber. Experiments carried out using acetone as the
organic peroxy radical precursor were performed with samples
labeled with 13C at the carbonyl carbon atom to enable CO2
formation from acetone to be distinguished from other sources.
Some additional experiments were carried out in which NO or
NO2 (7−10) × 1014 molecules cm−3 were added to the reaction
mixture to help determine the branching ratio for Cl-atom
attack. The chlorine photodissociates in the presence of UV
light to produce chlorine atoms (reaction R2) which then react
with the carbonyl compound and alcohol to produce an organic
peroxy radical and hydroperoxy radical in reactions R3−R6,
respectively.
Hellpointner and Gab.20
̈
GC-FID samples either were collected in a gastight syringe
and injected into the GC inlet or were flowed into the sample
loop of a six-port valve connected directly to the GC. The GC-
FID instrument at Fresno State (Agilent 6850) was fitted with a
30 m, 0.25 mm i.d., 0.5 μm film thickness HP5-MS column and
used helium as the carrier gas at a flow rate of 1 mL min−1. The
oven was held at −10 °C with liquid CO2 for 1 min and then
ramped at 5 °C min−1 to 30 °C and then at 30 °C min−1 to a
final temperature of 200 °C, which was held for 1 min. The
NCAR GC-FID instrument (Agilent 6980N) used the same
column, carrier gas, and flow rates as the Fresno State instru-
ment and a similar oven temperature program (−20 °C for
1 min; ramped at 5 °C min−1 to 30 °C and then at 30 °C min−1
to a final temperature of 200 °C, held for 1 min).
Cl2 + hν → 2Cl
(R2)
Additional experiments were carried out at Fresno State to
measure carboxylic acid concentrations within the chamber
using solid phase microextraction (SPME) fibers coupled with
GC−MS, as first described by Orzechowska and co-workers.21
The fibers were conditioned prior to their first use by heating
them to 220 °C for 30 min in the inlet of a GC/MS (HP 5890/
5973). Samples were then pumped from the chamber at a flow
rate of 1 L/min into a 1 L pyrex sampling chamber fitted with a
septum to allow for the introduction of the SPME fiber. The
70 μm Carbowax/divinylbenzene SPME fiber (Supelco) was
then exposed to the sample flow for 1 min. Samples were
thermally desorbed from the SPME fiber into the 0.75 mm
internal diameter inlet linear of the GC (HP 5890) at 200 °C.
To ensure complete desorption of the analytes, the fibers were
left in the GC inlet for the duration of the run. A 30 m Stabilwax
DA column (Restek) was used to separate the compounds present
using helium as the carrier gas at a flow rate of 1 mL min−1.
(R3)
Cl + R−H → R + HCl
R + O2 + M → RO2 + M
(R4)
R1R2CHOH + Cl → R1R2COH + HCl
(R5)
R1R2COH + O2 → R1R2CO + HO2
(R6)
The relative concentrations of RO2 and HO2 radicals within
the chamber were varied by adjusting the initial concentrations
of the aldehyde/ketone and alcohol. Changes in the con-
centrations of reactants and products were measured using a
combination of Fourier transform infrared (FTIR) spectroscopy,
high performance liquid chromatography (HPLC), gas
chromatography with flame ionization detection (GC-FID) and
gas chromatography−mass spectrometry (GC−MS). Reactants
6265
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
Table 1. Principal Chemical Reactions Used in Acuchem Simulations
a
b
reaction
rate coefficient/s−1 or cm3 molecule−1 s−1
1 × 10−4
Cl2 = Cl + Cl
HO2 Precursors
6 × 10−11
6 × 10−11
8.4 × 10−11
Cl + CH3OH(+O2) = CH2O + HCl + HO2
Cl + CH3C(O)CH2OH(+O2) = HCl + CH3C(O)CHO + HO2
Cl + CH3CH(OH)CH3(+O2) = HCl + CH3C(O)CH3 + HO2
Propanal Reactions
c
1 × 10−10
Cl + C2H5CHO(+O2) = HCl + C2H5C(O)O2
c
1 × 10−11
Cl + C2H5CHO(+O2) = HCl + RO2
YR1c × 2.2 × 10−11 (see text for details)
C2H5C(O)O2 + HO2 = OH + C2H5C(O)O + O2
C2H5C(O)O2 + HO2 = C2H5C(O)OH + O3
C2H5C(O)O2 + HO2 = C2H5C(O)O2H + O2
C2H5C(O)O = CO2 + C2H5O2
YR1b × 2.2 × 10−11 (see text for details)
YR1a × 2.2 × 10−11 (see text for details)
1 × 105
9 × 10−12
2 × 10−12
4 × 10−14
1.7 × 10−14
7.8 × 10−12
1.4 × 10−11
C2H5O2 + C2H5C(O)O2 = C2H5C(O)O + C2H5O + O2
C2H5O2 + C2H5C(O)O2 = C2H5C(O)OH + CH3CHO + O2
C2H5O2 + C2H5O2 = C2H5O + C2H5O + O2
C2H5O2 + C2H5O2 = C2H5OH + CH3CHO + O2
C2H5O2 + HO2 = C2H5O2H + O2
C2H5C(O)O2 + C2H5C(O)O2 = C2H5C(O)O + C2H5C(O)O + O2
Butanal Reactions
c
YR12a × 1.6 × 10−10 (see text for details)
Cl + C3H7CHO(+O2) = C3H7C(O)O2 + HCl
c
(1 − YR12a) × 1.6 × 10−10 (see text for details)
1.4 × 10−11
Cl + C3H7CHO(+O2) = RO2 + HCl
C3H7C(O)O2 + C3H7C(O)O2 = C3H7C(O)O + C3H7C(O)O + O2
C3H7C(O)O2 + RO2 = C3H7C(O)O + RO + O2
C3H7C(O)O2 + RO2 = C3H7C(O)OH + other products
RO2 + RO2 = products
9 × 10−12
2 × 10−12
4 × 10−12
YR1a × 2.2 × 10−11 (see text for details)
YR1b × 2.2 × 10−11 (see text for details)
C3H7C(O)O2 + HO2 = C3H7C(O)O2H + O2
C3H7C(O)O2 + HO2 = C3H7C(O)OH + O3
C3H7C(O)O2 + HO2 = C3H7C(O)O + OH + O2
RO2 + HO2 = ROOH + O2
YR1c × 2.2 × 10−11 (see text for details)
1.6 × 10−11
1 × 106
C3H7C(O)O = CO2 + other products
Acetone Reactions
2.1 × 10−12
Cl + CH3C(O)CH3 = CH3C(O)CH2 + HCl
5 × 10−12
CH3C(O)CH2 + O2 = CH3C(O)CH2O2
6 × 10−12
2CH3C(O)CH2O2 = 2CH3C(O)CH2O + O2
2 × 10−12
2CH3C(O)CH2O2 = CH3C(O)CHO + CH3C(O)CH2OH + O2
CH3C(O)CH2O = CH3C(O)O2 + CH2O
6 × 107
1 × 10−14
CH3C(O)CH2O + O2 = HO2 + CH3C(O)CHO
1 × 10−11
CH3C(O)CH2O2 + CH3C(O)O2 = CH3C(O)CH2O + CH3CO2 + O2
CH3C(O)CH2O2 + CH3C(O)O2 = CH3C(O)CHO + CH3C(O)OH + O2
CH3C(O)CH2O2 + CH3O2 = CH3C(O)CH2O + CH3O + O2
CH3C(O)CH2O2 + CH3O2 = CH3C(O)CHO + CH3OH + O2
CH3C(O)CH2O2 + CH3O2 = CH3C(O)CH2OH + CH2O + O2
CH3C(O)CH2O2 + HO2 = CH3C(O)CH2OOH + O2
CH3C(O)CH2O2 + HO2 = CH3C(O)CH2O + OH + O2
1 × 10−12
1.1 × 10−12
1.4 × 10−12
1.4 × 10−12
YR1a × 9 × 10−12 (see text for details)
YR1c × 9 × 10−12 (see text for details)
6266
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Table 1. continued
Article
a
b
reaction
rate coefficient/s−1 or cm3 molecule−1 s−1
Butanone Reactions
c
YR19a × 4.1 × 10−11 (see text for details)
Cl + CH3C(O)CH2CH3(+O2) = CH3C(O)CH(O2)CH3 + HCl
Cl + CH3C(O)CH2CH3(+O2) = RO2 + HCl
c
(1 − YR19a) × 4.1 × 10−11 (see text for details)
YR34a × 4 × 10−12 (see text for details)
(1 − YR34a) × 4 × 10−12 (see text for details)
2.9 × 10−12
1.9 × 10−12
