Total radical yields from tropospheric ethene ozonolysis

Article abstract of DOI:10.1039/c0cp02342f

The gas-phase reactions of ozone with alkenes can be significant sources of free radicals (OH, HO₂ and RO₂) in the Earth's atmosphere. In this study the total radical production and degradation products from ethene ozonolysis have been measured, under conditions relevant to the troposphere, during a series of detailed simulation chamber experiments. Experiments were carried out in the European photoreactor EUPHORE (Valencia, Spain), utilising various instrumentation including a chemical-ionisation- reaction time-of-flight mass-spectrometer (CIR-TOF-MS) measuring volatile organic compounds/oxygenated volatile organic compounds (VOCs/OVOCs), a laser induced fluorescence (LIF) system for measuring HO₂ radical products and a peroxy radical chemical amplification (PERCA) instrument measuring HO ₂ + ΣRO₂. The ethene + ozone reaction system was investigated with and without an OH radical scavenger, in order to suppress side reactions. Radical concentrations were measured under dry and humid conditions and interpreted through detailed chemical chamber box modelling, incorporating the Master Chemical Mechanism (MCMv3.1) degradation scheme for ethene, which was updated to include a more explicit representation of the ethene-ozone reaction mechanism. The rate coefficient for the ethene + ozone reaction was measured to be (1.45 ± 0.25) × 10^-18 cm³ molecules ^-1 s^-1 at 298 K, and a stabilised Criegee intermediate yield of 0.54 ± 0.12 was determined from excess CO scavenger experiments. An OH radical yield of 0.17 ± 0.09 was determined using a cyclohexane scavenger approach, by monitoring the formation of the OH-initiated cyclohexane oxidation products and HO₂. The results highlight the importance of knowing the [HO₂] (particularly under alkene limited conditions and high [O₃]) and scavenger chemistry when deriving radical yields. An averaged HO₂ yield of 0.27 ± 0.07 was determined by LIF/model fitting. The observed yields are interpreted in terms of branching ratios for each channel within the postulated ethene ozonolysis mechanism. the Owner Societies.

Full text of DOI:10.1039/c0cp02342f

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
PCCP
Cite this: Phys. Chem. Chem. Phys., 2011, 13, 11002–11015
www.rsc.org/pccp
PAPER
Total radical yields from tropospheric ethene ozonolysisw
a
a
b
c
Mohammed S. Alam, Marie Camredon,z Andrew R. Rickard, Timo Carr,
a
c
c
c
Kevin P. Wyche, Karen E. Hornsby, Paul S. Monks and William J. Bloss*
Received 1st November 2010, Accepted 8th March 2011
DOI: 10.1039/c0cp02342f
The gas-phase reactions of ozone with alkenes can be significant sources of free radicals
(
2 2
OH, HO and RO ) in the Earth’s atmosphere. In this study the total radical production and
degradation products from ethene ozonolysis have been measured, under conditions relevant to
the troposphere, during a series of detailed simulation chamber experiments. Experiments were
carried out in the European photoreactor EUPHORE (Valencia, Spain), utilising various
instrumentation including a chemical-ionisation-reaction time-of-flight mass-spectrometer
(
(
CIR-TOF-MS) measuring volatile organic compounds/oxygenated volatile organic compounds
VOCs/OVOCs), a laser induced fluorescence (LIF) system for measuring HO radical products
2
and a peroxy radical chemical amplification (PERCA) instrument measuring HO + SRO . The
2
2
ethene + ozone reaction system was investigated with and without an OH radical scavenger, in
order to suppress side reactions. Radical concentrations were measured under dry and humid
conditions and interpreted through detailed chemical chamber box modelling, incorporating the
Master Chemical Mechanism (MCMv3.1) degradation scheme for ethene, which was updated to
include a more explicit representation of the ethene–ozone reaction mechanism.
The rate coefficient for the ethene + ozone reaction was measured to be (1.45 ꢀ 0.25) ꢁ
ꢂ18
3
cm molecules s
ꢂ1 ꢂ1
1
0
at 298 K, and a stabilised Criegee intermediate yield of 0.54 ꢀ 0.12 was
determined from excess CO scavenger experiments. An OH radical yield of 0.17 ꢀ 0.09 was
determined using a cyclohexane scavenger approach, by monitoring the formation of the
OH-initiated cyclohexane oxidation products and HO
2
. The results highlight the importance
of knowing the [HO ] (particularly under alkene limited conditions and high [O ]) and scavenger
3
2
chemistry when deriving radical yields. An averaged HO yield of 0.27 ꢀ 0.07 was determined
2
by LIF/model fitting. The observed yields are interpreted in terms of branching ratios for each
channel within the postulated ethene ozonolysis mechanism.
2
1
. Introduction
in urban environments. However, another key fate for
alkenes is reaction with ozone, which may form the dominant
loss pathway depending on local conditions and the time of
Methyl-substituted ethenes (C –C ) are emitted to the atmo-
2
6
sphere from a wide range of sources, which include combus-
tion processes and vehicle emissions. They are significant
primary pollutants in the boundary layer, contributing to the
total volatile organic compound (VOC) emissions in urban air.
Unsaturated VOCs may account for up to 30% of the total
3
day. Alkene ozonolysis leads to the production of a wide
range of functionalized oxygenated products, including carbonyls,
organic acids and hydroperoxides, and also serves as a
1,4
non-photolytic source of OH, HO2 and RO radicals. In
2
addition, the gas-phase ozonolysis of biogenic alkenes, notably
C10 monoterpenes and C15 sesquiterpenes, is known to lead to
the production of semi-volatile oxygenated compounds that
1
OH sink and can contribute significantly to ozone formation
5
,6
a
have been detected in secondary organic aerosol (SOA).
School of Geography, Earth & Environmental Sciences,
General mechanistic details for the reactions of simple unsaturated
hydrocarbons with ozone are thought to be relatively well known.
However, there are quantitative details of the reaction pathways
that need to be determined in order to accurately understand
University of Birmingham, Edgbaston, Birmingham B15 2TT, UK.
E-mail: w.j.bloss@bham.ac.uk
National Centre for Atmospheric Science, School of Chemistry,
University of Leeds, Leeds LS2 9JT, UK
Department of Chemistry, University of Leicester,
Leicester LE1 7RH, UK
b
c
7
their implications for the atmosphere. The role of alkene
w Electronic supplementary information (ESI) available. See DOI:
ozonolysis reactions in the tropospheric HOx (OH + HO2)
budget has been established by a number of studies (ref. 3 and 7
and refs within). For example, during both urban and rural
case studies in Los Angeles and a forested site in Alabama,
1
0.1039/c0cp02342f
z Present address: LISA, UMR CNRS/INSU 7583, Universite
Est Creteil et Universite Paris Diderot, Institut Pierre Simon Laplace,
4010 Creteil Cedex, France.
´
Paris
´
´
9
´
1
1002 Phys. Chem. Chem. Phys., 2011, 13, 11002–11015
This journal is c the Owner Societies 2011

View Article Online
it was illustrated that, under certain conditions, alkene ozonolysis
3
(reaction 2 in Fig. 1), rearrangement through a dioxirane
structure (reaction 3), which can decompose to various products
(3a–3d), possibly via a ‘hot’ acid intermediate, or stabilisation
(reaction 4) followed by bimolecular reactions, e.g. 4a–4c. The
syn-mono and di-methyl substituted CI’s are thought to
decompose predominantly through isomerisation via a five-
membered transition state to give an excited hydroperoxide
species which subsequently decomposes to give OH and a
x
can be the predominant HO source. In an investigation of
the chemistry occurring in a forested region of north-western
Greece, observationally constrained box model simulations
show the significance of the ozonolysis of biogenics as an
8
important OH source. Under polluted urban conditions,
alkene ozonolysis accounted for 46% of the daytime radical
production in the summer and 62% during winter at an urban
9
site in Birmingham, UK. More recently, alkene ozone reac-
7
vinoxy radical. However, even in the absence of the so called
tions were reported to be indirectly responsible for 33% of the
peak radical production found during the day in Mexico
‘‘hydroperoxide’’ mechanism, OH production is observed,
with the decomposition/isomerisation channels 2 and 3a
postulated as the likely sources. There are a number of quanti-
tative indirect OH yield measurements from ethene ozonolysis
available in the literature, obtained through the use of radical
1
0
City and, under summertime conditions in Santiago de
Chile, contributed B24% to the total daytime primary OH
1
1
radical source. The radical formation potential from the
ozonolysis of unsaturated hydrocarbons is therefore of signifi-
cant interest as it has a substantial influence on the HO
1
8,19
20,21
tracer
tions reported at pressures between 10–60 Torr and at short
and scavenger
species, plus direct observa-
x
1
,3
22,23
budget in both urban and rural environments.
timescales (ca. 10 ms).
challenging, as the rate coefficient for the ethene + ozone reac-
Such direct observations are
Ethene (ethylene, H CQCH ), the simplest alkene, is widely
2
2
ꢂ18
3
cm molecule
ꢂ1 ꢂ1 24
used in industry, where it is an important precursor to a wide
variety of other organic chemicals. In terms of its mass
emission, ethene is one of the most significant VOCs released
into the environment, found in urban atmospheres at up to
tion (1.59 ꢁ 10
s ) is substantially slower
than that for other alkenes (e.g. ca. 2 orders of magnitude
2
4
slower than cis/trans-2-butene). This, coupled to the low OH
yield, results in a low steady state [OH] which is difficult to
observe directly. The high barrier to reaction calculated for
2
,12
parts per billion (ppbV) levels.
It has been estimated that
2
5
26
exposure to as little as 10 ppbV of ethene for one hour each
week may lead to an increased lifetime risk of cancer amounting
channel 2, together with pressure independent OH yields,
leads to the hypothesis that it is the fragmentation of the ‘hot’
acid (3a) that is the dominant source of OH, rather than
formation via the four-membered transition state (2).
1
3
to approximately 70 per 100 000 persons. In terms of the
ozonolysis reaction mechanism, the simple structure of ethene
means that its reaction with ozone is unique, in that it cannot
The current recommended OH yield from ethene ozonolysis
24
at atmospheric pressure is 0.16. Measurements of HO
2
7
follow OH production via the hydroperoxide mechanism.
The carbonyl oxide [CH
2
OO]* (also known as a Criegee
and/or potential RO
production of these contributors to HO
impact to direct OH production, as much ethene oxidation
occurs in moderate to high NO environments with rapid
NO-mediated radical cycling) is uncertain. Recently, HO
2
production, however, are scarce—
1
4
intermediate), formed from ethene ozonolysis is also a
product formed from the ozonolysis of all terminal alkenes,
including the important monoterpenes b-pinene and limonene,
as well as isoprene, the dominant non-methane hydrocarbon
emitted into the atmosphere. Understanding of the fate and
kinetics of the gas-phase reaction between ethene and ozone
therefore underpins interpretation of experimental measure-
ments of radical production from biogenic alkene ozonolysis.
Gas-phase ethene ozonolysis is believed to proceed via the
x
(which have a similar
x
2
radical measurements from ethene ozonolysis have been
performed using matrix isolation with electron spin resonance
2
7
(MI-ESR)
(PERCA),
and peroxy radical chemical amplification
with yields of 0.39 ꢀ 0.03 and 0.38 ꢀ 0.02
2
8,29
respectively reported at atmospheric pressure, a factor of
24
3 greater than those inferred from indirect studies. Radical
yields have also been indirectly inferred for a range of alkenes,
from long-duration (24–48 h) experiments performed in
1
4
Criegee mechanism, illustrated in Fig. 1. Ozonolysis is
initiated by a concerted electrophilic 1,3-cycloaddition of ozone
across the double bond to form a primary ozonide (POZ) (1a).
This intermediate is high in energy and rapidly decomposes,
giving rise to a vibrationally excited carbonyl oxide and a
stable (primary) carbonyl molecule, formaldehyde in the case
3
0
the SAPHIR chamber; yields of 0.00 ꢀ 0.05 for OH, and
were determined from C and O
0.50 ꢀ 0.25 for HO
2
2
H
4
3
temporal profiles (with and without a radical scavenger present),
with indications that the radical yields may be dependent upon
humidity. The production of H atoms has also been reported
2
of ethene. The carbonyl oxide [CH OO]*, hereafter referred to
as a Criegee intermediate (CI), has often been referred to as
carbonyl oxide, dioxymethylene or peroxymethylene biradical,
owing to its zwitterionic/biradical electronic structure. However,
ab initio calculations have indicated that the ground state of
3
1
(in the absence of O ) with a yield of 0.076 ꢀ 0.060 (at 5 Torr),
2
which may result from the decomposition of the formyl radical
(HCO) formed via channels 2 and 3a and/or via the ‘hot’ acid
channel (3b). Currently, atmospheric mechanisms (e.g. the
Master Chemical Mechanism—MCM) employ yields which
are largely inferred through the observation of associated stable
the [CH
2
OO]* CI is not the zwitterion, and may be regarded
1
5–17
simply as a biradical,
as illustrated in Fig. 1. The carbonyl
and CI co-products produced from the exothermic decomposition
of the POZ (1b) possess a significant amount of vibrational
excitation, enabling further unimolecular reactions of the
excited CI to occur (but the energy involved is not sufficient
for the decomposition of the carbonyl compound to occur).
Potential fates of the CI include: direct rearrangement
and decomposition via a four-membered transition state
3
2,33
products using assumed mechanisms.
(MCMv3.1, http://mcm.leeds.ac.uk/MCM/) uses a value of
0.13 for both OH and HO yields, similar to the earlier OH
yield recommendation of 0.12.
The MCM
2
3
4
In this paper we present the results of a detailed study of
ethene ozonolysis, performed under tropospherically relevant
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 11002–11015 11003

