A kinetics and mechanistic study of the OH and NO2 initiated oxidation of cyclohexa-1,3-diene in the gas phase

Article abstract of DOI:10.1039/b417525e

The kinetics and products of the OH and NO₂-initiated oxidation of cyclohexa-1,3-diene have been investigated at 296 K and 700 Torr using long path FTIR spectroscopy. Relative rate methods were employed using the photolysis of cyclohexa-1,3-diene/CH₃ONO/NO/air mixtures to measure k(OH + cyclohexa-1,3-diene) = (1.68 ± 0.43) × 10^-10 cm ³ molecule-¹ s^-1. From the pseudo-first order decay of cyclohexa-1,3-diene in the presence of excess NO₂, a value of k(NO₂ + cyclohexa-1,3-diene) = (1.75 ± 0.15) × 10^-18 cm³ molecule^-1 s^-1 was derived. An upper limit of k ≤ 7 × 10^-21 cm³ molecule^-1 s^-1 was established for the reaction of NO with cyclohexa-1,3-diene. Benzene was observed as a product of both the OH and NO₂ initiated oxidation, providing evidence of H atom abstraction in both reactions. Assuming the reaction of cyclohexadienyl radicals (C ₆H₇) with O₂ produces benzene as the sole organic product, the results are consistent with abstraction channel branching ratios of (8.1 ± 0.2)% and (1.5 ± 0.4)%, respectively. The results also indicate that C₆H₇ reacts with NO₂, with a relative rate coefficient k(C₆H₇ + NO ₂)/k(C₆H₇ + O₂) = (1.8 ± 0.5) × 10⁵, and that this partially forms benzene, with a branching ratio of (27 ± 7)%. The stoichiometry and products of the NO₂ reaction were investigated in the absence of O₂, in the presence of O₂, and in the presence of O₂ and NO. Reaction mechanisms consistent with the observations are presented. In the presence of NO and O ₂, the NO₂-initiated chemistry leads to NO-to-NO ₂ conversion, and the formation of HO_x radicals in significant yield, (0.79 ± 0.05), such that cyclohexa-1,3-diene removal occurs by reaction with both NO₂ and OH. HCOOH was detected as a product in this system, providing evidence for significant formation of stabilised C₆ α-hydroxyperoxy radicals from the OH-initiated chemistry, and their subsequent reaction with NO. An estimate of ca. 500-1000 s^-1 is made for their decomposition rate, based on the [NO]-dependence of the HCOOH yields. The implications of the results are discussed within the context of the atmospheric chemistry of conjugated dienes. The Owner Societies 2005.

Full text of DOI:10.1039/b417525e

View Article Online / Journal Homepage / Table of Contents for this issue
R E S E A R C H P A P E R
A kinetics and mechanistic study of the OH and NO2 initiated
oxidation of cyclohexa-1,3-diene in the gas phase
M. E. Jenkin,*a M. P. Sulbaek Andersen,b M. D. Hurley,c T. J. Wallington,c F. Taketanid
and Y. Matsumid
a Department of Environmental Science and Technology, Imperial College London, Silwood
Park, Ascot, Berkshire, SL5 7PY, UK. E-mail: m.jenkin@imperial.ac.uk
b Department of Chemistry, University of Copenhagen, Universitetsparken 5,
DK-2100 Copenhagen, Denmark
c Ford Research Laboratory, Ford Motor Company, SRL-3083, PO Box 2053, Dearborn,
MI 48121-2053, USA
d Solar-Terrestrial Environment Laboratory and Graduate School of Science,
Nagoya University, Honohara 3-13, Toyokawa, Aichi 442-8507, Japan
Received 17th November 2004, Accepted 17th January 2005
First published as an Advance Article on the web 3rd February 2005
The kinetics and products of the OH and NO2-initiated oxidation of cyclohexa-1,3-diene have been investigated
at 296 K and 700 Torr using long path FTIR spectroscopy. Relative rate methods were employed using the
photolysis of cyclohexa-1,3-diene/CH3ONO/NO/air mixtures to measure k(OH þ cyclohexa-1,3-diene) ¼
(1.68 ꢀ 0.43) ꢁ 10ꢂ10 cm3 moleculeꢂ1
s
ꢂ1. From the pseudo-first order decay of cyclohexa-1,3-diene in the
presence of excess NO2, a value of k(NO2 þ cyclohexa-1,3-diene) ¼ (1.75 ꢀ 0.15) ꢁ 10ꢂ18 cm3 moleculeꢂ1 sꢂ1
was derived. An upper limit of k r 7 ꢁ 10ꢂ21 cm3 moleculeꢂ1 sꢂ1 was established for the reaction of NO with
cyclohexa-1,3-diene. Benzene was observed as a product of both the OH and NO2 initiated oxidation, providing
evidence of H atom abstraction in both reactions. Assuming the reaction of cyclohexadienyl radicals (C6H7) with
O2 produces benzene as the sole organic product, the results are consistent with abstraction channel branching
ratios of (8.1 ꢀ 0.2)% and (1.5 ꢀ 0.4)%, respectively. The results also indicate that C6H7 reacts with NO2, with a
relative rate coefficient k(C6H7 þ NO2)/k(C6H7 þ O2) ¼ (1.8 ꢀ 0.5) ꢁ 105, and that this partially forms benzene,
with a branching ratio of (27 ꢀ 7)%. The stoichiometry and products of the NO2 reaction were investigated in
the absence of O2, in the presence of O2, and in the presence of O2 and NO. Reaction mechanisms consistent
with the observations are presented. In the presence of NO and O2, the NO2-initiated chemistry leads to NO-to-
NO2 conversion, and the formation of HOx radicals in significant yield, (0.79 ꢀ 0.05), such that cyclohexa-1,3-
diene removal occurs by reaction with both NO2 and OH. HCOOH was detected as a product in this system,
providing evidence for significant formation of stabilised C6 a-hydroxyperoxy radicals from the OH-initiated
chemistry, and their subsequent reaction with NO. An estimate of ca. 500–1000 sꢂ1 is made for their
decomposition rate, based on the [NO]-dependence of the HCOOH yields. The implications of the results are
discussed within the context of the atmospheric chemistry of conjugated dienes.