2.4 × 10−12
2CH3C(O)CH(O2)CH3 = 2CH3C(O)CH(O)CH3 + O2
2CH3C(O)CH(O2)CH3 = other products
RO2 + RO2 = RO + RO + O2
RO2 + RO2 = products
CH3C(O)CH(O2)CH3 + RO2 = CH3C(O)CH(O)CH3 + RO + O2
CH3C(O)CH(O2)CH3 + RO2 = products
1.6 × 10−12
c
1 × 107
CH3C(O)CH(O)CH3(+O2) = CH3CHO + CH3C(O)O2
CH3C(O)CH(O2)CH3 + HO2 = CH3C(O)CH(O2H)CH3 + O2
CH3C(O)CH(O2)CH3 + HO2 = CH3C(O)CH(O)CH3 + OH + O2
YR1a × 1 × 10−11 (see text for details)
YR1c × 1 × 10−11 (see text for details)
MAC Reactions
c
(1 − YR21a) × 3 × 10−10 (see text for details)
Cl + CH3C(CH2)CHO(+O2) = CH3C(O2)(CH2Cl)CHO
Cl + CH3C(CH2)CHO(+O2) = RO2 + HCl
c
YR21a × 3 × 10−10 (see text for details)
2.4 × 10−12
2CH3C(O2)(CH2Cl)CHO = 2CH3C(O)(CH2Cl)CHO + O2
RO2 + RO2 = RO + RO + O2
1.0 × 10−11
2.4 × 10−12
CH3C(O2)(CH2Cl)CHO + RO2 = CH3C(O)(CH2Cl)CHO + RO + O2
CH3C(O)(CH2Cl)CHO(+O2) = CH3C(O)CH2Cl + HO2 + CO
CH3C(O2)(CH2Cl)CHO + HO2 = CH3C(O2H)(CH2Cl)CHO + +O2
CH3C(O2)(CH2Cl)CHO + HO2 = CH3C(O)(CH2Cl)CHO + +OH + O2
OH + CH3C(CH2)CHO(+O2) = CH3C(O2)(CH2OH)CHO
OH + CH3C(CH2)CHO(+O2) = RO2 + H2O
c
1 × 107
YR1a × 1 × 10−11 (see text for details)
YR1c × 1 × 10−11 (see text for details)
c
1.5 × 10−11 (see text for details)
c
1.5 × 10−11 (see text for details)
YR1a × 1 × 10−11 (see text for details)
YR1c × 1 × 10−11 (see text for details)
CH3C(O2)(CH2OH)CHO + HO2 = CH3C(O2H)(CH2OH)CHO + +O2
CH3C(O2)(CH2OH)CHO + HO2
= CH3C(O)(CH2OH)CHO + +OH + O2
RO2 + HO2 = ROOH + O2
RO2 + HO2 = ROH + O3
RO2 + HO2 = RO + OH + O2
YR1a × 2.2 × 10−11 (see text for details)
YR1b × 2.2 × 10−11 (see text for details)
YR1c × 2.2 × 10−11 (see text for details)
MVK Reactions
YR27a × 2 × 10−10 (see text for details)
(1 − YR27a) × 2 × 10−10 (see text for details)
YR34a × 4 × 10−12 (see text for details)
(1 − YR34a) × 4 × 10−12 (see text for details)
Cl + CH2CHC(O)CH3 + O2 = CH2ClCH(O2)C(O)CH3
Cl + CH2CHC(O)CH3 + O2 = RO2
2CH2ClCH(O2)C(O)CH3 = 2CH2ClCH(O)C(O)CH3 + O2
2CH2ClCH(O2)C(O)CH3 = products
2.9 × 10−12
1.9 × 10−12
2.4 × 10−12
RO2 + RO2 = RO + RO + O2
RO2 + RO2 = products
CH2ClCH(O2)C(O)CH3 + RO2 = CH2ClCH(O)C(O)CH3 + RO + O2
CH2ClCH(O2)C(O)CH3 + RO2 = products
c
1.6 × 10−12
1 × 107
CH2ClCH(O)C(O)CH3(+O2) = CH2ClCHO + CH3C(O)O2
CH2ClCH(O2)C(O)CH3 + HO2 = CH2ClCH(O2H)C(O)CH3 + O2
CH2ClCH(O2)C(O)CH3 + HO2 = CH2ClCH(O)C(O)CH3 + OH + O2
RO2 + HO2 = ROOH + O2
YR1a × 1 × 10−11 (see text for details)
YR1c × 1 × 10−11 (see text for details)
1.4 × 10−11
Other Reactions
1.4 × 10−11
1.1 × 10−11
CH3C(O)O2 + CH3C(O)O2 = CH3CO2 + CH3CO2 + O2
CH3C(O)O2 + CH3O2 = CH3CO2 + CH3O + O2
6267
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Table 1. continued
Article
a
b
reaction
rate coefficient/s−1 or cm3 molecule−1 s−1
9 × 10−12
Other Reactions
CH3C(O)O2 + HO2 = CH3C(O)OOH + O2
CH3C(O)O2 + HO2 = CH3C(O)OH + O3
4 × 10−12
9 × 10−12
CH3C(O)O2 + HO2 = CH3CO2 + OH + O2
CH3O2 + CH3O2 = CH3OH + CH2O + O2
CH3O2 + CH3O2 = CH3O + CH3O + O2
CH3O2 + HO2 = CH3OOH + O2
HO2 + HO2 = H2O2 + O2
2.2 × 10−13
1.3 × 10−13
5.2 × 10−12
3 × 10−12
2 × 10−15
CH3O + O2 = HO2 + CH2O
7.3 × 10−11
5.5 × 10−12
Cl + CH2O = HCO + HCl
HCO + O2 = HO2 + CO
c
c
1 × 108
CH3CO2(+O2) = CH3O2 + CO2
4.5 × 10−11
1 × 10−10
1 × 10−10
Cl + CH3C(O)CHO(+O2) = HCl + CH3C(O)O2 + CO
Cl + CH3C(O)CH2O2 = CH3C(O)CH2O + ClO
Cl + CH3C(O)CH2OOH = HCl + OH + CH3C(O)CHO
a
b
Additional secondary reactions with minimal effects on the model fits are not shown. Rate coefficients are taken from refs 22, 26, 39−41. RO2 +
c
RO2 rate coefficients are estimated from these references for reactions where data are unavailable. The reaction consists of more than one
elementary reaction step. The rate coefficient listed corresponds to the rate limiting step.
The column was initially held at 50 °C for 4 min and was
then ramped at 12 °C min−1 to a final temperature of 200 °C.
This temperature was held for 3 min, resulting in a total run
time of 19.5 min. The mass spectrometer (HP 5973) was
operated in electron impact ionization mode (at 70 eV) and
signals were collected in the selective ion monitoring (SIM)
mode using m/z 59 and m/z 73 for all of the acids.
reactions to provide the best fit to the experimental product
yields measured in the absence of an HO2 precursor.
Simulations were then run in which the appropriate RO2 +
HO2 branching ratios were systematically varied. Best fits and
uncertainties are estimated “by eye”. The uncertainties in the
reported branching ratios are determined by propagating the
estimated uncertainty in the model fit to the experimental data
with the uncertainties in the experimental product yields.
Branching Ratios for Cl + Carbonyls. Chlorine atoms
abstract hydrogen atoms from carbonyl compounds to produce
an alkyl radical. However, because chlorine atoms are not
highly selective,22 reactions R3 and R4 result in a mixture of
organic peroxy radicals and the branching ratios for several of
the aldehydes/ketones studied are not well established. Because
the product yields predicted by the chemical model are
sensitive to these branching ratios (see below), additional
experiments were performed to determine the yields of the
desired organic peroxy radicals in reactions R3 and R4 for
selected organic peroxy radical precursors. The branching ratios
determined in these experiments are summarized in Table 2
RESULTS AND DISCUSSION
■
Product yields, defined as Δ[product]/−Δ[reactant], were
determined in each experiment from linear least-squares fits to
plots of [product] against −Δ[reactant]. Reported uncertainties
in these yields are one standard deviation of the slope of these
fits. Branching ratios for the RO2 + HO2 reaction were then
determined by fitting the measured product yields to the output
from a detailed chemical model (Acuchem19) in which the
branching ratios are adjustable parameters. The key reactions in
the reaction mechanism used in these simulations are shown in
Table 1. Rate data are not available for a number of the RO2 +
RO2 reactions of potential importance in these experiments, and
were estimated from literature values for structurally similar radicals.
Additional secondary reactions were included in the model but were
found to have a minor impact on the model fits.