View Article Online
Fig. 1 Schematic representation of the ethene ozonolysis reaction system. Adapted from ref. 4 and 7.
conditions and as a function of humidity, in a large atmospheric
simulation (smog) chamber, the EUropean PHOto-REactor,
EUPHORE. HO2 and HO2 + SRO2 were observed using
laser induced fluorescence (LIF) and peroxy radical chemical
amplification (PERCA) techniques, respectively. These observa-
tions, from experiments with/without radical scavengers
employed, together with measurements from a suite of other
instrumentation, were then used to derive yields for stable and
radical products (from the fast ozonolysis process: CI forma-
tion and decomposition) using a detailed chemical mechanism
based upon the MCM to account for the subsequent and
secondary chemical reactions. We then interpret these yield
data to derive branching ratios for the various channels in the
reaction mechanism. Finally, a discussion comparing our
results with the existing literature and associated atmospheric
implications is presented.
2. Experimental
2
.1 European photoreactor facility (EUPHORE)
and instrumentation
The ethene ozonolysis experiments were carried out in the
European Photoreactor Facility (EUPHORE) in Valencia,
1
1004 Phys. Chem. Chem. Phys., 2011, 13, 11002–11015
This journal is c the Owner Societies 2011

View Article Online
Spain as part of the TRAPOZ (Total RAdical Production
and degradation Products from alkene OZonolysis) project.
EUPHORE is a large scale atmospheric simulation chamber,
which is used extensively for studying the mechanisms of
photochemical atmospheric processes. More in depth details
were removed by cross- and self-reaction. (b) Excess CO,
2 4 3 2
C H /O /CO: designed to convert OH to HO by reaction with
2
excess CO. This allowed the total (sum of) OH and HO2
1
production to be monitored via measurement of HO , and
2
simplified the interpretation of RO data, as any OH produced
2
3
5,36
of the chamber and its instrumentation are given elsewhere.
from ethene ozonolysis was scavenged by CO, suppressing side
3
Briefly, it consists of two 197 m simulation chambers, formed
from fluorine–ethene–propene (FEP) Teflon foil (127 mm
thickness) fitted with housings that exclude ambient light.
The chamber is fitted with large horizontal and vertical fans
to ensure rapid mixing. In this study a range of analytical
instruments was used (Table 1) including a laser induced
fluorescence (LIF) system for the direct measurement
2
reactions. HO removal was therefore dominated by self-reac-
tion. Additionally, the stabilised Criegee intermediates (SCIs)
were also effectively scavenged by CO to form HCHO and
4
CO2. (c) Excess cyclohexane, C
2
2 4 3 6
H /O /c-C H12: designed to
obtain an indirect OH yield by monitoring products formed
4
3
from the OH + cyclohexane reaction, (as the steady state
[OH] from ethene ozonolysis was below the detection limit of
3
7,38
5
ꢂ3
of HOx radicals.
amplifier (PERCA) and a chemical ionisation reaction time-
In addition, a peroxy radical chemical
the LIF system (ca. 5 ꢁ 10 molecules cm ). (d) Increased
humidity: experiments (a) and (b) were performed with
graduated addition of water to the chamber, to examine the
humidity dependence of the radical yields and the effect of the
3
9
4
of-flight mass-spectrometer (CIR-TOF-MS) were deployed
0
for the measurement of HO + SRO and volatile organic/
2
2
oxygenated volatile organic compounds (VOCs/OVOCs),
2
SCI + H O reaction.
respectively.
2
.3 Instrumental accuracy
2
.2 Experimental approach
The results reported in this study were derived using a range
of measurement instrumentation (Table 1). Ozone was
monitored by both FT-IR and UV-absorption spectroscopy;
the two methods were in very good agreement. CO was
measured using a commercial IR monitor; the levels observed
were in good agreement with those calculated from the
quantities admitted to the chamber. Ethene was monitored
by FT-IR with the measurement accuracy limited by the
uncertainty of the reference cross sections available and fitting
process, conservatively estimated to be 10%. Chamber dilution
significantly affects the ethene and ozone time profiles (B60%
of the ethene loss is due to dilution), but is well constrained by
All experiments were performed with the chamber housing
ꢂ6 ꢂ1
closed (i.e. dark conditions, j(NO
2
) o 2 ꢁ 10
s ), at near
atmospheric pressure and ambient temperature, and conducted
under NO -free conditions. In the absence of NO and sunlight,
chamber wall radical production has been shown to be negligible.
x
x
41
For most experiments, humidity was low (in simulation chamber
terms: dew point ca. ꢂ45 1C/75 ppmV H O). The experimental
2
procedure, starting with a clean flushed chamber, was to add SF
(as a dilution tracer), followed by ozone (250–500 ppbV) and
6
in certain cases an OH scavenger (CO or cyclohexane, in
concentrations such that Z95% of any OH produced was
scavenged). To initiate the reaction, a known aliquot of ethene
6
the measured first order decay of SF (as monitored by the
(250–500 ppbV) was injected into the chamber. The evolution of
FT-IR). HCHO was monitored by FT-IR, CIR-TOF-MS and
a formaldehyde monitor (Aerolaser; Hantzsch reaction); the
data from all techniques were in good agreement—as the
FT-IR beampath traverses the chamber, while the monitor
and CIR-TOF-MS sampled from a single point, this served to
confirm that the reactants were well mixed (on the timescale of
the ethene–ozone decay).
reactants and products was then monitored (Table 1) over
timescales of 1–3 hours. The chamber mixing time was of the
order of 2–3 minutes.
Four types of ozonolysis experiments were carried out: (a)
simple, C H /O : designed to produce OH, HO and potentially
RO2 radicals. OH radicals were near-exclusively removed
2
4
3
2
(
chemically) by reaction with ethene, whereas HO and RO2
2
The CIR-TOF-MS was calibrated after the experiments
using a variety of methods, these included: (i) for cyclohexanone,
direct admission of a known gas standard (BOC Special
Gases, UK) to the instrument inlet line, following dynamic
dilution (calibration accuracy = 20%); (ii) sampling of
‘‘in-house’’ calibration material produced by the volumetric
dilution and liquid injection of cyclohexanone (Sigma Aldrich,
purity Z99%) into Tedlar sample bags (calibration accuracy =
26%) and (iii) for formaldehyde, production of gas standards
using certified permeation tubes (Eco-Scientific, UK), diluted,
appropriately humidified and delivered to the CIR-TOF-MS
sample inlet by a commercial calibration unit (Kintec, model:
491; calibration accuracy = 28%). Overall uncertainties
for formaldehyde and cyclohexanone measurements made
by CIR-TOF-MS were calculated to lie within the ranges
a
Table 1 Experimental techniques used during this work
Instrumentation
Target species
OH, HO
Laser Induced Fluorescence (LIF)
Peroxy Radical Chemical Amplifier (PERCA)
Fourier Transform Infra Red Spectroscopy,
Nicolet Magna 550 (FT-IR)
2
(HO
2
+ SRO
3
,
2
)
C
2
H
4
, O
HCHO, SF
VOCs/OVOCs
6
Chemical ionisation Reaction Time-of-flight
Mass Spectrometer (CIR-TOF-MS)
High Performance Liquid Chromatography
UV/FLD (HPLC)
Hydroperoxides
NO, NO , NO
Nitric Oxide(s) Analyser, CLD770
2
x
2
(chemiluminescence/photolytic NO )
Ozone Analyser ML9810 (UV absorption)
CO Analyser TE48C (Hantzsch reaction)
Dew Point Hygrometer, Walz
O
3
CO
2
H O
HCHO
HCHO Analyser AL4001
1
.8–16.5 and 0.4–0.7 ppbV, respectively. CIR-TOF-MS
a
Radical and compound specific measurement uncertainties and
instrument accuracies are discussed in detail in Section 2.3.
measurement uncertainty values comprise the combination
of both instrument precision and calibration accuracy,
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 11002–11015 11005