product studies have been carried out on the OH-initiated
1. Introduction
oxidation of buta-1,3-diene,13–15 isoprene,8,16–20 cis-penta-1,3-
Conjugated dienes play a major role in the chemistry of the
atmosphere. Isoprene (2-methylbuta-1,3-diene) is emitted in
substantial quantities from biogenic sources, and has the high-
est global mass emissions of all non-methane hydrocarbons.1
Conjugated dienes (including cyclohexa-1,3-diene) have been
identified as components of vehicle exhaust or fuel vapour,2–4
and the detection of several conjugated dienes in the urban
environment has been reported.5 A detailed knowledge of the
atmospheric chemistry of conjugated dienes is therefore a
necessary prerequisite to understanding tropospheric chemistry
on a number of geographical scales.
It is well established that conjugated dienes react rapidly
with OH radicals, NO3 radicals and ozone.6 The reactions with
NO2 have also received attention due to their importance
under conditions typically employed in laboratory chamber
systems,7–9 in studies of vehicle exhaust, and their possible role
in thermal initiation of free radical catalysed NO-to-NO2
conversion during urban winter pollution episodes.3,10 Kinetics
data have been reported for reactions of OH and NO2 with a
number of conjugated dienes,6,11,12 and comparatively detailed
diene15 and hexa-2,4-diene.12 In addition, Ohta21 has observed
benzene as a minor product of the OH-initiated oxidation of
cyclohexa-1,3-diene.
Information on the products and mechanisms of the
NO2-initiated oxidation has been reported for the conjugated
dienes, cyclohexa-1,3-diene,3,7 1-methylcyclopenta-1,3-diene3
and hexa-2,4-diene,12 and for the fully alkyl-substituted mono-
alkene, 2,3-dimethylbut-2-ene.22 Atkinson et al.7 reported evi-
dence for the formation of a number of oxidised organic
nitrogen compounds from the reaction of NO2 with cyclo-
hexa-1,3-diene, and demonstrated that OH radicals were
formed when NO was added to the system. Subsequently,
Shi and Harrison3 reported that the NO2-initiated oxidation
of cyclohexa-1,3-diene and 1-methylcyclopenta-1,3-diene un-
der simulated urban episode conditions leads to oxidation of
NO to NO2. Jenkin et al.12 observed evidence for the formation
of oxidised nitrogen compounds from the NO2-initiated oxida-
tion of hexa-2,4-diene, and for supplementary chain removal
(attributed to reaction with OH) when NO was present in the
system. A detailed mechanism was also developed which was
1194
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5

View Article Online
fully consistent with the observations. In addition, Niki et al.22
observed the formation of a number of oxidised organic
nitrogen compounds and acetone from the NO2-initiated
oxidation of the monoalkene, 2,3-dimethylbut-2-ene.
In the present paper, we report the results of an investigation
of the kinetics and products of the reactions of OH and NO2
with cyclohexa-1,3-diene, using long-path FTIR spectroscopy.
The main objectives of this work were to gain further informa-
tion on the mechanisms of the reactions, and to confirm and
quantify the formation of HOx radicals upon reaction of
cyclohexa-1,3-diene with NO2.
2. Experimental
All experiments were performed in the Ford 140 L Pyrex
chamber, interfaced with a Mattson Sirius 100 FTIR spectro-
meter, which is described in detail elsewhere.23 The chamber is
equipped with 22 fluorescent blacklamps (GE F40BLB), emit-
ting near UV radiation in the range 300–450 nm. For the OH-
initiated studies, OH radicals were thus generated by the
photolysis of CH3ONO in the presence of O2, by the well-
established mechanism:
Fig. 1 Loss of cyclohexa-1,3-diene versus cyclohexene, during photo-
lysis in the presence of CH3ONO, in 700 Torr air. Open circles are raw
data from single illumination experiments; filled circles are corrected
data from single illumination experiments; open squares are raw data
from multiple illumination of single mixture.
CH3ONO þ hn - CH3O þ NO
CH3O þ O2 - HCHO þ HO2
HO2 þ NO - OH þ NO2
(1)
(2)
(3)
limited to o5% of the cyclohexa-1,3-diene removed. Fig. 1
shows both the corrected and uncorrected data. The combina-
tion of low NO2 concentrations and comparatively short
photolysis times (r ca. 1 min) also eliminated the significant
complications which could otherwise result from the formation
of ozone and NO3 radicals, and their reaction with cyclohexa-
1,3-diene.
Also shown in Fig. 1 are the results of an experiment with
multiple illuminations of a single mixture, which demonstrate a
progressively increasing contribution to cyclohexa-1,3-diene
removal being made by reaction (6), with probable additional
removal also resulting from the formation of ozone (and
possibly NO3 radicals) following NO2 photolysis.
The NO2-initiated studies were carried out in the dark.