Table 2. Experimental Branching Ratios for Selected Cl +
Carbonyl Reactions
The sensitivity of the model output to the rate coefficients
used in the chemical mechanism was evaluated by systemati-
cally varying the values for the reaction rate constants by up to
a factor of 10. For the majority of the reactions included in the
model, the predicted concentrations are virtually independent
of the rate coefficients used. The model output is, however,
highly sensitive to the branching ratios used for the major
RO2 + RO2 and RO2 + HO2 branching ratios. In the absence of
an HO2 source, the observed product yields are dominated by
the RO2 + RO2 reactions for the major RO2 radicals generated
in the initial steps of the mechanism. In the presence of a high
concentration of an HO2 source, the products are largely
determined by the RO2 + HO2 reaction. Model fits were carried
out by first adjusting the branching ratios for RO2 + RO2
reaction
branching ratio
ref
YR12a = 0.65, YR12b = 0.08,
YR12c = 0.13, YR12d = 0.14
this work
Cl + CH3CH2CH2C(O)H
YR12a ≥ 0.42, YR12b < 0.25,
27
YR12c = 0.21, YR12d ≤ 0.12
YR12a = 0.66 0.04
YR19a = 0.64 0.04
26
this work
Cl + CH3C(O)CH2CH3
YR19a = 0.73 0.09
29
30
YR19a = 0.76 0.07, YR19b
=
0.031 0.006, YR19c
=
0.225 0.025
YR21a = 0.2, YR21b = 0.8
this work
34
Cl + MAC
Cl + MVK
YR21a = 0.25, YR21b = 0.47
YR27a = 0.65
this work
6268
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
Figure 1. Reaction pathways leading to the formation of major products from the Cl/OH radical initiated photoxidation of carbonyls in the absence
of an HO2 source: (a) butanal + Cl/OH; (b) butanone + Cl; (c) methyl vinyl ketone + Cl; (d) methacrolein + Cl. Some reactants and products are
not shown. Labeled reactions are the initial steps in a sequence of reactions leading to the products shown.
and are discussed in detail below. The major reaction pathways
and products are summarized in Figure 1.
oxygen, butanoyl peroxy radicals are produced with a high
yield.
Cl + Butanal. The yield of peroxy butanoyl nitrate
(PBN) from the photolysis of Cl2/butanal/NO2/air mix-
tures was measured by FTIR. Because PBN is not com-
mercially available, a reference spectrum was generated in
situ by the photolysis of isopropyl nitrite (IPN)/butanal/
NO/NO2/air mixtures. IPN photolyzes to produce
isopropoxy radicals and NO (reaction R7), which leads to
the formation of OH radicals and acetone via reactions R8
and R9.
CH3CH2CH2CHO + OH(+O2 + M)
→ CH3CH2CH2C(O)O2 + H2O(+M)
(R10)
These peroxy radicals then react with NO2 to produce PBN
(reaction R11).
CH3CH2CH2C(O)O2 + NO2 + M
→ CH3CH2CH2C(O)O2NO2 + M
(R11)
CH3CH(ONO)CH3 + hv → CH3CH(O)CH3 + NO
(R7)
In the experiments, PBN is the only reaction product
observed in the IR absorption spectrum during the early stages
of the reaction. An IR absorption cross section of 7.0 × 10−19 cm−1
at 1835 cm−1 was used to calibrate for PBN, which is consistent
with the cross sections of PPN and PiBN measured in this
lab. This corresponded to a PBN yield of 85%. Monedero and
co-workers24 recently reported a significantly lower cross
section for this species (3.6 × 10−19 cm−1 at 1834 cm−1)
using NO3 to oxidize butanal to PBN. However, the present
study is consistent with the higher cross sections for PANs
measured previously in this laboratory.
CH3CH(O)CH3 + O2 → CH3C(O)CH3 + HO2
(R8)
HO2 + NO → OH + NO2
(R9)
The OH radicals then react with butanal (reaction R10).
It is well established that the reaction between hydroxyl
radicals and small linear aldehydes occurs to a large
extent by H-atom abstraction from the acyl group (e.g.,
see ref 23 and references therein), and so in the presence of
6269
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
The major pathways leading to the observed reaction pro-
ducts are shown schematically in Figure 1a). Chlorine atoms
may abstract a hydrogen atom from any of the four carbon
atoms of butanal to produce an alkyl or acyl radical that reacts
with oxygen, leading to the formation of four different peroxy
radicals (reactions R12a−R12d).
reduced because of scavenging of propoxy radicals by NO
and NO2.
CH3CH2CH2O2 + NO(+O2)
→ CH3CH2CHO + NO2 + HO2
(R15)
Reaction R14c is potentially competitive with the reaction of
this alkoxy radical with O2, which would result in the formation
of multifunctional products. However, on the basis of the study
of a structurally similar radical,25 reaction R14c is expected to
be the major pathway, accounting for over 80% of the products.
The peroxy radical generated in reaction R14c is expected to
form formaldehyde and CO via reaction with NO followed by
decomposition and reaction with O2.
CH3CH2CH2C(O)H + Cl + O2(+M)
→ CH3CH2CH2C(O)O2 + HCl(+M)
(R12a)
→ CH3CH2CH(O2)C(O)H + HCl(+M) (R12b)
→ CH3CH(O2)CH2C(O)H + HCl(+M) (R12c)
→ CH2(O2)CH2CH2C(O)H + HCl(+M) (R12d)
O2CH2C(O)H + NO → OCH2C(O)H + NO2
(R16)
In the presence of excess NO2, the butanoyl peroxy radicals
react almost exclusively via reaction R11 to produce PBN. The
measured yield of PBN in these experiments is therefore equi-
valent to the branching ratio for reaction R12a, i.e., 0.65 0.1.
Mixtures containing Cl2, butanal, NO, and air were then
photolyzed, and product yields were determined using a com-
bination of FTIR spectroscopy (for all products) and GC-FID
(for carbonyl compounds). The measured yields for CO, CO2,
propanal, and acetaldehyde in these experiments are 0.21
0.02, 0.7 0.09, 0.26 0.04, and 0.13 0.02, respectively. The
reported propanal and acetaldehyde yields are averages of the
values measured by FTIR and GC-FID. The organic peroxy
radicals formed in reaction R4 react with NO, when present, to
form alkoxy radicals (reactions R13a−R13d).
OCH2C(O)H → CH2O + HCO(+O2)
→ CH2O + CO + HO2
(R17)
The peroxy radical formed in reaction R14d will react with NO
and then decompose to produce CO2 and a hydroxypropyl
radical
CH2(OH)CH2CH2C(O)O2 + NO
→ CH2(OH)CH2CH2C(O)O + NO2
→ CH2(OH)CH2CH2+CO2
(R18)
which likely leads to formation of the multifunctional species,
HOCH2CH2CHO. Thus in the presence of NO, propanal is
produced via channels R12a and R12b, CO2 is produced via
channels R12a and R12d, CO is produced via channels R12b
and R12c, and acetaldehyde is produced via channel R12c.
The yield of acetaldehyde in these experiments is expected to
be about 80% of the branching ratio for reaction R12c.25 Because
CO is only expected to be formed from R12b and R12c, the
CH3CH2CH2C(O)O2 + NO
→ CH3CH2CH2C(O)O + NO2
(R13a)
CH3CH2CH(O2)C(O)H + NO
→ CH3CH2CH(O)C(O)H + NO2
(R13b)
branching ratio for reaction R12b can be estimated as YCO
−
CH3CH(O2)CH2C(O)H + NO
Y
acetaldehyde/0.8 (which includes a correction for the 20% of
radicals formed in reaction R13c that are expected to react with
O2 rather than via reaction R14c). Finally, the branching ratio
→ CH3CH(O)CH2C(O)H + NO2
(R13c)
for reaction R12d can be estimated from 1 − (YR12a + YR12b
R12c) using the branching ratios for reactions R12a−R12c as
determined above. Combining these results together gives the
following branching ratios: YR12a = 0.65, YR12b = 0.05, YR12c
+
CH2(O2)CH2CH2C(O)H + NO
→ CH2(O)CH2CH2C(O)H + NO2
Y
(R13d)
=
These species are then expected to fragment or isomerize
0.16, and YR12d = 0.14. There are two previously published
values for YR12a. Singh and co-workers26 measured the yield of
butanoyl chloride from the photolysis of butanal/Cl2/N2
mixtures and reported a branching ratio of 0.68 for reaction
R12a, in excellent agreement with this work. Wu and Mu
measured27 product yields in a series of experiments using
butanal/Cl2/NO/air mixtures in a study that is analogous to
the work reported here. Branching ratios of ≥0.42, <0.25, 0.21,
and ≤0.12 are reported for reactions R12a−R12d, respectively.
The disagreement between these values and those reported in
this study is due to the differences in the product yields
measured. Wu and Mu measured a much lower yield of CO2
compared to this work, leading to a lower estimate for the
branching ratio for reaction R12a than in this work. The
authors also report a significantly higher yield of acetaldehyde,
leading to a larger branching ratio for reaction R12c compared
to the value reported here. Though the origins of these
discrepancies are not clear, we note that our acetaldehyde and
propanal yields were measured by two independent techniques
(reactions R14a−R14d).
CH3CH2CH2C(O)O(+O2) → CH3CH2CH2O2 + CO2
(R14a)
CH3CH2CH(O)C(O)H(+O2)
→ CH3CH2CHO + CO + HO2
(R14b)
CH3CH(O)CH2C(O)H(+O2)
→ CH3CHO + O2CH2C(O)H
(R14c)
CH2(O)CH2CH2C(O)H(+O2)
→ CH2(OH)CH2CH2C(O)O2
(R14d)
The propyl peroxy radicals produced in reaction R14a will react
with NO to form propoxy radicals, which may then react with
O2 to form propanal; however, the yield of propanal can be
6270
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
(FTIR and GC-FID), and that the concentrations of CO, CO2,
and PBN were directly measured (by FTIR). In contrast, Wu
and Mu used GC-FID for all their analysis, and measured CO
and CO2 following their conversion to methane, which
potentially introduced greater uncertainties. Further, the
branching ratios measured here are more consistent with the
trend in branching ratios for C2−C5 aldehydes,26 and are in
excellent agreement with our previous measurements using a
different chemical scheme.