View Article Online
accounting for measurement reproducibility and known
systematic uncertainty. The cyclohexyl-hydroperoxide obser-
vations, performed by HPLC, had an estimated uncertainty
of 30%.
ozonide (reactions 1b–4 in Fig. 1) were assumed to proceed
near-instantaneously on the 1–2 hour timescale of the experi-
ments; that is, reactions 1b, 2, 3a–c and 4 were not assigned
individual rate constants—rather the POZ and [CH OO]*
2
The radical measurements were performed by LIF (HO2)
and PERCA (SHO + RO ). The LIF system was calibrated
before, during and after each measurement campaign using the
were assumed to decompose rapidly to form radical products
(yields optimised as described below) and stable species, or the
2
2
2
stabilised CH OO, via reaction 4. Rate constants for the
4
4
O photolysis/ozone actinometry approach; calibrations
H
2
bimolecular reactions of the SCI (reactions 4a–c) were taken
directly from the MCM. The MCMv3.1 cyclohexane photo-
oxidation mechanism was also updated and extended, as
outlined below. The chamber box model simulations were
were constant to within a few percent. The estimated uncertainty
in the LIF data from a single calibration is 27% (combined
3
7,45
systematic error and precision)
although this may be
5
1
regarded as conservative as the precision contribution from
calibration will, to a certain extent, average out across the
experiments. The instrument was calibrated as a function of
humidity—in keeping with previous results (e.g. Creasey
integrated using FACSIMILE.
The initial conditions for each of the five experiments, as
shown in Table 2, were used for the initialization of the
corresponding box models. Temperature, relative humidity
and dilution rates were averaged over the duration of each
experiment, as the variation in these parameters was minimal.
The simulations were initialized at the time point at which the
maximum measured ethene mixing ratio was observed.
Four analytical stages were performed, in each case to
determine the overall yields of specific products from the
overall fast ozonolysis reaction (i.e. the CI formation/
decomposition chemistry), which are then interpreted in terms
of branching ratios for each channel within the postulated
mechanism.
4
6
et al.) a reduction in the instrument sensitivity is observed
as humidity is initially increased from near-dry conditions, but
the response flattens out within the range of humidities
employed here.
The PERCA was calibrated using methyl iodide photolysis
followed by subsequent methyl radical reaction with oxygen to
generate methylperoxy radicals, thus determining the chain
4
7
length of the chemical amplification. An NO
device was used to determine the sensitively of the Scintrex
Luminol chemiluminescence detector) to the NO product of
2
permeation
(
2
4
8
the amplification as per Fleming et al. The accuracy of
peroxy radical measurements with the Leicester dual channel
2
.4.1 Reaction rate coefficient (k
) optimisation.
O +ethene
3
4
9
The recommended rate coefficient for ethene + ozone
1
PERCA is discussed in detail elsewhere. The overall uncertainty
of peroxy radical measurements is 38% (at 2s) from
a combination of uncertainties associated with the radical
,24
(
k
) has an estimated uncertainty of 30%,
and the
O +ethene
3
model representation of the measured ethene decay could
be improved by optimising kO3+ethene, to simulate (fit) the
observed ethene and ozone time profiles. This was achieved by
using the FT-IR observations of ethene and ozone and by
minimising the sum of squares of residuals between the
measured and modelled results within FACSIMILE. Scavenger
experiment types (b) and (c) were used to determine the rate
coefficient, to suppress the OH + ethene reaction. Under these
conditions, the loss of ethene can be considered to be due to
dilution (known) and reaction with ozone only, allowing the
reaction rate coefficient between ethene and ozone to be
determined.
calibration, NO
2
quantification and humidity correction.
2
.4 Box model and mechanism optimisation
As HO
x
radicals are highly reactive in the experimental
system, results were interpreted through detailed chemical
box modelling, drawing upon the ethene photochemical
degradation scheme in the Master Chemical Mechanism
3
2,33,50
(
MCMv3.1);
in order to account for secondary chemical
processes. The MCM is a near explicit chemical mechanism,
describing the detailed gas-phase chemical processes involved
in the atmospheric degradation of important primary emitted
VOCs. The complete gas-phase photo-oxidation mechanism
for ethene was extracted (including a suitable set of inorganic
reactions) from the MCM and incorporated into a chamber
specific box model. The MCM ethene photooxidation
mechanism was updated to include a more explicit representa-
tion of the ethene–ozone reaction mechanism (Fig. 1) and
chamber processes (dilution). Within the model, the
rapid reaction steps occurring after formation of the primary
2.4.2 SCI branching ratio calculation. In the simple
ethene + ozone system, type (a) experiment, HCHO is formed
via the decomposition of the POZ, the SCI + CO reaction 4b
and the SCI + H
also formed as a first- and multi-generation product of the
OH + ethene reaction (e.g. from HOCH CH O decomposition).
2
O reaction 4c—see Fig. 1—and is
2
2
In addition the two loss pathways of HCHO in the simple,
type (a) system were dilution and reaction with OH. In the
Table 2 Initial conditions for EUPHORE experiments
e
Expt. type Ethene/ppbV Ozone/ppbV CO/ppmV Cyclohexane/ppmV Relative humidity/% Dilution /s (10 )
ꢂ1
ꢂ5
Temp/K Duration/min
a
5
5
4
4
2
01.1
22.8
90.5
33.1
64.8
478.3
486.7
469.0
455.6
242.3
0.753
570
0.281
633
504
0
0
18.5
0
0.3
0.3
0.3
0.2
6.75
7.21
4.64
5.04
4.88
293.1
295.2
296.2
300.4
302.8
130
80
200
100
90
b
c
b
b+d
0
29.0
a
b
Simple ethene and ozone. With added CO. With added cyclohexane. With added water. Derived from decay of SF
c
d
e
6
, monitored by FT-IR.
1
1006 Phys. Chem. Chem. Phys., 2011, 13, 11002–11015
This journal is c the Owner Societies 2011

View Article Online
Fig. 2 Formaldehyde production as a function of ethene consumed
for an excess CO experiment (Table 2). Data have been corrected for
dilution, and determine an overall yield of HCHO with respect to
ethene of 1.54.
excess CO experiments, the production of HCHO arises from
the decomposition of the POZ (1b) and the SCI + CO
reaction (4b), while the loss of HCHO is due to dilution only.
It can be assumed from the postulated mechanism (Fig. 1) and
5
2,53
previous studies,
that the primary formaldehyde yield,
formed directly from the decomposition of the POZ, is unity;
correspondingly yields in excess of 1 are believed to reflect
HCHO formation via reaction 4b, allowing the SCI branching
ratio to be determined, assuming reaction with CO competes
effectively with all other possible bimolecular fates of the SCI
under the experimental conditions. The HCHO yield was
obtained from regression of the HCHO production vs. C H
2
4
consumption, following correction for dilution (Fig. 2).
2
.4.3 OH yield optimisation. The steady-state OH concen-
trations generated in the ethene–ozone system were below the
detection limit of the LIF system. OH yields were determined
indirectly employing the excess cyclohexane system. The cyclo-
hexane experiments were designed to scavenge Z 95% of any
OH produced from ethene ozonolysis, generating products
4
3,54
illustrated in Fig. 3a.
The cyclohexanone formed arises
from OH reacting with cyclohexane (including both OH
formed directly from the ozonolysis reaction, and that produced
indirectly via HO
cyclohexyl-hydroperoxide formed and the fate of the cyclo-
hexylperoxy radical, c-C , depend upon the concentra-
tions of HO and RO , and identity of the RO , present in
2 3
+ O ). The ratio of cyclohexanone/
Fig. 3 (a) Cyclohexane-OH oxidation scheme (adapted from the
MCMv3.1 and ref. 57–59). (b) Schematic representation of the major
decomposition (ring-opening) pathways of the cyclohexoxy radical
chemistry under zero NOx conditions. See Table S1 (ESIw) for a full
list of the cyclohexoxy reactions used in the simulations.
6 11 2
H O
2
2
2
the system, as these compete for reaction with c-C H O .
11
6
2
Cyclohexanone ‘‘yields’’ from OH + cyclohexane are there-
fore not necessarily expected to be consistent between chemical
systems, or indeed within the same chemical system at different
observations for cyclohexanone, cyclohexyl-hydroperoxide
and HO
2
data whilst optimising the ratio k
6 7
/k and OH yield
(
peroxy radical) concentration levels, with the key being
the relative rates of the cyclohexylperoxy + RO and
HO reactions. Using the cyclohexanone (as measured
by CIR-TOF-MS), cyclohexyl-hydroperoxide (HPLC) and
steady state HO (LIF) observations, plus the ethene and O
data, the ratio k /k (where k = k6a + k6b + k6c; see Fig. 3a)
for the fate of the peroxy radical c-C H O and the OH yield
and constraining C
2
H
4
and O to their observed levels. The
3
MCMv3.1 cyclohexane mechanism was updated to include a
more explicit representation of the OH-initiated ring opening
chemistry, accounting for the decomposition/isomerisation
reactions of the cyclohexoxy radical as shown in Fig. 3a and b.
The atmospheric chemistry of the cyclohexoxy radical
involves competition between unimolecular decomposition
2
+
2
2
3
6
7
6
6
11
2
can be obtained. This was achieved by minimising the sum of
the square of the residuals between the simulations and
via ring-opening (8b) and reaction with O (8a)—unimolecular
2
isomerisation (8c) is found to be negligible owing to ring strain
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 11002–11015 11007