Infrared spectra of reagents and products were derived from
32 co-added interferograms with a spectral resolution of 0.5
cmꢂ1, and an analytical path length of 27.1 m. All experiments
were carried out at 296(ꢀ2) K and 700 Torr total pressure.
The organic reagents, cyclohexa-1,3-diene and cyclohexene
( Z 99%) were obtained from Aldrich Chemical Company.
NO, NO2, O2, N2 and air were obtained from Michigan Airgas
at research grade purity. CH3ONO was synthesised by the
dropwise addition of 50% H2SO4 to methanol saturated with
sodium nitrite. It was dried by passing through a column of
CaCl2, purified by fractional distillation and stored in the dark
at 296 K prior to use. Reference spectra were obtained by
expanding calibrated volumes into the chamber. Unless other-
wise specified, all quoted errors are two standard deviations.
A linear regression of the corrected data in Fig. 1 passes
through the origin, within the experimental uncertainties, and
gives k4/k5 ¼ 2.48 ꢀ 0.12. Using k5 ¼ (6.77 ꢀ 1.69) ꢁ 10ꢂ11 cm3
moleculeꢂ1 sꢂ1 (as recommended by Calvert et al.6) gives k4 ¼
(1.68 ꢀ 0.43) ꢁ 10ꢂ10 cm3 moleculeꢂ1 sꢂ1. This result is in good
agreement with the previously reported measurements of k4 ¼
(1.67 ꢀ 0.33) ꢁ 10ꢂ10 cm3 moleculeꢂ1 sꢂ1 by Atkinson et al.,7,24
and k4 ¼ (1.50 ꢀ 0.36) ꢁ 10ꢂ10 cm3 moleculeꢂ1 sꢂ1 by Peeters
et al.,25 with the quoted values based upon the rate coefficient
for the reference reaction, k (OH þ isoprene) ¼ (1.02 ꢀ 0.20) ꢁ
3. Results and discussion
3.1. Kinetics of the reaction of OH with cyclohexa-1,3-diene
The kinetics of reaction (4) were measured relative to those of
reaction (5):
10ꢂ10 cm3 moleculeꢂ1 ꢂ1, as recommended by Calvert et al.6
s
OH þ cyclohexa-1,3-diene - products
OH þ cyclohexene - products
(4)
(5)
Initial reagent concentrations were 6.82–6.97 mTorr of
cyclohexa-1,3-diene, 38.6–113 mTorr of CH3ONO, and 6.67–
24.9 mTorr of cyclohexene in 700 Torr of air diluent.
The observed loss of cyclohexa-1,3-diene versus that of the
cyclohexene following UV irradiation of cyclohexa-1,3-diene/
cyclohexene/CH3ONO/air mixtures, is shown in Fig. 1. The
raw data required correction to account for the loss of cyclo-
hexa-1,3-diene via reaction with NO2 accumulating in the
system:
3.2. Products and mechanism of the reaction of OH with
cyclohexa-1,3-diene
The products generated during the photolysis of cyclohexa-1,3-
diene/CH3ONO/air mixtures were investigated in a series of
single illumination experiments. Initial reagent concentrations
were 6.82–22.6 mTorr of cyclohexa-1,3-diene and 38.6–113
mTorr of CH3ONO in 700 Torr of air diluent. The only
oxidation product which could be positively identified was
benzene. Following minor corrections for cyclohexa-1,3-diene
removal via reaction with NO2 (as described in the previous
subsection), a molar yield of (8.1 ꢀ 0.2)% was determined for
benzene (see Fig. 2), which agrees well with the value of 8.9%
reported previously by Ohta.21
NO2 þ cyclohexa-1,3-diene - products
(6)
To minimise the contact time with NO2 present, each data
point was derived from a single illumination experiment. To
obtain a value of k4, the data were corrected using a value of
k6 ¼ 1.75 ꢁ 10ꢂ18 cm3 moleculeꢂ1 sꢂ1 (as determined in the
present study: see section 3.3), the elapsed time during the
experiment, and the NO2 concentration in the chamber (mea-
sured using FTIR spectroscopy). In this way, corrections were
As described by Ohta,21 benzene is formed following allylic
H atom abstraction in cyclohexa-1,3-diene, followed by reac-
tion of the resultant cyclohexadienyl radical (C6H7) with O2
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
1195

View Article Online
of the branching ratio for H atom abstraction, k4a/k4. Addi-
tional, but conflicting, information is provided by product
studies reported by Peeters et al.25 and Vereecken and
Peeters,26 based on the measurement of H2O formed in reac-
tion (4a). Peeters et al.25 report an upper limit H2O yield of
10%, which is consistent with reaction (7) proceeding predo-
minantly as shown, and with the benzene yield providing a
good estimate of k4a/k4. However, Vereecken and Peeters26 cite
a further unpublished H2O yield of (17 ꢀ 3)% from their
laboratory, indicative of as larger value of k4a/k4 and the
existence of additional channels for reaction (7). In the absence
of other confirmatory studies, however, the mechanistic ap-
praisal in the current paper assumes that reaction of C6H7 with
O2 exclusively yields benzene, and that the benzene yield
defines the value of k4a/k4.
The reaction of OH with cyclohexa-1,3-diene is expected to
be dominated by addition of OH, and some probable features
of the subsequent oxidation mechanism in the presence of NO
are illustrated in Fig. 3. These are largely defined by analogy
with those discussed previously for hexa-2,4-diene.12 The
branching ratios displayed in the figure for the various reaction
channels have either been estimated, or are defined from
experimental observations discussed in the present paper.