Cl + Butanone. The acetaldehyde yield from the photolysis
of Cl2/butanone/NO/air mixtures was measured using GC-
FID because the IR analysis of acetaldehyde is compli-
cated by other absorption bands. Yields of CO2 were also
measured using FTIR. The yields of acetaldehyde and CO2
are 0.68 0.04 and 0.6 0.1, respectively. The acetaldehyde
yield includes a correction for the secondary reaction of Cl
with acetaldehyde, which is significantly faster than the
corresponding reaction with butanone, and which results in
loss of about 30% of the acetaldehyde formed by the end of
the experiments.
mixtures measured using FTIR are 0.33 0.01, 0.72 0.04,
and 0.4 0.01, respectively. The formation of peroxymetha-
cryloyl nitrate (MPAN) was also observed late in the reaction.
Experiments conducted on the photolysis of MAC/Cl2/NO2/
air mixture indicated a yield of 15−20% MPAN, corresponding
to abstraction at the −CHO group.
The major pathways and products for the reaction between
Cl and MAC are summarized in Figure 1c. The reaction is
expected to proceed via either H-atom abstraction from the acyl
group or addition to the terminal end of the double bond. The
yield of acetyl chloride, which is expected to be formed from
internal addition, was less than 2%.
The subsequent reactions of these products lead to the
formation of CO2 from reaction R21a followed by R22,
and CO and chloroacetone from reaction R21b followed
by R23.
The major reaction pathways and reaction products are
shown in Figure 1b. The reaction between Cl and butanone in
the presence of oxygen leads to the formation of three peroxy
radicals.
CH3C(O)CH2CH3 + Cl + O2(+M)
→ CH3C(O)CH(O2)CH3 + HCl(+M) (R19a)
→ CH2(O2)C(O)CH2CH3 + HCl(+M) (R19b)
→ CH3C(O)CH2CH2(O2) + HCl(+M) (R19c)
The peroxy radical formed in reaction R19a reacts with NO to
produce an alkoxy radical that subsequently decomposes to
form acetaldehyde.
The vinyl radical generated in reaction R22 will produce CO2
with a yield of 1; thus the yield of CO2 is expected to be double
the branching ratio for reaction R21a. A previous FTIR product
yield study of the chlorine atom initiated oxidation of MAC in
the absence of NOx measured yields of CO, chloroacetone and
CH3C(O)CH(O2)CH3 + NO
→ CH3C(O)CH(O)CH3 + NO2
→ CH3C(O) + CH3CHO
(R20)
The alkoxy radicals generated from reaction of NO with the
peroxy radicals produced in reactions R19b and R19c have
been studied theoretically by Ferenac et al.28 Their study shows
that the major fate of the alkoxy radical from R19b,
OCH2C(O)CH2CH3, is decomposition, which likely leads
mainly to formation of acetaldehyde, CH2O and CO2. The
Ferenac et al. study suggests that the alkoxy radical formed via
channel R19c, CH3C(O)CH2CH2O, mainly undergoes reac-
tion with O2, with lesser contributions from isomerization and
decomposition, leading mainly to multifunctional species that
are not identified with the analytical techniques used in this
work. Because it is expected that the fraction of the Cl/
butanone reaction occurring via (R19b) is quite small,29 the
yield of acetaldehyde can then be equated with the branch-
ing ratio to pathway R19a, i.e., 0.64 0.04. This branching
ratio has been previously determined twice by measuring
the yield of 3-chlorobutanone from the photolysis of mix-
tures of chlorine, butanone and nitrogen. The literature
values reported for the branching ratio of (R19a) using this
CO2 of 0.52
0.04, 0.42
0.05, and 0.23
0.02, res-
pectively.31 In a similar experiment, Canosa-Mas et al.32
reported yields for chloroacetone and CO of 0.41 0.04 and
0.75 0.06. The product yields measured in these experiments
are consistent with those measured here, particularly
considering the fact that the previous studies were conducted
in the absence of NOx where yields of CO, CO2, and chloro-
acetone would likely be reduced due to the occurrence of RO2/
RO2 reactions.
The chemistry in these systems is complicated by the
reaction of MAC with OH radicals, which are generated in the
reaction between HO2 and NO and in the radical channel of
the RO2 + HO2 reaction R1c. Acuchem calculations indicate
that the steady state concentration of OH may exceed that of
Cl and that under the experimental conditions reported here up
to half of the MAC consumed may result from reaction with
OH radicals rather than Cl-atoms. Addition of OH to the
terminal end of the double bond (equivalent to the Cl-initiated
reaction 21b) is expected to lead to the formation of hydroxy-
acetone via a reaction analogous to reaction R23. This product
is observed with a yield of roughly 10%. H-atom abstraction by
OH is equivalent to reaction R21a and leads to the same set of
reaction products. Acuchem simulations were carried out in
approach are 0.73
agreement with this work.
0.0929 and 0.76
0.07,30 in good
Cl + Methacrolein (MAC). The yields of CO2, CO, and
chloroacetone from the photolysis of MAC/Cl2/NO/air
6271
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
which the abstraction and addition branching ratios for reaction
with both Cl and OH were adjusted. Using branching ratios for
(R21a) and (21b) of 0.2 and 0.8, respectively, and assuming
that the rates of abstraction by and addition of OH are the
same,33 the modeled yields of CO2, CO, and chloroacetone are
0.65, 0.65, and 0.36, respectively, somewhat different from the
measured yields. The model predicts a hydroxyacetone yield of
0.25, substantially higher than the experimental result, but
given the complexity of and uncertainties in the secondary
chemistry, further attempts to refine the model were not made.
Nonetheless, the MPAN yields are in reasonable agreement
with the recent measurements by Kaiser and co-workers,34 who
report branching ratios of 0.24 and 0.47 for reactions R21a and
21b, respectively.
isomerize, ultimately forming multifunctional products that are
not readily analyzed using the instrumentation available in
this study. The yield of formyl chloride, potentially formed
from decomposition of this radical, was less than 2%. Thus,
chloroacetaldehyde is a quantitative marker for R27a, and its
yield of 0.65 is therefore equivalent to the branching ratio for
reaction R27a. Acuchem simulations predict that reaction
with OH radicals accounts for less than 8% of the loss of
MVK, and so this chemistry has a relatively small impact on
the observed product distributions. The yield of glycolalde-
hyde was less than 7%, which is consistent with the estimate
from the model (assuming a 70% yield of glycolaldehyde
from OH + MVK).
RO2 + HO2 Branching Ratios. Acyl Peroxy + HO2. Acyl
peroxy radicals are produced from the reaction between
Cl-atoms and aldehydes in the presence of oxygen.
Cl + Methyl Vinyl Ketone (MVK). Product yields from the
photolysis of MVK/Cl2/NO/air mixtures were measured by
FTIR. The measured yields of chloroacetaldehyde, carbon
dioxide, formaldehyde, and peroxyacetyl nitrate are 0.65 0.04,
0.66
0.07, 0.30
0.03, and 0.19
0.03, respectively.
Chloroacetaldehyde is not commercially available and was
synthesized in situ by the photolysis of mixtures containing
ethene, chlorine, and air. Chlorine atoms add to the ethene
double bond resulting in the formation of (2-chloroethyl)-
peroxy radicals (reaction R24), which then react with peroxy
radicals (reaction R25) and oxygen (reaction R26) to produce
chloroacetaldehyde.
Cl + CH2CH2(+O2 + M) → CH2ClCH2O2
CH2ClCH2O2 + RO2 → CH2ClCH2O + RO + O2
(R24)
The major reaction pathways and products for the acylperoxy
radicals studied in this work are summarized in Figure 2. In the
absence of NOx, these peroxy radicals may react either with
other organic peroxy radicals (reaction R30) or with hydro-
peroxy radicals (reaction R1′).
(R25)
CH2ClCH2O + O2 → CH2ClCHO + HO2
(R26)
As the reaction proceeds, the chloroacetaldehyde generated
in reactions R24−R26 reacts with chlorine to form secondary
products. However, in the early stages of the reaction,
chloroacetaldehyde is the only product observed in the IR
spectrum. A previous study of the Cl + MVK reaction in the
absence on NOx from this laboratory reported a yield for
chloroacetaldehyde of 0.75 0.08,31 in good agreement with
this work.
The Cl + MVK reaction pathways are summarized in Figure 1d.
The reaction is expected to proceed via Cl-atom addition to the
double bond.
The product of reaction R27a will then react with NO to
produce an alkoxy radical that is expected to decompose to
produce chloroacetaldehyde and an acyl radical that either will
form carbon dioxide and formaldehyde or will generate
peroxyacetyl nitrate.