View Article Online
5
5,56
ꢂ18
3
ꢂ1 ꢂ1
s at 295.2 K,
(
radical + O
Fig. 3a).
The rate coefficients for the cyclohexoxy
reaction, 8a, and the unimolecular decomposition
1.33, 1.37 and 1.55 (ꢁ10 ) cm molecules
2
296.2 K and 300.4 K respectively. This corresponds to a value
ꢂ18
3
cm molecules s at 298 K which
ꢂ1 ꢂ1
of the cyclohexoxy radical, 8b, used in this study were 2.5 ꢁ
of (1.45 ꢀ 0.25) ꢁ 10
ꢂ14
(ꢂ300/T) 57
ꢂ13
(ꢂ6026/T) 58
exp , respectively,
1
0
exp
,
and 3.8 ꢁ 10
was subsequently used for all simulations performed. The
indicated uncertainty is combined precision (1s) and systematic
uncertainty of the instrumentation, see Section 2.3. This value
is in good agreement with previous studies (see Table 3), and is
which correspond to a yield of 0.43 for cyclohexanone from
reaction 8a under atmospheric conditions. This is found to be
5
9
in reasonable agreement with Orlando et al. (2000):
5
8
24
0
.36 ꢀ 0.06, and good agreement with Welz et al. (0.43),
approximately 10% lower than the IUPAC recommendation,
4
3
Atkinson et al. (0.42 ꢀ 0.05), and with the average yield of
which has an estimated uncertainty of ꢀ30%. The rate
constant for ethene ozonolysis is relatively slow in comparison
to other small chain alkenes, such that the primary loss
of ethene in the chamber is from dilution rather than
chemical loss.
58–62
0
.41 ꢀ 0.08 derived from literature studies,
calculated using
the Atkinson (2007) recommended rate coefficient for secondary
peroxy radical reactions with O in all cases, as above.
57
2
2
.4.4 HO optimisation. The HO yield was determined by
2 2
The mean total HCHO yield (i.e. production via both
pathways 1b and 4b) calculated from the excess CO experi-
ments (Table 2) was 1.54 ꢀ 0.12, corresponding to a branching
using FT-IR measurements for ethene and ozone and LIF
observations for HO and minimising the sum of the square of
2
the residuals between the simulated and observed concentra-
tions. Fig. 1 illustrates that the formation of OH is likely to be
accompanied by a formyl radical from reactions 2 and 3a,
which under the experimental conditions would react rapidly
ratio for the SCI (Y ) of 0.54. The reported uncertainty is
SCI
the combined precision (1s) and systematic uncertainty of
the instrumentation, see Section 2.3. The HCHO yield was
obtained from regression of the HCHO concentration as a
function of the change in ethene concentration, after correc-
tion for dilution (Fig. 2). The value is at the upper end of the
previous reported literature range (Table 3). Fig. 4 shows a
with O
therefore gives the overall YOH and also accounts for a frac-
tion of the HO yield; as the overall YOH is calculated during
2 2
forming HO + CO. The sum of channels 2 and 3a
2
stage 3, channels 2 and 3a can be combined and quantified
within the model. In this representation of the reaction
mechanism, the remaining channel producing HO2 in the
ethene–ozone system to be quantified is 3b, where 2 H atoms
2 4 3
comparison between the observed and modelled C H , O and
HCHO mixing ratios after optimisation of the ethene + ozone
rate constant, and before/after optimisation of the SCI yield.
Increasing the SCI yield from 0.37 (as specified in MCMv3.1)
to 0.54 (this study) markedly improves agreement between the
HCHO observations and the simulation, to within 3%.
An overall OH yield (YOH) for ethene ozonolysis in the
presence of cyclohexane (OH scavenger experiment type (c),
see Table 2) of 0.17 ꢀ 0.09 was obtained (shown in Table 4),
attributed to the combined branching ratios (2) + (3a). The
OH yield was determined by fitting the cyclohexanone, cyclo-
react near-instantaneously with O
was optimised within the model to best reproduce the observed
HO concentrations. The sum of the combined branching
ratios 2 + 3a (determined in stage 3—OH yield optimisation)
and 3b, determine the overall yield of HO
2 2
to form 2 HO . Channel 3b
2
2
.
3
. Results
The ethene + ozone reaction rate coefficients, kO3+ethene,
2
hexyl-hydroperoxide and HO observations, as described in
Section 2.4 and shown in Fig. 5. The fit to the OH-initiated
determined from the scavenger experiments (Table 2) were
Table 3 Comparison of reaction rate coefficient (kO3+ethene) and yield of stabilised Criegee intermediate CH
studies
2
OO from this work and previous
3
Rate constant/cm molecule
ꢂ1 ꢂ1
s
ꢂ18
at 298 K (ꢁ10
)
Yield of SCI (YSCI
)
SCI scavenger type
Reference
This study
MCMv3.1 (IUPAC)
a
(
(
(
(
(
1.45 ꢀ 0.25)
1.59 ꢀ 0.30)
1.45 ꢀ 0.10)
1.37 ꢀ 0.08)
2
4
89
Bahta et al.
Treacy et al.
90
91 c
1.59 ꢀ 0.30)
Atkinson and Arey
Qi et al.
b
28
1
.44
a
0
0
0
0
0
0
0
0
0
0
0
0
.54 ꢀ 0.12
CO
This study
MCMv3.1 (IUPAC)
Atkinson and Lloyd
2
4
5
.37
.40
8
67
.35 ꢀ 0.05
.37 ꢀ 0.02
.38
HCHO
HCHO
HCHO
Niki et al.
8
Kan et al.
1
66
Su et al.
Hasson et al.
Hatakeyama et al.
92
.39 ꢀ 0.11
.39 ꢀ 0.053
.47 ꢀ 0.05
.50 ꢀ 0.04
.51
H
SO
2
O and HCOOH
65
2
8
7
Total yield of decomposition products
HCOOH
Total yield of decomposition products
Horie and Moortgat
7
0
9
Neeb et al.
Neeb et al.
Horie et al.
6
9
3
.52
CF
3
C(O)CF
3
a
The error of kO +ethene and YSCI was determined from the uncertainty of the instrumentation and the standard deviation (1s) of repeated
experiments. Error not specified. Also recommendation of ref. 1.
3
b
c
1
1008 Phys. Chem. Chem. Phys., 2011, 13, 11002–11015
This journal is c the Owner Societies 2011

View Article Online
Fig. 4 FT-IR Observed temporal profiles of C
2 4 3
H (circles), O
Fig. 5 Observed temporal profiles of cyclohexanone (c-C
squares, from CIR-TOF-MS), cyclohexyl-hydroperoxide (c-C
grey triangles, from HPLC) and steady state [HO ] (black circles, from
6
H
10O—open
(squares) and HCHO (triangles) plus model simulations (solid lines)
6
H11OOH—
for an excess CO scavenger experiment type (b) (see Table 2). Model
2
simulations for optimised kO3+ethene with SCI branching ratio of
LIF) plus optimised model simulations (lines); data determine the OH
yield.
0.54 (solid line) and 0.37 (current IUPAC recommended value)
(dashed line) – see text.
cyclohexane oxidation products (c-C H O and c-C H OOH),
11
ratios (2), (3a) and (3b) for the mechanism illustrated in Fig. 1
can be assigned. The YOH of 0.17 also corresponds to the total
yield of HCO formed alongside OH; this HCO reacts near
instantaneously with O to form HO . The remaining branching
6
10
6
HO , ethene and ozone gave a k /k ratio of 0.44, which corres-
2
6
7
ꢂ12
3
cm molecules s )
ꢂ1 ꢂ1
ꢂ1 ꢂ1 63,64
s .
ponds to the lower limit (ca. 5.0 ꢁ 10
ꢂ11
3
of literature values for k
7
of 1.70 ꢁ 10 cm molecules
2
2
The MCMv3.1 RO permutation rate coefficient (for 4C
2
3
6
ratio, channel 3b forming 2 HO , is therefore 0.05.
2
ꢂ12
3
cm molecules s was used for k
ꢂ1 ꢂ1
alkyl) of 2.5 ꢁ 10
The observed decrease in YHO2 when increasing the relative
humidity from 0.2 to 29.0% is illustrated in Fig. 6. A slightly
larger decrease was seen in the observed HO2 + SRO2
measurements made by the PERCA instrument in comparison
3
3
overall, while the branching ratios for reactions 6a, 6b and 6c
used in the simulations were 0.6, 0.2 and 0.2 respectively, in
6
good agreement with Rowley et al. who determined ratios of
1
0
.58 and 0.42 for 6a and (6b + 6c) respectively.
The HO2 yields (YHO2) obtained for the four types of
to the LIF measurements for HO , when increasing the
2
humidity. RO concentrations (which would presumably be
2
experiments are shown in Table 5. The apparent decrease in
YHO2 in the presence of excess CO (experiment type b) is
discussed in Section 4. For the direct system (C H /O ) an
HOCH CH O , formed from OH + ethene) are expected to
2
2
2
be minimal in this system, and so the discrepancies between
the two independent methods are likely owing to instrumental
uncertainties or artefact(s) (as discussed in Section 2.3). The
LIF measurements are well simulated for the first B95 minutes
of the experiment, with an overall optimised yield of 0.10 for
2
4
3
HO
2
yield of 0.30 ꢀ 0.08 was determined; where the indicated
uncertainty arises from the uncertainty in the LIF system
measurement (see Section 2.3). The value of YHO2 obtained
is in reasonable agreement with the limited previous literature
HO , but the model could not reproduce the observed decrease
2
(
Table 5). Using the mean YHO2 from both the simple and
in HO seen after increasing the humidity. The decrease in the
2
cyclohexane experiments, 0.27 ꢀ 0.07, the product branching
simulated HO observed with increasing [H O] arises from the
2
2
humidity dependence of the HO
2
2
self reaction, as no H O
dependence was applied to the HO
2
yield in the model
Table 4 Comparison of OH formation yields from this work and
previous studies
Table 5 HO yields derived vs. experimental conditions
2
Yield of OH
Method
Reference
Scavenger/method
Simple C H + O
3
Yield of HO
2
Reference
a
0
0
.17 ꢀ 0.09
This Study
IUPAC
This study
.16
http://www.iupac-kinetic.
ch.cam.ac.uk/
http://mcm.leeds.ac.uk/MCM
Paulson et al.
Rickard et al.
Mihelcic et al.
Atkinson et al.
Kroll et al.
Wegener et al.
0.30 ꢀ 0.08
This study
This study
This study
This study
This study
This study
2
4
Excess CO (expt., 1) 0.10 ꢀ 0.03
Excess CO (expt., 2) 0.10 ꢀ 0.03
0
0
0
0
0
.13
MCMv3.1
Tracer
Tracer
1
8
a
.18 ꢀ 0.06
.14 ꢀ 0.07
.20 ꢀ 0.02
.12 ꢀ 0.06
Excess cyclohexane
Excess CO + H O
Average (simple +
0.24 ꢀ 0.12
0.05 ꢀ 0.01
0.27 ꢀ 0.07
1
9
2
2
7
MI-ESR
4
3
Cyclohexane
Low pressure LIF
Stoichiometry
Calculation
CO
excess cyclohexane)
MCMv3.1
MI-ESR
PERCA
Stoichiometry (dry)
Stoichiometry (wet)
2
3
ca. 0.14
0
0.13
http://mcm.leeds.ac.uk/MCM
Mihelcic et al.
3
0
27
.00 ꢀ 0.05
0.39 ꢀ 0.03
0.38 ꢀ 0.02
0.50 ꢀ 0.25
0.40 ꢀ 0.20
2
5
28
r0.05
Gutbrod et al.2
Gutbrod et al.
Fenske et al.
Qi et al.
Wegener et al.
Wegener et al.
1
30
0
0
.08 ꢀ 0.01
7
7
30
.22 ꢀ 0.06
Tracer
a
a
The uncertainty in YOH represents combined instrumental factors
51%) and uncertainty associated with the branching ratio for
decomposition of cyclohexoxy radical.
The uncertainty in YHO2 represents combined uncertainty from the
instrumentation used and uncertainty associated with the branching
ratio for decomposition of cyclohexoxy radical.
(
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 11002–11015 11009