The structure–reactivity relationships reported by Peeters and
coworkers25,26 for the reactions of OH with conjugated dienes
suggest that addition of OH to the terminal and central carbon
atoms of the diene system are of comparable importance. In
the case of terminal addition, the resultant radical possesses an
allyl resonance, such that there are two possible positions for
subsequent addition of O2. Given that each of the peroxy
radicals potentially formed is ‘secondary’, it seems reasonable
to assume that the probability of O2 addition is equivalent at
both positions.
Fig. 2 Formation of benzene relative to loss of cyclohexa-1,3-diene at
296 K in 700 Torr O2, during photolysis of CH3ONO/cyclohexa-1,3-
diene/O2 mixtures. Data have been corrected for loss of cyclohexa-1,3-
diene by reaction with NO2.
(see also Fig. 3):
OH þ cyclohexa-1,3-diene - C6H7 þ H2O
C6H7 þ O2 - C6H6 þ HO2
(4a)
(7)
Assuming that C6H7 reacts exclusively with O2 under the
experimental conditions, and that this proceeds only by
the channel shown, the yield of benzene provides a measure
Various peroxy radicals (RO2), generated following the
initial attack of OH are expected to react with NO via two
possible channels:
RO2 þ NO - RO þ NO2
RO2 þ NO - RONO2
(8a)
(8b)
The illustrated branching ratios, k8a/k8 ¼ 0.79, are based on
the results of the analysis presented in Section (3.4.3).
The subsequent fate of the RO radicals, formed from
channel (8a), is a major factor in determining the distribution
of oxidised products. The b-hydroxy-cyclohexenyl-oxy radi-
cals, (I) and (III), are expected to undergo rapid C–C scis-
sion,27 leading to the formation of hex-2-enedial, as shown in
Fig. 3. Key steps in this process are the reactions of O2 with the
a-hydroxy radicals (II) and (IV), which are discussed further in
Section (3.4.3). The d-hydroxy-cyclohexenyl-oxy radical (V) is
expected to react mainly with O2 to produce 4-hydroxy-cyclo-
hex-2-enone.
Although the inferred major reaction products (other than
benzene) in Fig. 3 were not positively identified, IR bands were
observed which were consistent with the formation of one or
more products containing carbonyl and nitrate functional
groups.
3.3. Kinetics of the reactions of NO2 with cyclohexa-1,3-diene
The kinetics of reactions (6) were studied by monitoring the
loss of cyclohexa-1,3-diene in the presence of excess NO2.
NO2 þ cyclohexa-1,3-diene - products
(6)
Fig. 3 Schematic representation of the mechanism initially adopted
for the reaction of OH with cyclohexa-1,3-diene in the presence of NO
and O2. Illustrated branching ratios are either estimated or defined in
the present paper (see text). Percentages are thus predicted molar yields
of boxed products relative to cyclohexa-1,3-diene lost. Where indi-
cated, modifications to the displayed mechanism are discussed further
in the text.
Initial reactant concentrations were 2.17–3.21 mTorr NO2
and 15.5–47.5 mTorr of cyclohexa-1,3-diene in 700 Torr of O2
or N2 diluent. NO2 was always in large excess and, as shown in
the insets in Fig. 4, the observed loss of cyclohexa-1,3-diene
followed pseudo first order kinetics. Fig. 4 shows the observed
pseudo first order loss rates of cyclohexa-1,3-diene as a function
1196
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5

View Article Online
in 700 Torr N2 diluent. Based on this observation, we derive an
upper limit value of k r 7 ꢁ 10ꢂ21 cm3 moleculeꢂ1 sꢂ1 for the
reaction of NO with cyclohexa-1,3-diene.
3.4. Mechanism of the reaction of NO2 with
cyclohexa-1,3-diene
3.4.1. Experiments in N2 diluent. The relative removal of
NO2 and cyclohexa-1,3-diene in 700 Torr N2 diluent was
investigated under conditions when neither reagent was present
in large excess. The results of five experiments, shown in Fig. 5,
indicate removal of (1.95 ꢀ 0.06) molecules of NO2 for every
cyclohexa-1,3-diene molecule removed, which is consistent
with a mechanism involving the sequential addition of two
NO2 molecules to cyclohexa-1,3-diene. Based on the above
discussion, we infer that this predominantly yields the 3,6-
dinitro-substituted cyclohexene, although smaller amounts of
the 3,4-dinitro isomer may also be formed. FTIR product
features were observed at 758, 1400, 1574, 1683 and 1704
cmꢂ1, consistent with a dinitrocyclohexene product. The 758
cmꢂ1 feature can be assigned to an NO2 scissors mode and the
feature at 1400 cmꢂ1 can be attributed to NO2 symmetric
stretch. The feature at 1574 cmꢂ1 is characteristic of asym-
metric NO2 stretch and, finally, the features at 1683 and 1704
cmꢂ1 can be assigned to CQC stretching modes, of which the
latter presumably is the symmetric mode.
Fig. 4 Pseudo first order loss rate of cyclohexa-1,3-diene versus NO2
concentration, observed in cyclohexa-1,3-diene /NO2 mixtures in 700
Torr of either N2 (filled symbols) or air (open symbols) diluent. The
inset shows representative decays of cyclohexa-1,3-diene in the pre-
sence of either 14.0 (circles), 28.1 (squares), or 46.5 (triangles) mTorr
of NO2.