The alkoxy radical formed from the product of reaction R27b
in the corresponding reaction is likely to react with O2 or
For both reactions, one of the channels (reactions R1′c and
R30a) results in the formation of an acyl alkoxy radical that will
6272
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
Figure 2. Reaction pathways leading to the formation of major products from the reaction of carbonyl-containing peroxy radicals with HO2 and
RO2: (a) CH3CH2C(O)O2; (b) CH3(CH2)2C(O)O2 and CH3CH2CH(O2)CHO; (c) CH3C(O)CH2CH2O2; (d) CH3C(O)CH(O2)CH3;
(e) CH3C(CH2)C(O)O2 and CH3C(O2)(CH2Cl)CHO; (f) CH2ClCH(O2)C(O)CH3. Reactions R1′c and R1″c are the initial steps in a sequence
of reactions leading to the products shown.
rapidly decompose to produce carbon dioxide and an alkyl
peroxy radical (reaction R31).
both RO2 + RO2 and RO2 + HO2 reactions. However, if the
HO2 concentration is sufficiently high, the peroxy radicals
generated in reactions R29a and R29b react almost exclusively
with HO2. Under these conditions, the relative yields of the
peracid, carboxylic acid, and CO2/alkyl hydroperoxide should
be the same as the branching ratios for reactions R1′a, R1′b, and
R1′c, respectively. Further, the sum of product yields from these
three channels should equal the branching ratio for H-atom
abstraction from the acyl group by chlorine atoms. In practice,
this limiting behavior cannot be reached because it requires
extremely high HO2 precursor concentrations, and the
branching ratios are determined by using the Acuchem model
to describe the dependence of the product yields on the HO2
concentration. The key chemical reactions used are summarized
in Table 1. The branching ratios for the two acylperoxy radicals
studied (propionyl peroxy and butanoyl peroxy radicals) are sum-
marized in Table 3, and the experimental results are discussed
in detail below.
The alkylperoxy radicals formed in reaction R31 may also react
with RO2 or HO2 radicals. The reaction with organic peroxy
radicals will lead to the formation of alcohols and carbonyls,
whereas reaction with hydroperoxy radicals is expected to
generate an alkyl hydroperoxide in high yield (reaction R32).
At lower HO2 precursor concentrations, the product
distribution is complex as the observed products result from
6273
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
Table 3. Experimental Branching Ratios for Carbonyl-Containing RO2 + HO2 Reactions
branching ratios
reaction
YR1a
YR1b
YR1c
ref
0.75 0.13
0.25 0.13
this work
CH3C(O)CH2O2 + HO2
0.33 0.10
nm
0.67 0.20
0.15 0.08
0.15
14
16
nm
17
0.35 0.1
0.25 0.1
0.29 0.1
0.4 0.1
this work
C2H5C(O)O2 + HO2
0.24 0.15
0.47 0.15
this work
C3H7C(O)O2 + HO2
0.42 0.1
nm
0.58 0.1
this work
17
CH3C(O)CH(O2)CH3 + HO2
a
nm
0.2
0.2 0.1
0.8 0.2
0.8 0.2
0.4 0.16
this work
CH2ClCH(O2)C(O)CH3 + HO2
CH2ClC(CH3)(O2)CHO + HO2
CH3C(O)O2 + HO2
0.2 0.2
this work
14
0.4 0.16
0.2 0.08
0.38 0.12
0.12 0.04
0.43 0.10
0.5 0.2
0.2
15
17
17
nm
nm
nm
nm
C6H5C(O)O2 + HO2
a
Measurement is for all RO2 radicals generated from Butanone + Cl. nm = not measured
α-Carbonyl Peroxy + HO2. α-Carbonyl peroxy radicals are
produced from the reaction of ketones with chlorine atoms in
the presence of oxygen (reaction R33a).
The α-carbonyl alkoxy radical produced in reactions R34a and
R1″c will fragment to produce a carbonyl compound and an
acyl peroxy radical (reaction R35).
At high HO2 concentrations, the relative yields of the
hydroperoxide and aldehyde/ketone produced in reactions
R1″a and R1″c (followed by reaction R35), respectively, are the
same as the branching ratios for these reactions. The sum of
these product yields is also equal to the yield of α-carbonyl
peroxy radicals produced in reaction R33a. In practice, for the
majority of the compounds investigated, the hydroperoxide
formed in reaction R1″a cannot be quantified. However, the
yield of the aldehyde/ketone formed in reaction R35 under
The major pathways and products for the α-carbonyl peroxy radicals
studied in this work are summarized in Figure 2. The peroxy radical
formed in reaction R33a may react with other organic peroxy
radicals (reaction R34) or with hydroperoxy radicals (reaction R1″).
these conditions is equal to the product of YR33a and YR1″c
.
Because YR33a is known, the branching ratios for reactions R1″a
and R1″c can still be determined, even when the hydroperoxide
yields are not measurable.
In experiments in which an alcohol is not added to the
reaction mixture, the aldehyde/ketone R1R2CO is generated in
reaction R35 following reactions R34a and R1″c. The yield of
this product thus depends on the concentrations of RO2 and
HO2, and the branching ratios for reactions R33a, R34a, and
R1″c. During the reaction, a mixture of different RO2 radicals is
present, but Acuchem simulations indicate that reaction R34 is
dominated by the self-reaction of α-carbonyl peroxy radicals.
The modeled product yields are therefore sensitive to the
branching ratio for this RO2 self-reaction, and modeling enables
the branching ratio for this reaction to be estimated. The
branching ratios for the RO2 + HO2 reactions are summarized
in Table 3, the calculated branching ratios for RO2 self-reaction
are given in Table 4, and individual systems studied are
discussed in detail in the following sections.
CH3CH2C(O)O2 + HO2. Product yields from the photolysis
of Cl2/propanal/methanol/air mixtures were measured using
a combination of FTIR (ethyl hydroperoxide, CO2, and
6274
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
Table 4. Experimental Branching Ratios for Selected RO2 + RO2 Reactions
reaction
branching ratio
YR34a = 0.66 0.15
2CH3C(O)CH(O2)CH3 = 2CH3C(O)CH(O)CH3 + O2
2CH3C(O)CH(O2)CH3 = other products
YR34b = 0.34 0.15
YR34a = 1.0 0.1
YR34b = 0
2CH2ClC(CH3)(O2)C(O)H = 2CH2ClC(CH3)(O)C(O)H + O2
2CH2ClC(CH3)(O2)C(O)H = other products
2CH2ClCH(O2)C(O)CH3 = 2CH2ClCH(O)C(O)CH3 + O2
2CH2ClCH(O2)C(O)CH3 = other products
YR34a = 1.0 0.2
YR34b = 0
Figure 3. Comparison of experimental and model product yields from the chlorine atom initiated oxidation of propanal/methanol mixtures. Filled
symbols represent FTIR measurements, and open symbols represent HPLC measurements. All model runs shown used the optimized branching
ratios for reaction R30 (see text for details).
propanoic acid) and HPLC (ethyl hydroperoxide and
perpropanoic acid). The measured yields are shown in
Figure 3. As [methanol]0:[propanal]0 increases from 0:1 to
4:1, the yields of ethyl hydroperoxide, propanoic acid, and
perpropanoic acid increase, while the carbon dioxide yield falls.
The reaction of propanal with Cl-atoms proceeds primarily via
the abstraction of the acyl hydrogen followed by addition of
oxygen to form propionyl peroxy radicals (reaction R29a,
where R1 = CH3 and R2 = H). In the absence of HO2 radicals,
the propionyl peroxy radicals primarily react with other organic
peroxy radicals to form the propionoxy radicals (reaction
R30a), which readily decompose to form carbon dioxide and
ethyl peroxy radicals (reaction R31).
The major reactions of the propionyl peroxy radicals are
summarized in Figure 2a. Propionyl peroxy radicals may react
with HO2 to form perpropanoic acid, propanoic acid, and
propionoxy radicals (reactions R1′a−R1′c, where R1 = CH3 and
R2 = H). Ethyl hydroperoxide is generated in the reaction
between ethyl peroxy radicals and HO2 (reaction R32). On the
basis of this mechanism, perpropanoic acid and propanoic acid
are markers for reactions R1′a and R1′b, respectively. The CO2
and ethyl hydroperoxide yields should approach the yield of
reaction R1′c as the methanol (and therefore HO2) concen-
tration increases.
The measured product yields were simulated using an
Acuchem model, as shown in Figure 3. Branching ratios of
0.35 0.1, 0.25 0.1, and 0.4 0.1 for YR1′a, YR1′b, and YR1′c
,
respectively, are obtained in this analysis. We are unaware of
any previous measurements of these branching ratios for
comparison.
CH3(CH2)2C(O)O2 + HO2. Product yields were measured as a
function of [methanol]0/[butanal]0 using a combination of
FTIR (CO2 and butanoic acid), HPLC (propyl hydroperoxide),
GC-FID (propanal), and GC−MS (butanoic acid), as shown in
Figure 4. The yields of propyl hydroperoxide, butanoic acid,
and perbutanoic acid increase, while the yields of CO2 and
propanal decrease as the initial concentration of methanol
increases. The chemistry of the butyroxy peroxy radicals is
summarized in Figure 2b. The behavior of propyl hydro-
peroxide, butanoic acid, perbutanoic acid, and CO2 can be
qualitatively explained by reactions R29a−R32 (where R1 =
C2H5 and R2 = H). Propanal can be generated in two ways. In
the absence of HO2, the propyl peroxy radicals formed in
reaction R31 can react with peroxy radicals to form propoxy
6275
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
Figure 4. Comparison of experimental and model product yields from the chlorine atom initiated oxidation of butanal/methanol mixtures. Filled
symbols represent FTIR measurements, open symbols represent HPLC measurements, half-filled symbols represent GC-FID (for propanal) and
GC−MS (for butanoic acid) measurements. All model runs shown used the optimized branching ratios for reaction R30 (see text for details).
radicals, which subsequently react with oxygen to form
propanal.
carbon atoms, it is possible that the true perbutanoic acid yield
is higher than the value measured in these experiments and the
branching ratio for reaction R1′a should be considered a lower
limit. However, we are unaware of any literature values for
these branching ratios for comparison.