View Article Online
0
.18 ꢀ 0.06 and 0.14 ꢀ 0.07 respectively, both in good
2
1
agreement with this study. Gutbrod et al. use excess CO as
an OH scavenger, reporting a YOH of 0.08 ꢀ 0.01 by detecting
CO2 as a product of the CO + OH reaction. Ab initio
calculations made prior to their study, however, estimate a
2
5
43
YOH of r0.05. Atkinson and co workers measured an OH
yield of 0.12 from ethene ozonolysis using cyclohexane as
an OH scavenger, which is also found to be in reasonable
agreement with this work, but has a reported uncertainty of a
factor of 1.5 (i.e. ꢀ0.06) owing to the unknown HO
tration in the system, and therefore uncertain contribution of
the HO + O reaction to the total OH. Previous studies using
2
concen-
2
3
cyclohexane as an OH scavenger have monitored the produc-
tion of cyclohexanone + cyclohexanol (or just cyclohexanone
itself) relative to the extent of alkene ozonolysis in order to deter-
mine an OH yield, based upon a known (or measured) yield
of cyclohexanone (+cyclohexanol) from the OH-initiated
oxidation of cyclohexane. The yield of OH is then determined
from eqn (9):
Fig. 6 Temporal profile of HO
PERCA, black triangles) plus model simulations before (dashed line,
base case MCMv3.1 chemistry) and after (solid line) optimising the
HO yield to the LIF data, for an excess CO scavenger experiment
type (b), see Table 2). H O was added to increase the humidity from
.2% to 29% RH over a 26 minute period from 97 min (dotted lines).
with increased RH reflects secondary
self-reaction) only.
2 2 2
(LIF, grey circles) and HO + SRO
(
2
(
2
0
½
cyclohexanoneꢃðþ½cyclohexanolꢃÞ=D½alkeneꢃ
cyclohexanoneꢃðþ½cyclohexanolꢃÞ=D½cyclohexaneꢃ
ð9Þ
Modelled variation in HO
chemical effects (HO
2
YOH ¼
½
2
mechanism. The humidity dependence of the HO
2
+ HO
reaction is well defined with an uncertainty of approximately
0%; the reaction rate constant would have to be in error by
2
2
4
There have been numerous studies calculating the yield of
cyclohexanone + cyclohexanol from the OH-initiated oxidation
43,61,72,73
of cyclohexane, giving a range of values between 0.50–0.88.
3
a factor of 2 to account for the observed trend in isolation,
suggesting that YHO2 in the ethene ozonolysis reaction
decreases with increased humidity, under the conditions of
these experiments (i.e. in the presence of excess CO).
More recently, YOH has been determined by monitoring the
evolution of the parent alkene relative to the production of
5
4,73,74
cyclohexanone only.
from the OH-initiated oxidation of cyclohexane have also been
A range of yields of cyclohexanone
1
,43,54,61,73
reported: 0.26–0.53.
nism of the self reaction of CH
c-C , suggests that the yield of cyclohexanone from
the OH-initiated oxidation of cyclohexane should always
However, a study of the mecha-
4
. Discussion
3
OCH , analogous to
2 2
O
6 11 2
H O
The SCI yield (YSCI) obtained of 0.54 ꢀ 0.12 is slightly larger
6
5
than that reported by Hatakeyama et al., (0.39 ꢀ 0.05), who
7
5
be close to 0.5. The range in the reported yields is likely
to arise from the varying flux through the HO reaction
reaction 7 in Fig. 3a) between studies, as the yield of
measured the yield of sulphuric acid (H SO ) formed from the
4
2
ozonolysis of ethene in the presence of SO . In this study it was
2
2
(
assumed that the SCI was scavenged by reaction with SO
2
,
cyclohexyl-hydroperoxide is dependent on the HO concentra-
2
forming SO , which was further converted to H SO aerosol.
4
3
2
6
6
67
tion, in competition with reactions 6a, 6b and 6c. This, coupled
with the decomposition/isomerisation reactions of the cyclo-
hexoxy radical (reactions 8b and 8c in Fig. 3), results in a less
than unity yield of cyclohexanone and cyclohexanol. The
yields observed will depend upon the relative concentrations
The yields determined by Su et al. and Niki et al. are also
lower than that obtained in this study (Table 3). These studies
used long path FT-IR spectroscopy to measure the yield
of what was thought to be hydroxy methyl formate (HMF,
HOCH
2 2
OCHO) as a product from the reaction of CH OO +
2 2 2
of RO and HO radicals (and RO speciation) present, and
HCHO. However, this product was later identified as being
6
8
will therefore be dependent on the specific reaction system and
reagent concentrations used. Thus, this scavenger technique
may result in a large variation in YOH if applied to systems
hydroperoxy methyl formate (HPMF, HOOCH OC(O)H).
2
The YSCI value obtained here is, however, in good agreement
9,70
6
with the studies of Neeb and co-workers
HPMF yield of 0.51 in the presence of excess HCOOH
who report a
where the HO and/or RO present differ from those under
2
2
7
exploiting the reaction SCI + HCOOH), and deduce the
0
which the cyclohexanone/cyclohexanol yields from cyclo-
hexane were originally determined.
In this work, owing to the slow ethene + O rate coefficient
(
total yield of decomposition products from the excited CI,
6
respectively. The YSCI has also been found to be pressure
9
3
and the low OH yield, approximately 40% of the OH
formed in the overall ethene + ozone system comes from
reaction (10):
dependent, where it is reported that a significant fraction
(
0.20 ꢀ 0.03) of the initial [CH
2
OO]* formed is thermally
‘cold’’ at low pressure, determined from extrapolation of
7
1
‘
measurements over the range 13 mbar to 1.5 bar.
HO
2
+ O
3
- OH + 2O
2
(10)
The calculated YOH of 0.17 ꢀ 0.09 is consistent with the
1
recent literature (Table 4). Paulson et al. and Rickard et al.
8
19
If YOH were calculated using eqn (9) (i.e., from the relationship
between the cyclohexanone formed versus the amount of
employed OH radical tracer methods to obtain yields of
1010 Phys. Chem. Chem. Phys., 2011, 13, 11002–11015
1
This journal is c the Owner Societies 2011