Nitrous acid (HONO) and benzene were also detected as
trace products. This provides evidence that the reaction of NO2
with cyclohexa-1,3-diene possesses a minor H atom abstraction
channel, with benzene formed from the reaction of the cyclo-
hexadienyl radical with NO2:
of the NO2 concentration in the chamber. As seen in the figure,
there was no discernible difference between the kinetic data
obtained in either O2 or N2 diluent. Linear least squares
NO2 þ cyclohexa-1,3-diene - HONO þ C6H7
C6H7 þ NO2 - C6H6 þ HONO
(6a)
(9a)
analyses of the data give a second order rate constant k6
s
ꢂ1. Quoted errors are
¼
(1.75 ꢀ 0.15) ꢁ 10ꢂ18 cm3 moleculeꢂ1
two standard deviations from the regression analysis. The
measurement reported herein is in good agreement with the
values of k6 ¼ (1.90 ꢀ 0.38) ꢁ 10ꢂ18 and (1.78 ꢀ 0.22) ꢁ 10ꢂ18
cm3 moleculeꢂ1 sꢂ1 reported by Ohta et al.8 and Atkinson
et al.,7 respectively.
Benzene was detected with high sensitivity, and its yield was
well determined as (0.41 ꢀ 0.02)% (see Fig. 6). HONO con-
centrations were close to the detection limit and not as well-
defined, but consistent with a yield of ca. 2%. This is in
contrast to our previous study of the reaction of NO2 with
hexa-2,4-diene,12 in which HONO was not formed in detect-
able concentrations.
The reported trends in the kinetics database for reactions of
NO2 with alkenes and dienes6,11 are consistent with the reac-
tions occurring by an addition mechanism, and provide some
insights into the likely distribution of radicals initially gener-
ated from the reactions of NO2 with the conjugated dienes. As
discussed previously,12 the significantly higher reactivity of
conjugated dienes in general, compared with monoalkenes or
unconjugated dienes, suggests that addition of NO2 to terminal
carbons of a diene system is favoured, because this generates
resonance-stabilised nitroalkenyl radicals. This also allows the
unpaired electron to be located remotely (i.e. d) from the
deactivating influence of the –NO2 group. Thus, the order of
magnitude higher reactivity of hexa-2,4-diene compared with
buta-1,3-diene may reflect the fact that the d-substituted loca-
tion is secondary in the hexa-2,4-diene system, and therefore
significantly more stabilised than the corresponding primary
radical formed in the buta-1,3-diene system. The factor of five
higher reactivity of cyclohexa-1,3-diene compared with hexa-
2,4-diene, may reasonably be interpreted in terms of the release
in ring strain in the cyclic diene upon reaction with NO2 (a
similar effect has also been noted for reactions of O3 with cyclic
and acyclic conjugated dienes28). Based on the above reason-
ing, it therefore seems likely that the subsequent reactions of
the nitrohexenyl radical formed from larger conjugated dienes
proceed mainly via addition (e.g. of O2) at the position d to the
NO2 group, and this is assumed in the mechanistic appraisal
for cyclohexa-1,3-diene in the following sections.
3.4.2. Experiments in O2 diluent. The relative removal of
NO2 and cyclohexa-1,3-diene was also investigated in a series
of experiments in 700 Torr O2 diluent. The results, shown in
Fig. 5 Observed relative loss of NO2 and cyclohexa-1,3-diene in 700
Torr N2 diluent. Line is linear regression of the data: D[NO2] ¼ (1.95 ꢀ
0.06)D[cyclohexa-1,3-diene] ꢂ (0.2 ꢀ 0.4). Respective initial concentra-
tions of NO2 and cyclohexa-1,3-diene were: filled squares, 32.4 and 37.5
mTorr; open squares, 23.4 and 27.4 mTorr; triangles, 42.4 and 12.8
mTorr; filled circles, 13.8 and 14.3 mTorr; open circles, 15.5 and
2.4 mTorr.
The reactivity of cyclohexa-1,3-diene with NO was also
investigated. No discernible reaction (r2% loss of cyclo-
hexa-1,3-diene) was observed over the period of 1 h in a
mixture of 8.05 mTorr 1,3-cyclohexadiene and 27 mTorr NO
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
1197

View Article Online
The peroxynitrate is expected to be rapidly equilibrated with
the peroxy radical. At greater extents of reaction, therefore, the
slow removal of the peroxy radical by its self reaction leads to
the progressive loss of the nitrocyclohexenylperoxy nitrate, and
the generation of alternative organic products, most of which
contain only one oxidised nitrogen group:
NRO2 þ NRO2 - NRO þ NRO þ O2
NRO2 þ NRO2 - NROH þ NRꢂHO þ O2
(11a)
(11b)
Based on the products typically observed from the self
reactions of peroxy radicals (e.g., Lesclaux,29) the reaction is
expected to generate both 4-nitro-cyclohex-2-enol (NROH)
and 4-nitro-cyclohex-2-enone (NRꢂHO) via a terminating
channel (11b), and the corresponding oxy radical via a propa-
gating channel (11a). Under the experimental conditions, the
oxy radical would be expected to react mainly with O2 (provid-
ing another route to 4-nitro-cyclohex-2-enone), although the
reaction with NO2 to form a ‘nitrate’ product, 3-nitro-6-
nitrooxy-cyclohexene (NRONO2), may also contribute, parti-
cularly in the experiments with higher NO2 concentrations:
Fig. 6 Formation of benzene relative to loss of cyclohexa-1,3-diene in
NO2-initiated experiments in 700 Torr N2 diluent. Line is linear
regression of the data: [benzene] ¼ (0.0041 ꢀ 0.0002)D[cyclohexa-1,3-
diene] ꢂ (0.001 ꢀ 0.002). Initial reagent concentrations are given in the
caption to Fig. 5.