CH3CH2CH2O2 + RO2 → CH3CH2CH2O + RO + O2
(R36)
CH3C(O)CH2O2 + HO2. The yields of CO2 and 13CO2
measured by FTIR from the photolysis of Cl2/methanol/
acetone and Cl2/methanol/13C-2-acetone mixtures, respec-
tively, are shown in Figure 5. The CO2 yield decreases from
0.7 in the absence of methanol to 0.36 at higher methanol
concentrations. These measurements are consistent with the
CO2 yields previously determined in this lab.14 In contrast, the
13CO2 yield from the labeled acetone decreases from 0.55 in the
absence of methanol to 0.13 for [CH313C(O)CH3]0:[CH3OH]0 =
2:1. In the presence of methanol, unlabeled CO2 is also observed
with a yield of 0.35.
Relative yields of methyl hydroperoxide (MHP), peracetic
acid (PAA), and acetonyl hydroperoxide (AHP) are also shown
in Figure 5. In the absence of methanol, the MHP yield is 50%
higher than that of AHP, but when methanol is added to the
reaction mixture, the ratio YMHP/YAHP falls to 0.5. Similarly,
CH3CH2CH2O + O2 → CH3CH2CHO + HO2
(R37)
As the initial methanol concentration is increased, the propyl
peroxy radicals will preferentially react with HO2 to form
propyl hydroperoxide (reaction R32), and propanal formation
via reactions R36 and R37 will become less significant. Propanal
may also be formed via H-atom abstraction by chlorine atoms
at the α-position (reaction R12b), followed by reaction with
other peroxy radicals to form the corresponding alkoxy radical
(reaction R38), which can then decompose via reaction R14b.
CH3CH2CH(O2)C(O)H + RO2
→ CH3CH2CH(O)C(O)H + RO + O2
(R38)
The 2-peroxy butanal radicals produced in reaction R12b may
also react with HO2, which may result in the formation of
2-hydroperoxy butanal.
Y
PAA/YAHP is 0.8 in the absence of methanol, but falls to 0.26
when [methanol]0/[acetone]0 is 0.25 or higher.
The results of Acuchem simulations for this chemical system
are shown in Figure 4. The optimal fits to the data give
branching ratios for reactions R1′a−R1′c of 0.24 0.15, 0.29
0.1, and 0.47 0.15, respectively. Because the HPLC system is
not optimized for hydroperoxides containing more than three
The chemistry of the acetonyl peroxy radical is summarized
in Figure 2c). The reaction proceeds via H-atom abstraction
followed by oxygen addition to form acetonyl peroxy radicals
(reaction R33a, where R1 = R2 = H and R3 = CH3). AHP is
6276
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
Figure 5. Comparison of experimental and model product yields from the chlorine atom initiated oxidation of acetone/methanol mixtures. Filled
symbols represent FTIR measurements, and open symbols represent HPLC measurements. All model runs shown used the optimized branching
ratios for reaction R34 (see text for details).
formed from the reaction of this peroxy radical with HO2 in
reaction R1″a. Alternatively, acetonoxy radicals can be formed
from the reaction of acetonyl peroxy radicals with organic
peroxy radicals (reaction R34a) or with hydroperoxy radicals
(reaction R1″c). This alkoxy radical then decomposes
unimolecularly via reaction R35 to form formaldehyde and
acetyl peroxy radicals. The acetyl peroxy radicals may then form
peracetic acid or acetic acid by reacting with HO2 in reactions
R1′a and R1′b, respectively. Alternatively, they may form
acetoxy radicals by reacting with RO2 (reaction R30a) or with
HO2 (reaction R1′c), which will result in the formation of
carbon dioxide and MHP via reactions R31 and R32.
The branching ratio for reaction R1′c is 0.4 for acetyl peroxy
radicals,14 and so the CO2 yield at high methanol concentra-
tions is expected to be lower than this value. In the unlabeled
experiments, the high methanol CO2 yield of 0.36 suggests that
reaction R1″ occurs almost exclusively through channel R1″c.
However, the large increase in the acetonyl hydroperoxide
yield relative to peracetic acid and methyl hydroperoxide is
consistent with a large branching ratio for reaction R1″a.
The labeled acetone experiments provide an explanation for
the discrepancy. The much lower 13CO2 yield in the presence
of methanol implies that reaction R1″a is the primary pathway
for the acetonyl peroxy + HO2 reaction. In these experiments,
there is no clear primary source of unlabeled CO2, suggesting
that a significant fraction of the CO2 in the unlabeled
experiments is not formed from acetone chemistry. Additional
experiments were carried out in which nitrogen/oxygen and
nitrogen/oxygen/chlorine mixtures were photolyzed within the
chamber. In the latter experiment, approximately 3 × 1012
molecules cm−3 of CO2 was formed on the time scale of a
typical acetone experiment. The results are consistent with a
reaction between Cl-atoms and the reactor walls as a source of
carbon dioxide in these experiments. Because acetone reacts
relatively slowly with chlorine atoms, typically less than 10% of
the acetone has reacted at the end of experiments in which
methanol is present. Thus the relatively small amount of CO2
produced in this heterogeneous reaction translates into an
apparently large product yield in these experiments.
The product yields measured in these experiments were
fitted to the outputs of a detailed chemical model with the
branching ratios for reactions R1″a and R1″c treated as
adjustable parameters. The results of this analysis are shown
in Figure 5. 13CO2 yields from 13C-acetone are consistent with
YR1″a:YR1″c = 0.8 0.1:0.2 0.1. The relative yields of MHP,
PAA, and AHP suggest a slightly higher branching ratio for the
radical channel (YR1″a:YR1″c = 0.7
0.15:0.3
0.15). The
branching ratios are reported as the average of these two values
6277
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
in Table 3 (0.75 0.13 and 0.25 0.13 for reactions R1″a and
R1″c, respectively).
does) and it is not a source of acetaldehyde (as simple alcohols
such as ethanol and isopropanol are).
Key reactions leading to the formation of observed products
are summarized in Figure 2d. As shown in Figure 6, the
acetaldehyde yield is essentially independent of the hydroxy-
acetone concentration in these experiments. Acetaldehyde is
formed from the peroxy radical generated in reaction R33a
(where R1 = R3 = CH3 and R2 = H). This species may react
either with an organic peroxy radical (reactions R34a and 34b)
or with a hydroperoxy radical (reactions R1″a and R1″c). The
alkoxy radical formed in reactions R34a and R1″c may then
fragment to form acetaldehyde (reaction R35) and acetyl
peroxy radicals. The acetaldehyde yield at high HO2 concen-
trations is thus indicative of the fraction of peroxy radicals
reacting via pathway R1″c. Note that we again assume the
branching to reaction R19b is small and thus does not
contribute significantly to the formation of acetaldehyde.
HPLC analysis identified peracetic acid and methyl hydro-
peroxide as reaction products both in the absence of HO2
precursor and at high HO2 precursor concentrations. This
observation is consistent with the production of acetyl peroxy
radicals via reaction R1″c followed by reaction R35. No other
hydroperoxides, such as the species generated in reaction R1″a
are observed. Although it is possible that other hydroperoxides
are not formed in these reactions, it seems more likely that
organic hydroperoxides containing more than three carbon
atoms are too nonpolar to be detected with the HPLC using
dilute acid as the eluent.
There are three previous measurements of the branching
ratios for this reaction. Jenkin et al.16 photolyzed mixtures of
acetone, methanol, chlorine and benzene in air. The researchers
measured phenol production, which is generated in the reaction
between OH and benzene,and determined the branching ratio
for reaction R1″c using a chemical model. Dillon and Crowley17
used laser induced fluorescence (LIF) to directly monitor OH
concentrations in flash photolysis experiments. Both studies
measured YR1″a:YR1″c of 0.85:0.15, in excellent agreement with
this work. However, a previous study from this laboratory14
using the same experimental approach as in this work con-
cluded that YR1″a:YR1″c is 0.33:0.67. Two experimental observa-
tions were used to reach this conclusion. First, in the presence
of methanol, a relatively large CO2 yield was observed. As
discussed above, most of this CO2 can be attributed to
heterogeneous chemistry rather than reactions involving
acetone. Second, the measured yield of acetonyl hydroperoxide
was low compared to those for peracetic acid and methyl
hydroperoxide. This result is in contrast to the measurements
in this work, in which the acetonyl hydroperoxide yield is much
higher than that of the other hydroperoxides. It is likely that a
contaminated HPLC column in the original experiments may
have led to the loss of acetonyl hydroperoxide during the
analysis, resulting in the reported observations. This destruction
of larger organic hydroperoxides by degraded HPLC columns
has been observed in our lab previously.