View Article Online
ethene reacted), while adopting an average cyclohexanone
yield of 0.528 from recent cyclohexane + OH oxidation
that due to reaction with ethene) in excess CO experiments,
deriving an HO
2
yield, YHO2 of 0.50 ꢀ 0.25. The large
5
4,73
studies;
obtained. Allowing for the OH formation from HO + O , an
an overall OH yield of 0.09 ꢀ 0.02 would be
uncertainty reflects the small contribution of the HO + O
2
3
reaction to the total consumption of O , but is in agreement
3
2
3
effective YOH of 0.05 for the direct decomposition of
CH OO]* would be obtained, a factor of 3 smaller than the
(within uncertainty) with this work. The PERCA HO2 +
SRO data obtained here are consistent (within uncertainty)
[
2
2
value determined through the model optimisation approach.
This highlights the importance of considering the detailed
2
with the HO observations for the ethene/ozone/CO system
(Fig. 6), and indicate that organic peroxy radical levels in this
system are small. This is in accord with our understanding
of the mechanism (Fig. 1), and the presence of an OH
scavenger (CO). The PERCA data also (independently)
display the same trend with added water vapour as the LIF
scavenger chemistry as well as the concentration of HO
2
under
]/[ethene] conditions. Mihelcic et al., determined a
OH of 0.20 ꢀ 0.02, by measuring HOCH CH using the
MIESR technique, as this peroxy radical is a direct product of
the C + OH reaction. This slightly larger YOH may reflect
secondary OH formation from HO + O ; for the present
2
7
high [O
3
Y
2
2 2
O
H
2 4
2
HO observations, over and above that anticipated from the
humidity dependence of the HO self-reaction, as discussed
2
2
3
work, if the secondary OH formation via reaction (10) is
combined with the 0.17 OH yield determined for the direct
further below. The decrease in YHO2 seen when increasing
relative humidity from 0.2 to 29.0% is in qualitative agreement
3
0
isomerisation/decomposition of [CH OO]*, (calculated by OH
2
with Wegener et al., who report decreases, albeit with large
uncertainty, in HO2 with enhanced humidity for the
yield optimisation simulations, in which the HO
2
+ O
3
Y
reaction is accounted for), an OH formation (primary and
secondary) yield of 0.24 would be obtained.
The first direct evidence for OH formation in the ethene
ozonolysis of ethene (0.50 ꢀ 0.25 to 0.40 ꢀ 0.20), propene
(1.50 ꢀ 0.75 to 1.15 ꢀ 0.60) and isobutene (2.00 ꢀ 1.00 to
1.60 ꢀ 0.08).
2
2
27
ozonolysis system was obtained by Donahue et al. from LIF
observations at a total pressure of 5 Torr, with a reported YOH
of 0.43 ꢀ 0.20. Subsequent studies from the same laboratory
confirmed that the OH radical formation was not pressure
Mihelcic et al. measured YHO2 by MIESR in the presence
and absence of CO, and found that the overall YHO2 increased
from 0.39 to 0.66 in the presence of excess CO (arising from
the total HOx radical yield, OH + HO ). The observed
2
2
3,76
dependent,
ponding to prompt OH production over short timescales
and reported a refined YOH of 0.14, corres-
increase in YHO2 in the presence of CO is in contrast to this
work, where a decrease in YHO2 was seen (after accounting
for additional HO production via OH + CO). Mihelcic
2
3
(
B30 ms) at pressures between 10–60 Torr. This suggests
that OH is not formed by the direct dissociation of the excited
CI [CH OO]*, but rather via fragmentation of a ‘hot’ formic
2
2
7
et al., however, used large initial concentrations of ethene
(ca. 125–350 ppmV) to ensure that the ozonolysis kinetics were
2
acid intermediate, as shown in reaction 3a in Fig. 1. In
contrast, however, the pressure dependence of OH yields
3
pseudo-first-order with respect to O , and used a reaction rate
coefficient, kO3+ethene, which is 40% smaller than the currently
accepted value for their analysis, stating that the use of a larger
value (such as current evaluations recommend) for kO3+ethene
would result in the radical yields decreasing by B40%.
The apparent decrease in YHO2 observed in the presence of
excess CO is not consistent with the mechanism presented in
Fig. 1, which suggests that the YHO2 would remain the same
7
was also investigated by Fenske et al., who reported an
7
increase in OH formation with decreasing pressure from
0
.22 ꢀ 0.06 at 760 Torr to 0.61 ꢀ 0.18 at 20 Torr. The authors
proposed an alternative biradical pathway at low pressures,
but could not preclude that the ‘hot’ formic acid formed from
the dioxirane could serve as a pressure dependent source of
7
6
OH. Kroll et al. attributed this increase in OH production
at low pressures to bimolecular reactions of reactive inter-
mediates, which could be significant over longer timescales. To
date, direct measurements of OH from ethene ozonolysis at
atmospheric pressure using LIF have not been achieved; thus,
although there is strong evidence that OH is formed from
the fragmentation of ‘hot’ formic acid at lower pressures
while the observed [HO ] increases due to the formation of
2
2
HO via the reaction of OH + CO. In order to account for this
behaviour, we consider three possible explanations: Firstly,
bimolecular reactions with a partially stabilised CI. There are
three potential channels that form HO
of the CI to form OH + HCO (HCO + O
2
; direct decomposition
2
- HO + CO )
2
2
(reaction (2) in Fig. 1), and two branching ratio products from
the fragmentation of ‘hot’ formic acid forming OH + HCO
(reaction 3a) and CO + 2H (H + O + M - HO + M)
(
10–60 Torr) and short time scales (on the order of ms) the
lack of direct measurements at atmospheric pressure, both at
short and long timescales, means that OH formation from the
direct dissociation of the excited CI [CH OO]* cannot be ruled
2
2
2
(reaction 3b). It has been shown that OH formation from
alkene ozonolysis can result from both prompt formation
from a vibrationally excited CI and formation from the
decomposition of a partially ‘‘thermalised’’ CI, at longer
2
out. Fig. 1 illustrates that the production of OH may result
from both channels 2 and 3a; the data obtained here cannot
distinguish between these routes and so channels 2 and 3a are
summed, giving an overall OH yield.
7
,76
timescales.
Thus, there is the possibility for bimolecular
reactions to occur with the partially thermalised CI interrupting
the decomposition process, particularly in the presence of
excess CO, as illustrated in Fig. 7. This could also account
for the sub-unity OH yield observed for the substituted alkene
2
The HO yields (YHO2) calculated for the different types of
experiments are shown in Table 5, where the values obtained,
together with the previous literature, suggest that the yield of
2
4
HO
2
currently used in the MCMv3.1 (0.13) is an underesti-
species 2,3-dimethyl-2-butene,
where the hydroperoxide
3
mate. Wegener et al. exploited reaction 10 to evaluate HO2
0
channel should be dominant with an expected OH yield of 1,
but for which many studies measure a yield of o1, giving rise
formation from the additional ozone turnover (over and above
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 11002–11015 11011

View Article Online
2
4
to a recommended yield of 0.88. This hypothesis has not
to our knowledge) been tested for the CH OO CI and/or
HCHO or 1–10 ppmV of HCOOH were used in these studies,
8
2
(
2
in order to scavenge the SCI; Wolff et al. confirmed the
formation of HPMF in the ozonolysis of ethene in the
presence of excess HCOOH, but also observed HPMF in
the absence of HCOOH when using very large ethene concen-
trations. Under the conditions of our experiments, FAN
would not be expected to arise from the decomposition of
HPMF formed from SCI + HCOOH, as the SCI would be
expected to react with the excess CO present, rather than
HCOOH and/or HCHO.
with CO as the scavenger, but was partially tested for the
ozonolysis of 2-methyl-2-butene in the presence of the Criegee
7
scavengers H O, SO , butanone and acetic acid. The authors
8
2
2
concluded that bimolecular reactions with the studied scavengers
were too slow to inhibit OH formation. The proposed
bimolecular reaction of CO with the partially stabilised CI
would, however, only account for half of the overall observed
2
HO reduction in the CO scavenger experiments, if it is
assumed that as an upper limit 100% of the OH formation
from ethene ozonolysis at atmospheric pressure comes from
Thirdly, the similar YHO2 obtained for ethene–ozone experi-
ments and those in which cyclohexane was present may indicate
that the presence of SCI scavengers could be responsible for
the decrease in YHO2. This could imply the occurrence of an
independent HO2 channel from the SCI, which may be
2
the direct decomposition of [CH OO]* and not from the ‘hot’
formic acid. Scavenging of the partially stabilised CI would
also contribute to the observed HCHO yield, as a product of
the CO + SCI reaction, altering the overall SCI yield inferred.
A second possible explanation for the observed decrease in
inhibited in the presence of CO and/or H O. Such a radical
2
pathway has been proposed in a theoretical study investigating
the unimolecular decomposition channels of hydroxymethyl
Y
HO2 in the presence of excess CO is reaction of the dioxirane
with CO resulting in the formation of formic anhydride
7
hydroperoxide (HOCH
CH OO + H O reaction. The authors, however, also state
2
OOH or HMHP), the product of the
83
9,80
79
In a previous study, Kuhne et al. proposed a
(
FAN).
2
2
mechanism for the formation of FAN, shown in Fig. 8, and
reported an increase in its yield in the presence of excess CO.
This channel would compete with the isomerisation of the
dioxirane to the ‘hot’ formic acid, should the lifetime of the
that the water assisted decomposition of HMHP would yield
further radical products, which is in contrast to the results of
this study.
The branching ratio obtained for channel 3b of 0.05 (Fig. 1)
equates to a H atom yield of 0.10, which is in agreement
with the previously reported low pressure H atom yield of
8
1
dioxirane be sufficient. Kan et al. attributed the formation
of FAN to the decomposition of hydroxyl methyl formate
3
1
(
HMF, HOCH OCHO), a product of the reaction of
2
0.076 ꢀ 0.060. This suggests that H atoms are formed via the
hot acid channel, 3b, and not via the formyl radical, in
agreement with ab initio calculations on the mechanism of
6
6
SCI + HCHO, where Su et al. had previously reported a
FAN yield of 0.24 in the presence of excess CO, using FT-IR.
However, more recently the infra-red spectrum attributed
to HMF was reassigned to hydroperoxy methyl formate
8
4
HCO + O
channel 3b have been reported: 0.06–0.10;
are in agreement, within uncertainty, with this study. Horie
2
.
A range of branching ratio determinations for
1
,34,85–88
all of which
6
8
(
HPMF, HOOCH
2
OCHO), a product of the reaction of
8
7
SCI + HCOOH, and a combined yield of 0.18 was assigned
and Moortgat derived a branching ratio of 0.09 for channel
3b by determining the H-atom yield from the increase in CO
after addition of HCHO to the reaction system. Subsequently
6
for both HPMF and FAN. Relatively high levels of
9
2
–9 ppmV of ethene and ozone, and either 5–50 ppmV of
Fig. 7 Possible routes to OH formation from the vibrationally excited CI.
1
1012 Phys. Chem. Chem. Phys., 2011, 13, 11002–11015
This journal is c the Owner Societies 2011