NRO þ O2 - NRꢂHO þ HO2
NRO þ NO2 - NRONO2
Reaction (12) also generates HO2 radicals, which are seques-
tered in the form of HO2NO2 under the conditions of these
experiments:
(12)
(13)
Fig. 7, indicate that in the early stages of each experiment ca. 2
molecules of NO2 were consumed for every cyclohexa-1,3-
diene molecule, while in the later stages of the experiment the
stoichiometry is significantly less than 2. The initial removal of
two NO2 molecules is interpreted in terms of the dominant
formation of 4-nitrocyclohex-2-enylperoxy nitrate (NRO2-
NO2), via the following sequence (where ‘NR’ is used as
shorthand for the 4-nitrocyclohex-2-enyl group):
HO2 þ NO2 (þM) " HO2NO2 (þM)
(14)
HO2NO2 was identified as a product (see Fig. 8), confirming
the formation of HO2 radicals from the reaction of NO2 with
cyclohexa-1,3-diene. In addition to reaction (12), the minor
abstraction channel of the initiating reaction also potentially
provides a small prompt source of HO2 in the system, in
conjunction with benzene formation:
NO2 þ cyclohexa-1,3-diene (þO2) - NRO2
NRO2 þ NO2 " NRO2NO2
(6b)
(10)
FTIR product features were initially observed at 788, 1296,
1573 and 1724 cmꢂ1, consistent with the formation of a
nitrocyclohexenylperoxy nitrate. The 788, 1296 and 1573
cmꢂ1 features can be assigned to NO2 scissors, NO2 symmetric
stretch, and NO2 asymmetric stretching modes, while the
feature at 1724 cmꢂ1 can be attributed to CQC stretching.
With the exception of the position of one band, these results
confirm the observations of Atkinson et al.,7 who report
NO2 þ cyclohexa-1,3-diene - HONO þ C6H7
C6H7 þ O2 - C6H6 þ HO2
(6a)
(7)
As shown in Fig. 9, benzene formation was observed in the
NO2/cyclohexa-1,3-diene/O2 experiments. The yield was found
to vary systematically over the approximate range 0.6–1.3%,
with lower yields observed in experiments with higher concen-
trations of NO2. This is interpreted in terms of a competition
product features at 788, 1294, 1524 and 1721 cmꢂ1
.
Fig. 7 Examples of observed relative loss of NO2 and cyclohexa-1,3-
diene in 700 Torr O2 diluent. Initial concentrations of NO2 were 6.5
mTorr (circles), 14.0 mTorr (squares) and 56.4 mTorr (triangles).
Initial concentration of ca. 15 mTorr cyclohexa-1,3-diene used in all
experiments. Curved lines are simulations using assumed mechanism
(see text).
Fig. 8 Example of time dependence of NO2, cyclohexa-1,3-diene and
HO2NO2 in experiment performed in 700 Torr O2 diluent. Lines are
simulations using mechanism and parameters discussed in the text.
1198
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5

View Article Online
substituted peroxy radicals formed in diene systems.32
Although these parameter values are entirely reasonable, they
do not uniquely describe the profiles in Figs. 7 and 8; however,
it was clear that both decomposition of the peroxynitrate and
the self reaction of the associated peroxy radical have to be
comparatively rapid to provide a good description of the data.
The observed yields of benzene, and their trend with [NO2]/
[O2], allow the following well-defined values to be determined
for a number of parameters, as illustrated by the simulations
presented in Fig. 9: k6a/(k6a þ k6b) ¼ (0.015 ꢀ 0.004); k9a/(k9a
þ
k
9b) ¼ (0.27 ꢀ 0.07); k9/k7 ¼ (1.8 ꢀ 0.5) ꢁ 105. The assigned
errors are based on the ‘goodness of fit’ of the simulations, plus
an additional 10% uncertainty. The optimised values of the
branching ratios, k6a/(k6a þ k6b) and k9a/(k9a þ k9b), are also
consistent with the ca. 0.4% yield of benzene observed in the
experiments performed in the absence of O2 (Fig. 6). There are
no previous determinations of any of these parameters with
which to compare the present results. However, the value of the
rate coefficient ratio, k9/k7, for C6H7 is of a similar magnitude
to those reported for a limited number of hydroxy-substituted
cyclohexadienyl radicals formed during the oxidation of aro-
matic hydrocarbons,33–35 which lie in the range (0.4–1.4) ꢁ 105.
The optimised branching ratios, k6a/(k6a þ k6b) and k9a/
(k9a þ k9b), constrain the prompt contribution to HO2 forma-
tion in the system. In order to simulate the observed concen-
trations of HO2NO2 (e.g. Fig. 8), it was also necessary to
include the secondary source of HO2 from reaction (12),
discussed above. However, in agreement with the results pre-
sented previously for the hexa-2,4-diene system,12 only a very
small fraction (ca. 2%) of reaction (11) needs to proceed via the
propagating channel (11a) for the observed concentrations to
be accounted for. The gradual decrease in the concentration of
HO2NO2 in the latter stages of the experiments (e.g. Fig. 8)
results from its equilibration with HO2 and NO2 in reaction
(14), and the slow removal of HO2 by its self reaction and
reaction with nitrocyclohexenyl peroxy radicals, and the steady
decay of NO2. As shown in Fig. 8, the reaction mechanism
used in these simulations provides an adequate description of
the observed time dependence of HO2NO2 throughout the
experiment.