The results of chemical simulations carried out using
Acuchem are shown in Figure 6 using the chemical scheme
outlined in Table 1. The data are consistent with branching
ratios for reactions R1″c and R34a of 0.58 0.1 and 0.66
0.15, respectively, as reported in Tables 1 and 4, respectively.
Dillon and Crowley17 measured the branching ratio for R1″c for
the reaction of HO2 with the peroxy radicals generated from
the reaction of chlorine atoms with butanone in the presence of
oxygen using flash photolysis with LIF spectroscopy to monitor
OH concentrations. The branching ratio for R1″c is reported to
be about 0.2. If the assumption is made that only α-carbonyl
peroxy radicals generate OH in this reaction, the corresponding
branching ratio predicted in Dillon and Crowley's experiments
using the data from this study would be the product of the
CH3C(O)CH(O2)CH3 + HO2. Acetaldehyde yields were
measured from the photolysis of chlorine/butanone/hydrox-
yacetone/air mixtures as a function of [hydroxyacetone]0/
[butanone]0, as shown in Figure 6. Hydroxyacetone was used as
branching ratio for reaction R19a (0.64
0.04) and the
branching ratio for reaction R1″c (0.58
0.1), which gives
0.37 0.07. The branching ratios reported here are thus not in
quantitative agreement with that study. However, both studies
are consistent with a significant radical yield in this reaction.
CH2ClC(CH3)(O2)CHO + HO2. Typical product yield plots for
CO, CO2, and chloroacetone from the photolysis of MAC/
isopropanol/Cl2/air mixtures are shown in Figure 7a. Product
yields measured from all experiments are shown in Figure 7b.
The yields of these products do not change within the un-
certainties of the measurements as the initial concentration of
the HO2 precursor increases. Formaldehyde was also observed
with a yield of approximately 20% in all of the experiments.
Analysis of the reaction products by HPLC shows the
formation of AHP, PAA, and MHP in the presence of
isopropanol. PAA and MHP should be formed from acetyl
radicals formed from the methylvinyl radical produced in the
abstraction channel, (R21a). Because isopropanol reacts with
Cl-atoms to form acetone, it is likely that the observed acetonyl
hydroperoxide is a secondary product of isopropanol oxidation
and that hydroperoxides from MAC reactions are not observed.
Figure 6. Comparison of experimental and model product yields from
the chlorine atom initiated oxidation of butanone/hydroxyacetone
mixtures using GC-FID. All model runs shown used the optimized
branching ratios for reaction R34 (see text for details).
the HO2 source because it does not interfere with the
acetaldehyde signal in the GC-FID chromatogram (as methanol
6278
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
experiments, which is consistent with the 10% hydroxyacetone
from the Cl-atom initiated oxidation of MAC in the presence of
NO, suggesting that similar levels of OH are present in these
NO-free experiments.
Analysis of the reaction system is complicated by the
importance of the secondary OH-initiated chemistry occurring.
The three major peroxy radicals expected in these reactions are
the species resulting from Cl addition to MAC from R21b (Cl−
RO2), the corresponding peroxy radical from OH addition to
MAC (OH−RO2), and the acyl peroxy radical generated from
H-abstraction by either Cl (R21a) or OH. The observed
product distributions are potentially influenced by the
branching ratios for the reaction of all three peroxy radicals
with HO2. Because of the large number of variable parameters
in the model, a quantitative determination of all of the RO2 +
HO2 branching ratios in this system is not possible. Instead,
four scenarios were considered as shown in Figure 7b. In the
first simulation (run 1), the branching ratios for all three peroxy
radicals reacting with HO2 were assumed to be 0.2 and 0.8 for
(R1a) and (R1c), respectively. In run 2, YR1a and YR1c for Cl−
RO2 and OH−RO2 were set to the same values as in run 1, but
the YR1c for the acyl peroxy radical was reduced to zero. In run
3, YR1c was set to zero for Cl−RO2 and OH−RO2, whereas YR1a
and YR1c were restored to 0.2 and 0.8, respectively, for the acyl
peroxy radical. Finally, in run 4, YR1c was set to zero for all three
of the organic peroxy radicals.
As shown in Figure 7b, runs 1 and 2 both reproduce the
observed chloroacetone and CO yields well. The common
element in these simulations is the large radical channel (YR1c
)
for the Cl−RO2 and OH−RO2 radicals. Although the CO and
chloroacetone yields are insensitive to the branching ratios for
the acyl peroxy radical, run 1 does reproduce the magnitude of
the experimental CO2 better than run 2, implying that (R1c) is
also important for this organic peroxy radical. However,
because the acyl radical is the minor channel, the results are
not very sensitive to that branching ratio.
A possible alternative explanation for the observed products
is that the peroxy radical formed in reaction R21b (and the
analogous OH + MAC addition reaction) could rearrange and
decompose as shown in reaction R39.35 Because reaction R21b
is expected to be quite exothermic, these peroxy radicals could
undergo this rearrangement before they are collisionally stabilized.
Figure 7. Product yields from the chlorine atom initiated oxidation of
MAC/isopropanol mixtures using FTIR. (a) Typical product yield
plots. Lines are least-squares fits to all point. (b) Comparison of
experimental and model product yields (see text for details).
Though the Cl + acetone reaction is slow, the relatively low
concentrations of peroxides observed are consistent with
formation via this pathway.
This mechanism has been proposed to explain some of the OH
radical formation from ozone alkene reactions.36 Quantum
chemical and Master equation calculations on structurally
similar C2 and C3 peroxy radicals indicate that this rearrange-
ment reaction will not be significant at atmospheric pressure
and that collisional deactivation followed by bimolecular reac-
tion with organic peroxy or hydroperoxy radicals is the major
reaction pathway.36,37 Because the peroxy radical produced in
reaction R19b contains four carbon atoms, it should be
stabilized even more rapidly, and so reaction R39 is not
expected to be significant under the reaction conditions used in
this work.
Key reactions leading to observed reaction products are
summarized in Figure 2e). Chloroacetone and CO are pro-
duced via addition of Cl-atoms to MAC in the presence of
oxygen (reaction R21b). The resultant peroxy radicals are then
converted via reaction with RO2 or HO2 to alkoxy radicals
(reactions R34a and R1″c, respectively where R1 = CH3, R2 =
CH2Cl, and R3 = H), which subsequently decompose via
reaction R35 to form chloroacetone and HCO. The HCO frag-
ment reacts with oxygen to form the observed carbon mono-
xide. CO and hydroxyacetone are produced via an analogous
set of reactions initiated by the addition of OH to MAC. The
yield of CO is about 10% higher than chloroacetone in these
CH2ClCH(O2)C(O)CH3 + HO2. Chloroacetaldehyde product
yields were monitored as a function of [isopropanol]0/[MVK]0
in the absence of NOx and are shown in Figure 8. In the
6279
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
The α-carbonyl peroxy radicals studied in this work show
large variations in the branching ratios. For the small number of
species studied here, there is an increase in the branching ratio
for reaction R1″c from primary peroxy radicals (CH3C(O)-
CH2O2, YR1″c = 0.25) to secondary and tertiary peroxy radicals
(CH3C(O)CH(O2)CH3 and CH2ClCH(O2)C(O)CH3 and
CH2ClC(CH3)(O2)CHO, YR1″c = 0.6, 0.8, and 0.8, respec-
tively). This observation suggests that the degree of
substitution at the peroxy carbon center may play a role in
determining the branching ratios. Additional measurements are
clearly needed to better characterize the relationship between
radical structure and the branching ratios.
This study suggests that radical recycling via reactions R1′c
and R1″c is likely a general feature of all acyl peroxy and
α-carbonyl peroxy radicals, respectively. This implies that these
reactions may have a significant impact on radical concen-
trations in the troposphere under certain conditions. In
particular, we note that MAC and MVK are first generation
reaction products from the photo-oxidation of isoprene.
Because the peroxy radicals produced via the reaction of
MAC and MVK with OH and Cl are structurally very similar, it
is likely that the atmospheric reactions of MAC and MVK lead
to the formation of peroxy radicals with substantial branching
ratios for reaction R1″c. Therefore, this chemistry may help to
at least partially explain the unexpectedly high radical
concentrations that were measured in an isoprene-rich
environment in a recent field campaign.38
Figure 8. Comparison of experimental and model product yields from
the chlorine atom initiated oxidation of MVK/isopropanol mixtures
using GC-FID. All model runs shown used the optimized branching
ratios for reaction R34 (see text for details).
ASSOCIATED CONTENT
■
absence of isopropanol, a chloroacetaldehyde yield of 0.55 is
observed. As the HO2 precursor is added to the reaction mix-
ture, the product yield is observed to decrease slightly to 0.50.