View Article Online
79
Fig. 8 Proposed mechanism for the formation of formic anhydride (FAN) (after Kuhne et al.).
Table 6 Branching ratios derived for the CH OO* CI formed in the
2
ozonolysis of ethene
CO is assumed to reflect the reaction of CH
leading to CO and HCHO), determining an SCI yield of
.54 ꢀ 0.12.
An OH yield of 0.17 ꢀ 0.09 was derived from the
isomerisation/decomposition of CH OO* (combined branching
2
OO with CO
(
2
0
Reaction number
(see Fig. 1)
Branching
ratio
Reaction
2
(
(
(
(
(
2) + (3a)
CH
CH
CH
CH
CH
2
2
2
2
2
OO* - OH + HCO
OO* - CO + 2 H
OO* - CO + H
OO* - CO + H
OO* - CH OO
0.17
0.05
0.24
ratio of channels 2 + 3a, see Fig. 1), through numerical
optimisation to observed concentrations of C H OOH, cyclo-
hexanone, ethene, ozone and HO . The results highlight the
3b)
3c)
3d)
4)
2
a
2
O
6
11
2
2
2
2
0.54
importance of accounting for the HO abundance, particularly
2
a
Combined branching ratio for reactions (3c) and (3d).
under relatively high O
scavenger chemistry (to allow for RO
the O + HO contribution to the overall OH yield) when
deriving radical yields in scavenger experiments.
A mean HO
yield of 0.27 ꢀ 0.07 was determined in the
presence (0.24 ꢀ 0.12) and absence (0.30 ꢀ 0.08) of cyclo-
hexane, with an apparent decrease in HO yield observed in
3
/alkene conditions, and the detailed
2
+ HO coupling and
2
3
2
6
Neeb et al. showed that CO is also produced in the reaction
9
of SCI + HCHO, and assigned a branching ratio of 0.23 to
2
8
8
the combined channels 3b and 3d. Thomas et al. studied the
pressure dependence of the products of ethene ozonolysis,
using long path FT-IR and GC/UV, determining branching
ratios for channels 3b, 3c and 3d of 0.09, 0.31 and 0.13,
respectively. With the quantification of the branching ratios
determined during this study (0.17 for the combined branching
ratios 2 and 3a, 0.05 for 3b and 0.54 for 4) the remaining
branching ratios (3c and 3d) are combined to 0.24. The
branching ratios derived for reactions 1–4 are summarised in
Table 6.
2
the presence of both CO (to 0.10 ꢀ 0.03) and increased
humidity (to 0.05 ꢀ 0.01). Potential explanations for these
dependencies are: (i) bimolecular reaction of the partially
thermalised Criegee intermediate (CI) interrupting the
decomposition process, particularly in the presence of CO
scavenger; (ii) the reaction of CO + dioxirane, leading to
the formation of formic anhydride (FAN); (iii) an additional
HO
SCI, which competes with CO and H
consistent with competition between radical production and
bimolecular removal of the CH OO carbonyl oxide, indicating
2
production channel from bimolecular reactions of the
2
O. These results are
5
. Conclusion
2
The reaction between ethene and ozone has been investigated
in detail under ambient tropospheric boundary layer conditions
using a range of instrumentation, including FT-IR, CIR-TOF-MS,
PERCA and LIF. Ethene ozonolysis experiments in the
presence and absence of cyclohexane and CO were performed
with the results interpreted through detailed chemical box
modelling. A reaction rate coefficient for ethene + ozone
that the ‘‘real’’ HO2 yield from ethene ozonolysis under
atmospheric conditions (with higher humidity and in the
presence of other known CI co-reactants such as CO, SO2
and NO) may be lower than is observed in simple ethene + ozone
experimental systems.
From our results and previous studies, we suggest the
following rate constant/yields are incorporated into atmo-
spheric chemistry and air quality models where chemical detail
is required:
ꢂ18
3
ꢂ1 ꢂ1
(
k
O +ethene) of (1.45 ꢀ 0.25) ꢁ 10
cm molecules s at
3
2
98 K was calculated by optimising kO3+ethene to best simulate
the observed ethene and ozone decay traces. The postulated
mechanism and previous studies indicate that the primary
HCHO yield is unity; therefore, the averaged measured
HCHO yield of 1.54 ꢀ 0.12 in the presence of excess
3 2
O + ethene - 0.17 OH + 0.27 HO + 0.54 SCI + 1.00 HCHO
k
= 1.45 (ꢀ0.25) ꢁ 10ꢂ18 cm molecules
3
ꢂ1 ꢂ1
s at 298 K
O +ethene
3
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 11002–11015 11013

View Article Online
Data from alternative experimental approaches, in which
other potential product channels (for example, CO, CO ) are
19 A. R. Rickard, D. Johnson, C. D. McGill and G. Marston, J. Phys.
Chem. A, 1999, 103, 7656–7664.
2
2
2
2
2
0 A. A. Chew and R. Atkinson, J. Geophys. Res., 1996, 101,
8649–28653.
1 R. Gutbrod, S. Meyer, M. M. Rahman and R. N. Schindler, Int. J.
Chem. Kinet., 1997, 29, 717–723.
2 N. M. Donahue, J. H. Kroll, J. G. Anderson and K. L. Demerjian,
Geophys. Res. Lett., 1998, 25, 59–62.
3 J. H. Kroll, T. F. Hanisco, N. M. Donahue, K. L. Demerjian and
J. G. Anderson, Geophys. Res. Lett., 2001, 28, 3863–3866.
monitored and other parameters (e.g. pressure) are varied
would offer further mechanistic insights.
2
Acknowledgements
The authors would like to thank co-workers from CEAM and
Dr Roland Leigh, Dr Zoe Fleming and Dr Iain White from
¨
24 http://www.iupac-kinetic.ch.cam.ac.uk/, IUPAC Subcommittee on
Gas Kinetic Data Evaluation, 2006.
the University of Leicester for technical support during the
TRAPOZ experiments performed at EUPHORE and sub-
sequent data analysis, and Dr Mike Jenkin for discussion of
the cyclohexoxy radical decomposition chemistry. We are
grateful to two anonymous reviewers for their comments
which have improved the manuscript. The TRAPOZ project
was funded by the Natural Environment Research Council
2
5 R. Gutbrod, R. N. Schindler, E. Kraka and D. Cremer, Chem.
Phys. Lett., 1996, 252, 221–229.
26 J. H. Kroll, N. M. Donahue, V. J. Cee, K. L. Demerjian and
J. G. Anderson, J. Am. Chem. Soc., 2002, 124, 8518–8519.
7 D. Mihelcic, M. Heitlinger, D. Kley, P. Musgen and A. Volz-Thomas,
Chem. Phys. Lett., 1999, 301, 559–564.
8 B. Qi, K. Sato, T. Imarnura, A. Takami, S. Hatakeyama and
Y. Ma, Chem. Phys. Lett., 2006, 427, 461–465.
9 B. Qi, H. Y. Yang and Z. Q. Wang, Chin. J. Chem., 2007, 25,
2
2
2
3
(
Grant Ref NE/E016081/1).
1
082–1084.
0 R. Wegener, T. Brauers, R. Koppmann, S. R. Bares, F. Rohrer,
R. Tillmann, A. Wahner, A. Hansel and A. Wisthaler, J. Geophys.
Res., [Atmos.], 2007, 112, DOI: 10.1029/2006JD007531.
References
1
J. G. Calvert, R. Atkinson, J. A. Kerr, S. Madronich,
G. K. Moortgat, T. J. Wallington and G. Yarwood, The Mechanism
of Atmospheric Oxidation of the Alkenes, Oxford University Press,
New York, 2000.
31 J. H. Kroll, J. S. Clarke, N. M. Donahue, J. G. Anderson and
K. L. Demerjian, J. Phys. Chem. A, 2001, 105, 1554–1560.
32 S. M. Saunders, M. E. Jenkin, R. G. Derwent and M. J. Pilling,
Atmos. Chem. Phys., 2003, 3, 161–180.
2
T. B. Ryerson, M. Trainer, W. M. Angevine, C. A. Brock,
R. W. Dissly, F. C. Fehsenfeld, G. J. Frost, P. D. Goldan,
J. S. Holloway, G. Hubler, R. O. Jakoubek, W. C. Kuster,
J. A. Neuman, D. K. Nicks, D. D. Parrish, J. M. Roberts and
D. T. Sueper, J. Geophys. Res., [Atmos.], 2003, 108, 4249–4272.
S. E. Paulson and J. J. Orlando, Geophys. Res. Lett., 1996, 23,
33 M. E. Jenkin, S. M. Saunders and M. J. Pilling, Atmos. Environ.,
1997, 31, 81–104.
34 R. Atkinson, J. Phys. Chem. Ref. Data, 1997, 26, 215–290.
35 K. H. Becker, EUPHORE: Final Report to the European Commission,
Contract EV5V-CT92-0059, Bergische Universitat Wuppertal,
¨
3
4
5
Germany, 1996.
3
727–3730.
S. E. Paulson, M. Y. Chung and A. S. Hasson, J. Phys. Chem. A,
999, 103, 8125–8138.
36 K. H. Becker, In situ Euphore Radical Measurement (EUROPHAM):
final Report to the European Commission, Contract ENV4-CT95-0011,
¨
Bergische Universitat Wuppertal, Germany, 1999.
37 W. J. Bloss, J. D. Lee, C. Bloss, D. E. Heard, M. J. Pilling,
K. Wirtz, M. Martin-Reviejo and M. Siese, Atmos. Chem. Phys.,
2004, 4, 571–583.
1
M. Camredon, J. F. Hamilton, M. S. Alam, K. P. Wyche, T. Carr,
I. R. White, P. S. Monks, A. R. Rickard and W. J. Bloss, Atmos.
Chem. Phys., 2010, 10, 2893–2917.
6
M. Hallquist, J. C. Wenger, U. Baltensperger, Y. Rudich,
D. Simpson, M. Claeys, J. Dommen, N. M. Donahue,
C. George, A. H. Goldstein, J. F. Hamilton, H. Herrmann,
T. Hoffmann, Y. Iinuma, M. Jang, M. E. Jenkin, J. L. Jimenez,
A. Kiendler-Scharr, W. Maenhaut, G. McFiggans, T. F. Mentel,
A. Monod, A. S. H. Prevot, J. H. Seinfeld, J. D. Surratt,
R. Szmigielski and J. Wildt, Atmos. Chem. Phys., 2009, 9,
38 M. Siese, K. H. Becker, K. J. Brockmann, H. Geiger,
A. Hofzumahaus, F. Holland, D. Mihelcic and K. Wirtz, Environ.
Sci. Technol., 2001, 35, 4660–4667.
39 T. J. Green, C. E. Reeves, Z. L. Fleming, N. Brough,
A. R. Rickard, B. J. Bandy, P. S. Monks and S. A. Penkett,
J. Environ. Monit., 2006, 8, 530–536.
40 K. P. Wyche, R. S. Blake, A. M. Ellis, P. S. Monks, T. Brauers,
R. Koppmann and E. C. Apel, Atmos. Chem. Phys., 2007, 7, 609–620.
41 J. Zador, T. Turanyi, K. Wirtz and M. J. Pilling, J. Atmos. Chem.,
2006, 55, 147–166.
42 T. Brauers, J. Bossmeyer, H. P. Dorn, E. Schlosser, R. Tillmann,
R. Wegener and A. Wahner, Atmos. Chem. Phys., 2007, 7, 3579–3586.
43 R. Atkinson, S. M. Aschmann, J. Arey and B. Shorees, J. Geophys.
Res., [Atmos.], 1992, 97, 6065–6073.
44 U. Aschmutat, M. Hessling, F. Holland and A. Hofzumahaus, in
Physico-Chemical Behaviour of Atmospheric Pollutants, ed. G. Angeletti
and G. Reselli, European Commission, Brussels, 1994, vol. 2,
pp. 811–816.
45 W. J. Bloss, J. D. Lee, D. E. Heard, R. A. Salmon, S. J. B.
Bauguitte, H. K. Roscoe and A. E. Jones, Atmos. Chem. Phys.,
2007, 7, 4171–4185.
46 D. J. Creasey, P. A. HalfordMaw, D. E. Heard, M. J. Pilling and
B. J. Whitaker, J. Chem. Soc., Faraday Trans., 1997, 93,
2907–2913.
47 K. C. Clemitshaw, L. J. Carpenter, S. A. Penkett and M. E. Jenkin,
J. Geophys. Res., [Atmos.], 1997, 102, 25405–25416.
48 Z. L. Fleming, P. S. Monks, A. R. Rickard, D. E. Heard,
W. J. Bloss, P. W. Seakins, T. J. Still, R. Sommariva,
M. J. Pilling, R. Morgan, T. J. Green, N. Brough, G. P. Mills,
S. A. Penkett, A. C. Lewis, J. D. Lee, A. Saiz-Lopez and J. M. C.
Plane, Atmos. Chem. Phys., 2006, 6, 2193–2214.
5155–5235.
7
8
D. Johnson and G. Marston, Chem. Soc. Rev., 2008, 37, 699–716.
N. Carslaw, D. J. Creasey, D. Harrison, D. E. Heard,
M. C. Hunter, P. J. Jacobs, M. E. Jenkin, J. D. Lee,
A. C. Lewis, M. J. Pilling, S. M. Saunders and P. W. Seakins,
Atmos. Environ., 2001, 35, 4725–4737.
D. E. Heard, L. J. Carpenter, D. J. Creasey, J. R. Hopkins,
J. D. Lee, A. C. Lewis, M. J. Pilling, P. W. Seakins, N. Carslaw
and K. M. Emmerson, Geophys. Res. Lett., 2004, 31, L18112, DOI:
9
10.1029/2004GL020544.
1
0 R. Volkamer, P. Sheehy, L. T. Molina and M. J. Molina, Atmos.
Chem. Phys., 2010, 10, 6969–6991.
1
1 Y. F. Elshorbany, R. Kurtenbach, P. Wiesen, E. Lissi, M. Rubio,
G. Villena, E. Gramsch, A. R. Rickard, M. J. Pilling and
J. Kleffmann, Atmos. Chem. Phys., 2009, 9, 2257–2273.
2 V. Altuzar, S. A. Tomas, O. Zelaya-Angel, F. Sanchez-Sinencio
and J. L. Arriaga, Atmos. Environ., 2005, 39, 5219–5225.
3 M. Tornqvist, Environ. Health Perspect., 1994, 102, 157–160.
4 R. Criegee, Angew. Chem., Int. Ed. Engl., 1975, 14, 745–752.
5 D. Cremer, E. Kraka, M. L. McKee and T. P. Radhakrishnan,
Chem. Phys. Lett., 1991, 187, 491–493.
6 R. Gutbrod, E. Kraka, R. N. Schindler and D. Cremer, J. Am.
Chem. Soc., 1997, 119, 7330–7342.
7 M. Olzmann, E. Kraka, D. Cremer, R. Gutbrod and S. Andersson,
J. Phys. Chem. A, 1997, 101, 9421–9429.
1
1
1
1
1
1
1
49 D. M. Brookes, Characterisation of a PERCA instrument and
measurements of peroxy radicals during the West African Monsoon
2006, PhD thesis, University of Leicester, 2009.
8 S. E. Paulson, J. D. Fenske, A. D. Sen and T. W. Callahan,
J. Phys. Chem. A, 1999, 103, 2050–2059.
1
1014 Phys. Chem. Chem. Phys., 2011, 13, 11002–11015
This journal is c the Owner Societies 2011