Fig. 9 Examples of the formation of benzene relative to loss of
cyclohexa-1,3-diene in NO2-initiated experiments in 700 Torr O2
diluent. Lines are simulations using mechanism and parameters dis-
cussed in the text. Respective initial concentrations of NO2 and
cyclohexa-1,3-diene were: squares, 29.2 and 13.1 mTorr; circles, 14.0
and 14.3 mTorr; diamonds, 6.5 and 14.6 mTorr; broken line represents
the benzene yield in the absence of O2 (Fig 6).
between reactions (7) and (9) for cyclohexadienyl radicals, with
reaction (9) only partially forming benzene:
C6H7 þ NO2 - C6H6 þ HONO
(9a)
C6H7 þ NO2-other products (e.g. nitrocyclohexadiene)
(9b)
Further support for this interpretation comes from the mild
upward curvature in the yield plots which results from the
gradual depletion of NO2 during the course of each experi-
ment.
The system was simulated on the basis of the reaction
sequence discussed above using the FACSIMILE kinetics
integration package, with a number of parameters optimised
to provide a good description of the observations presented in
Figs. 7–9. The mechanism included reactions (6), (7), and (9)–
(14), and competing reactions of HO2 with itself and NRO2.
Unless otherwise stated, the rate coefficients for reactions of
NRO2 were assigned the same values as used and justified for
the analogous acyclic peroxy radical in the previous appraisal
of the reaction of NO2 with hexa-2,4-diene.12 Rate coefficients
for reaction (14) and the HO2 self reaction were based on the
recommendations of Atkinson et al.30
3.4.3. Experiments in the presence of NO and O2. The
system was further investigated in a series of experiments in
which NO was also present. The initial conditions of the ex-
periments are given in Table 1. The objective of these experi-
ments was to look for evidence of additional HOX-catalysed
removal of cyclohexa-1,3-diene when NO is present, resulting
from the formation of HO2 from the NO2 initiated chemistry,
as illustrated schematically in Fig. 10. In the presence of NO,
efficient conversion of NRO2 radical to the corresponding oxy
radical occurs via reaction (15a), with some simultaneous
radical termination occurring via reaction (15b):
The stoichiometry data in Fig. 7 and the time dependences of
the cyclohexa-1,3-diene and NO2 concentrations in Fig. 8, were
found to be reasonably well described with an equilibrium
coefficient for reaction (10) of K10 ¼ 4.2 ꢁ 10ꢂ12 cm3
moleculeꢂ1 (k10 ¼ 7.5 ꢁ 10ꢂ12 cm3 moleculeꢂ1
s
ꢂ1; k
¼
1.8 sꢂ1) and a self reaction rate coefficient, k11 ¼ 3.3 ꢁꢂ1100ꢂ12
cm3 moleculeꢂ1
s
ꢂ1. The values of k10 and k
are similar to
ꢂ10
NRO2 þ NO - NRO þ NO2
NRO2 þ NO - NRONO2
(15a)
(15b)
those reported for alkyl peroxy radicals such as C2H5O2,31 and
the value of k11 is based on those reported for hydroxy-
Table 1 Experimental conditions and results from NO2/cyclohexa-1,3-diene/NO/O2 experiments (296 K, 700 Torr)
a,b
Run [NO]0/mTorr [NO2]0/mTorr [diene]0/mTorr [O2]/Torr % Removal by OH
Benzene yield/%a HCOOH yield/%a a, b,c
A
B
C
D
E
F
a
2.5
2.7
0.09
0.09
0.11
0.22
0.62
2.19
4.0
3.9
700
700
700
700
700
72 ꢀ 1
68 ꢀ 1
72 ꢀ 2
70 ꢀ 1
68 ꢀ 1
0 ꢀ 2
6.1 ꢀ 0.4
6.4 ꢀ 0.3
5.5 ꢀ 0.6
5.0 ꢀ 0.6
5.1 ꢀ 0.5
0.45 ꢀ 0.04
2.7 ꢀ 0.3
2.5 ꢀ 0.5
3.3 ꢀ 0.6
3.5 ꢀ 0.6
4.9 ꢀ 0.7
trace
0.21
0.22
0.19
0.22
0.21
—
5.1
5.2
8.1
8.1
9.7
96.3
16.1
17.7
d
10
b
c
Determined from observations while NO present. Determined from loss rate required to account for observed diene decay. Optimised values
d
from simulation of data, assuming a ¼ b (see text). Balance N2.
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
1199

View Article Online
Fig. 12 Observed relative loss of NOx and cyclohexa-1,3-diene during
NO2-initiated oxidation, initially in the presence of NO (700 Torr O2
diluent). Data are shown for experiment A (circles) and for experiment
B (triangles). Regression analysis of ‘NO present’ data yields slope of
(0.42 ꢀ 0.04). Regression analysis of ‘NO consumed data’ yields slope
of (1.91 ꢀ 0.40).
Fig. 10 Schematic representation of the salient mechanistic features of
the NO2-initiated oxidation of cyclohexa-1,3-diene in the presence of
NO and O2. The initial sequence leads to partial formation of HOX
radicals and supplementary catalytic loss of cyclohexa-1,3-diene via
reaction with OH radicals. The mechanism also promotes NO-to-NO2
conversion and is therefore autocatalytic. ‘R’ is a collective representa-
tion of the organic groups in radicals generated from the OH-initiated
oxidation. Conversion of RO to HO2 (#) occurs by a variety of
pathways, as shown in Fig. 3 and discussed in the text. Chemistry
from the minor abstraction channels of the OH and NO2 reactions is
not shown.
with both NO2 and OH. On the basis of the measured NO2
concentrations and the value of k6 reported above in section
3.3, the observed decay rate of cyclohexa-1,3-diene during the
‘NO present’ period was consistent with between ca. 68 and
72% of its removal being due to reaction with OH in these
experiments (see Table 1). Consistent with this, benzene was
observed as a product, with a yield of (69 ꢀ 10)% of that
observed from the OH-initiated chemistry (Section 3.2), based
on the results of experiments A–E in Table 1.