Pathways leading to the formation of chloroacetaldehyde are
summarized in Figure 2f. Chloroacetaldehyde is produced via
the peroxy radical formed in reaction R27a. This species then
generates an alkoxy radical by reaction with RO2 or HO2
(reactions R34a and R1″c, respectively, where R1 = CH2Cl,
R2 = H, and R3 = CH3), followed by unimolecular decomposi-
tion (reaction R35). PAA and MHP are observed in the HPLC
chromatograms of all of the experiments. When isopropanol is
present in the reaction mixture, AHP is also observed at
concentrations that are several times smaller than that of PAA
and MHP. Because the yield of AHP relative to PAA and MHP
from acetone oxidation is expected to be greater than 1, the
observed PAA and MHP from experiments where [isopropa-
nol]0/[MVK]0 is greater than 0 are consistent with the forma-
tion of acetyl peroxy radicals from reaction R1″c followed
by R35.
S
* Supporting Information
Typical infrared spectra obtained in experiments to investigate
the Cl-atom initiated reactions of 2-13C acetone, methacrolein,
and methyl vinyl ketone in both the presence and absence of an
HO2 source. This information is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: 559-278-2420. Fax: 559-278-4402. E-mail: ahasson@
csufresno.edu.
Present Addresses
§UC Davis School of Medicine, 4610 X Street, Sacramento
California 95817, United States.
∥Department of Chemistry and Biochemistry, 607 Charles E.
Young Drive East, Box 951569, UCLA, Los Angeles California
90095, United States.
⊥UC Irvine School of Medicine, 1001 Health Sciences Road,
Irvine California 92697, United States.
Acuchem simulations were used to determine the branching
ratios for reactions R1″c and R34a using the reaction scheme
shown in Table 1. As shown in Figure 8, the optimal fit
corresponds to branching ratios for reactions R1″c and R34a of
0.8 0.2 and 1.0 0.2, respectively.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
SUMMARY AND CONCLUSIONS
This work was supported by the NASA Upper Atmosphere
Research Program (NNG06GE44G) and by a National
Oceanic and Atmospheric Administration award under
cooperative agreement # NA06OAR4810187. A.S.H. thanks
■
Within the mutual uncertainties, the branching ratios for the
reaction of propionyl peroxy and butanoyl peroxy radicals with
hydroperoxy radicals are the same. Further, these branching
ratios are similar to the literature values for the acetyl peroxy +
HO2 reaction. Collectively, these data suggest that the radical
channel (reaction R1′c) is generally important for linear acyl
peroxy radicals, and that the branching ratios for reactions
R1′a−R1′c are relatively insensitive to the carbon chain length.
́
NCARs Advanced Study Program for the award of a Faculty
Fellowship that enabled some of this work to be performed.
The National Center for Atmospheric Research is operated by
the University Corporation for Atmospheric Research, under
the sponsorship of the National Science Foundation.
6280
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281

The Journal of Physical Chemistry A
Article
(36) Kuwata, K. T.; Templeton, K. L.; Hasson, A. S. J. Phys. Chem. A
2003, 107, 11525−11532.
REFERENCES
■
(1) Atkinson, R. J. Phys. Chem. Ref. Data 1997, 26, 215.
(2) Wallington, T. J.; Japar, S. M. Chem. Phys. Lett. 1990, 167, 513−
518.
(37) Kuwata, K. T.; Hasson, A. S.; Dickinson, R. V.; Petersen, E. B.;
Valin, L. C. J. Phys. Chem. A 2005, 109, 2514−2524.
(38) Lelieveld, J.; Butler, T. M.; Crowley, J. N.; Dillon, T. J.; Fischer,
H.; Ganzeveld, L.; Harder, H.; Lawrence, M. G.; Martinez, M.;
Taraborrelli, D.; Williams, J. Nature 2008, 452, 737−740.
(39) Orlando, J. J.; Tyndall, G. S.; Vereecken, L.; Peeters, J. J. Phys.
Chem. A 2000, 104, 11578−11588.
(3) Wallington, T. J.; Japar, S. M. Chem. Phys. Lett. 1990, 166, 495−
499.
(4) Lightfoot, P. D.; Roussel, P.; Caralp, F.; Lesclaux, R. J. Chem. Soc.,
Faraday Trans. 1991, 87, 3213−3220.
(5) Wallington, T. J. J. Chem. Soc., Faraday Trans. 1991, 87, 2379−
2382.
(40) 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, 12157−12182.
(41) Jenkin, M. E.; Boyd, A. A.; Lesclaux, R. J. Atmos. Chem. 1998, 29,
267−298.
(6) Spittler, M.; Barnes, I.; Becker, K. H.; Wallington, T. J. Chem.
Phys. Lett. 2000, 321, 57−61.
(7) Elrod, M. J.; Ranschaert, D. L.; Schneider, N. J. Int. J. Chem. Kinet.
2001, 33, 363−376.
(8) Jackson, A. V.; Hewitt, C. N. Crit. Rev. Environ. Sci. Technol. 1999,
29, 175−228.
(9) Niki, H.; Maker, P. D.; Savage, C. M.; Breitenbach, L. P. J. Phys.
Chem. 1985, 89, 588−591.
(10) Moortgat, G. K.; Veyret, B.; Lesclaux, R. Chem. Phys. Lett. 1989,
160, 443−447.
(11) Horie, O.; Moortgat, G. K. J. Chem. Soc., Faraday Trans. 1992,
88, 3305−3312.
(12) Crawford, M. A.; Wallington, T. J.; Szente, J. J.; Maricq, M. M.;
Francisco, J. S. J. Phys. Chem. A 1999, 103, 365−378.
(13) Tomas, A.; Villenave, E.; Lesclaux, R. J. Phys. Chem. A 2001,
105, 3505−3514.
(14) Hasson, A. S.; Tyndall, G. S.; Orlando, J. J. J. Phys. Chem. A
2004, 108, 5979−6989.
(15) Jenkin, M. E.; Hurley, M. D.; Wallington, T. J. Phys. Chem.
Chem. Phys. 2007, 24, 3149−3162.
(16) Jenkin, M. E.; Hurley, M. D.; Wallington, T. J. Phys. Chem.
Chem. Phys. 2008, 29, 4274−4280.
(17) Dillon, T. J.; Crowley, J. N. Atmos. Chem. Phys. 2008, 8, 4877−
4889.
(18) Hasson, A. S.; Kuwata, K. T.; Arroyo, M. C.; Petersen, E. B. J.
Photochem. Photobiol. A 2005, 176, 218−230.
(19) Braun, W.; Herron, T. J.; Kahaner, D. K. Int. J. Chem. Kinet.
1988, 20, 51−62.
(20) Hellpointner, E.; Gab, S. Nature 1989, 337, 631.
̈
(21) Orzechowska, G. E.; Paulson, S. E. J. Phys. Chem. A 2005, 109,
5358−5365.
(22) Atkinson, R.; Cox, R. A.; Crowley, J. N.; Hampson, R. F.; Hynes,
R. G.; Jenkin, M. E.; Kerr, J. A.; Rossi, M. J.; Troe, J. Atmos. Chem.
Phys. 2006, 6, 3625−4055.
(23) Atkinson, R. Chem. Rev. 1986, 85, 69−201.
(24) Monedero, E.; Salgado, M. S.; Villanueva, F.; Martín, P.; Barnes,
I.; Cabanas, B. Chem. Phys. Lett. 2008, 465, 207−211.
̃
(25) Atkinson, R.; Tuazon, E. C.; Aschmann, S. M. Environ. Sci.
Technol. 2000, 34, 623−631.
(26) Singh, S.; Hernandez, S.; Ibarra, Y.; Hasson, A. S. Int. J. Chem.
Kinet. 2009, 41, 133−141.
(27) Wu, H.; Mu, Y. Int. J. Chem. Kinet. 2007, 39, 168−174.
(28) Ferenac, M. A.; Davis, A. J.; Holloway, A. S.; Dibble, T. S. J.
Phys. Chem. A 2002, 107, 63−72.
(29) Iwasaki, E.; Taketani, F.; Takahashi, K.; Matsumi, Y.;
Wallington, T. J.; Hurley, M. D. Chem. Phys. Lett. 2007, 439, 274−279.
(30) Kaiser, E. W.; Wallington, T. J.; Hurley, M. D. J. Phys. Chem. A
2009, 113, 2424−2437.
(31) Orlando, J., J.; Tyndall, G., S.; Apel, E., C.; Riemer, D. D.;
Paulson, S. R. Int. J. Chem. Kinet. 2003, 35, 334−353.
(32) Canosa-Mas, C. E.; Hutton-Squire, H. R.; King, M. D.; Stewart,
D. J.; Thompson, K. C.; Wayne, R. P. J. Atmos. Chem. 1999, 34, 163−
170.
(33) Tuazon, E. C.; Atkinson, R. Int. J. Chem. Kinet. 1990, 22, 591−
602.
(34) Kaiser, E. W.; Pala, I. R.; Wallington, T. J. J. Phys. Chem. A 2010,
114, 6850−6860.
(35) Asatryan, R.; Silva, G. d.; Bozzelli, J. W. J. Phys. Chem. A 2010,
114, 8302−8311.
6281
dx.doi.org/10.1021/jp211799c | J. Phys. Chem. A 2012, 116, 6264−6281