View Article Online
5
0 C. Bloss, V. Wagner, M. E. Jenkin, R. Volkamer, W. J. Bloss,
J. D. Lee, D. E. Heard, K. Wirtz, M. Martin-Reviejo, G. Rea,
J. C. Wenger and M. J. Pilling, Atmos. Chem. Phys., 2005, 5, 641–664.
1 A. R. Curtis and W. P. Sweetenham, Facsimile/Checkmat User’s
Manual, Harwell Laboratory, Oxfordshire, 1987.
2 E. Grosjean, J. B. DeAndrade and D. Grosjean, Environ. Sci.
Technol., 1996, 30, 975–983.
3 E. Grosjean and D. Grosjean, Environ. Sci. Technol., 1996, 30,
73 T. Berndt, O. Boge and F. Stratmann, Atmos. Environ., 2003, 37,
3933–3945.
74 R. Winterhalter, F. Herrmann, B. Kanawati, T. L. Nguyen,
J. Peeters, L. Vereecken and G. K. Moortgat, Phys. Chem. Chem.
Phys., 2009, 11, 4152–4172.
75 M. E. Jenkin, G. D. Hayman, T. J. Wallington, M. D. Hurley,
J. C. Ball, O. J. Nielsen and T. Ellermann, J. Phys. Chem., 1993,
97, 11712–11723.
76 J. H. Kroll, S. R. Sahay, J. G. Anderson, K. L. Demerjian and
N. M. Donahue, J. Phys. Chem. A, 2001, 105, 4446–4457.
77 J. D. Fenske, A. S. Hasson, S. E. Paulson, K. T. Kuwata,
A. Ho and K. N. Houk, J. Phys. Chem. A, 2000, 104,
7821–7833.
5
5
5
5
2036–2044.
4 T. L. Malkin, A. Goddard, D. E. Heard and P. W. Seakins, Atmos.
Chem. Phys., 2010, 10, 1441–1459.
5 R. Atkinson, Int. J. Chem. Kinet., 1997, 29, 99–111.
6 J. J. Orlando, G. S. Tyndall and T. J. Wallington, Chem. Rev.,
5
5
2
003, 103, 4657–4689.
78 D. Johnson, A. G. Lewin and G. Marston, J. Phys. Chem. A, 2001,
105, 2933–2935.
79 H. Kuhne, S. Vaccani, T. K. Ha, A. Bauder and H. H. Gunthard,
Chem. Phys. Lett., 1976, 38, 449–455.
80 S. Hatakeyama and H. Akimoto, Res. Chem. Intermed., 1994, 20,
503–524.
5
5
7 R. Atkinson, Atmos. Environ., 2007, 41, 8468–8485.
8 O. Welz, F. Striebel and M. Olzmann, Phys. Chem. Chem. Phys.,
2
008, 10, 320–329.
9 J. J. Orlando, L. T. Iraci and G. S. Tyndall, J. Phys. Chem. A,
000, 104, 5072–5079.
5
6
6
6
6
6
6
6
6
6
6
7
7
7
2
0 J. Platz, J. Sehested, O. J. Nielsen and T. J. Wallington, J. Phys.
Chem. A, 1999, 103, 2688–2695.
1 D. M. Rowley, P. D. Lightfoot, R. Lesclaux and T. J. Wallington,
J. Chem. Soc., Faraday Trans., 1991, 87, 3221–3226.
2 L. Zhang, K. A. Kitney, M. A. Ferenac, W. Deng and T. S. Dibble,
J. Phys. Chem. A, 2004, 108, 447–454.
3 A. A. Boyd, P. M. Flaud, N. Daugey and R. Lesclaux, J. Phys.
Chem. A, 2003, 107, 818–821.
4 D. M. Rowley, R. Lesclaux, P. D. Lightfoot, B. Noziere, T. J.
Wallington and M. D. Hurley, J. Phys. Chem., 1992, 96, 4889–4894.
5 S. Hatakeyama, H. Kobayashi and H. Akimoto, J. Phys. Chem.,
81 C. S. Kan, F. Su, J. G. Calvert and J. H. Shaw, J. Phys. Chem.,
1981, 85, 2359.
82 S. Wolff, A. Boddenberg, J. Thamm, W. V. Turner and S. Gab,
Atmos. Environ., 1997, 31, 2965–2969.
83 R. Crehuet, J. M. Anglada and J. M. Bofill, Chem.–Eur. J., 2001, 7,
2227–2235.
84 M. Martinez-Avila, J. Peiro-Garcia, V. M. Ramirez-Ramirez and
I. Nebot-Gil, Chem. Phys. Lett., 2003, 370, 313–318.
85 R. Atkinson and A. C. Lloyd, J. Phys. Chem. Ref. Data, 1984, 13,
315–444.
86 J. T. Herron and R. E. Huie, J. Am. Chem. Soc., 1977, 99,
5430–5435.
87 O. Horie and G. K. Moortgat, Atmos. Environ., Part A, 1991, 25,
1881–1896.
88 W. Thomas, F. Zabel, K. H. Becker and E. H. Fink, Proceedings of
the 1st European Symposium on Physico-Chemical Behaviour of
Atmospheric Pollutants, Varese, Italy, 1993, 207–212.
89 A. Bahta, R. Simonaitis and J. Heicklen, Int. J. Chem. Kinet., 1984,
16, 1227–1246.
1
984, 88, 4736.
6 F. Su, J. G. Calvert and J. H. Shaw, J. Phys. Chem., 1980, 84,
39–246.
2
7 H. Niki, P. D. Maker, C. M. Savage and L. P. Breitenbach,
J. Phys. Chem., 1981, 85, 1024.
8 P. Neeb, O. Horie and G. K. Moortgat, Chem. Phys. Lett., 1995,
2
46, 150–156.
9 P. Neeb, O. Horie and G. K. Moortgat, J. Phys. Chem. A, 1998,
02, 6778–6785.
0 P. Neeb, O. Horie and G. K. Moortgat, Int. J. Chem. Kinet., 1996,
8, 721–730.
1
90 J. Treacy, M. Elhag, D. Ofarrell and H. Sidebottom, Phys. Chem.
Chem. Phys., 1992, 96, 422–427.
91 R. Atkinson and J. Arey, Atmos. Environ., 2003, 37, S197–S219.
92 A. S. Hasson, G. Orzechowska and S. E. Paulson, J. Geophys. Res.,
[Atmos.], 2001, 106, 34131–34142.
2
1 S. Hatakeyama, H. Kobayashi, Z. Y. Lin, H. Takagi and
H. Akimoto, J. Phys. Chem., 1986, 90, 4131.
2 R. Atkinson and S. M. Aschmann, Environ. Sci. Technol., 1993, 27,
93 O. Horie, C. Schafer and G. K. Moortgat, Int. J. Chem. Kinet.,
1999, 31, 261–269.
1357–1363.
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 11002–11015 11015