Once NO is consumed, the removal rate of cyclohexa-1,3-
diene decreases, and the system reverts to that described
above in Section 3.4.2 with concentrations of both the diene
and NO2 decreasing. The contrast between the behaviour of
the system in the presence and absence of NO is further
illustrated in Fig. 12. When NO is present, the removal
stoichiometry, D[NOX]/D[cyclohexa-1,3-diene], is significantly
below unity. Once NO is consumed, a value of ca. 2 is
observed, consistent with the results presented in Fig. 7 at
low reagent conversions.
As illustrated in Fig. 10, the major termination steps when
NO is present are likely to be formation of the nitrocyclohex-
enyl nitrate (reaction (15b)), and the analogous formation of
nitrates from the reactions of the peroxy radicals generated
from the OH-initiated chemistry (reaction (8b)). Making the
assumption that these reactions alone are responsible for
radical termination in the presence of NO, and denoting the
branching ratio k15b/(k15a þ k15b) as ‘a’ and k8b/(k8a þ k8b) as
‘b’, the loss of cyclohexa-1,3-diene for every initiation event is
given by:
These reactions preclude the build up of the nitrocyclohex-
enylperoxy nitrate observed in the absence of NO, and also
provide the potential for significant generation of HOX radicals
from the subsequent reaction (12) of the oxy radical, NRO.
Because this chemistry converts NO to NO2, initial concen-
trations of NO2 were arranged to be at or close to zero.
Formation of NO2 occurs from the reaction,
NO þ NO þ O2 - NO2 þ NO2
allowing the chemistry in Fig. 10 to be initiated.
The time dependence of NO, NO2 and cyclohexa-1,3-diene
in one experiment is shown in Fig. 11, which is typical for that
observed in all experiments in 700 Torr O2 diluent. The
progressive conversion of NO to NO2 is accompanied by decay
of cyclohexa-1,3-diene, which is due to its removal by reaction
D[cyclohexa-1,3-diene] ¼ 1 þ (1 ꢂ a)/b
(i)
where the first term represent loss by reaction with NO2 and
second loss by reaction with OH. The loss of NOX for every
initiation event is given by:
D[NOX] ¼ 2a þ (1 ꢂ a) þ (1 ꢂ a) ¼ 2
(ii)
where the respective terms represent loss as nitrocyclohexenyl
nitrate, nitrocyclohexenyl ketone and the various RONO2
formed from reaction (8b). Combining eqn. (i) and (ii) there-
fore suggests that the removal stoichiometry when NO is
present is given by the expression:
Fig. 11 Time dependence of NO, NO2, cyclohexa-1,3-diene and
benzene in experiment performed in 700 Torr O2 diluent (Experiment
D, Table 1). Lines are simulations (see text). Grey line: mechanism
based on OH-initiated chemistry in Fig. 3 (with a ¼ b ¼ 0.19); black
line: mechanism with expanded a-hydroxy þ O2 chemistry given in
Fig. 14 (with a ¼ b ¼ 0.22).
D[NOX]/D[cyclohexa-1,3-diene] ¼ 2b/(1 ꢂ a þ b)
(iii)
The data presented in Fig. 12 indicate that D[NOX]/D[cyclo-
hexa-1,3-diene] ¼ (0.42 ꢀ 0.04), allowing a first indication of
1200
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5

View Article Online
the values of a and b. In the absence of quantified yields of the
nitrocyclohexenyl nitrates formed from reaction (15b), and the
nitrates formed from reaction type (8b), it is not possible to
estimate a and b independently: however, making the assump-
tion that a ¼ b in eqn. (iii), provides a value of (0.21 ꢀ 0.02) for
each of the parameters.
In conjunction with eqn. (i), any of the a, b pairs which
satisfy eqn. (iii) indicate that (0.79 ꢀ 0.08) of cyclohexa-1,3-
diene removal was due to reaction with OH when NO was
present in the system. This is slightly greater than that deter-
mined from the observed removal kinetics of cyclohexa-1,3-
diene (see Table 1). The fact that the two estimates are
not in better agreement gives an indication that the idealised
mechanism presented in Fig. 10 does not fully describe the
chemistry.
The experiments were therefore simulated using a chemical
mechanism containing a detailed representation of the OH and
NO2 initiated chemistry, as described above and in the pre-
vious subsections. This allowed a full description of the
chemistry resulting from both the addition and abstraction
channels to be included, potential competing reactions for the
intermediate peroxy and oxy radicals to be considered, and
inclusion of other relevant inorganic chemical processes (e.g.,
the reaction of OH with NO2) in the analysis. The mechanism
was initially based on that presented previously for hexa-2,4-
diene (see Table 1 of ref. 12), and rate coefficients and branch-
ing ratios not explicitly identified in the present paper (e.g. for
reactions of the peroxy radicals) were assigned the same values
as those used for the analogous acyclic species in the previous
appraisal.
Fig. 13 Formation of HCOOH relative to loss of cyclohexa-1,3-diene
in NO2-initiated experiments in 700 Torr O2 diluent with NO present.
Lines are simulations using mechanism and optimised parameters
discussed in the text. Run A – filled circles (heavy line); Run B – open
circles (light line); Run C – open squares (light line); Run D – filled
squares (heavy line); Run E – triangles (heavy line).
competitive reaction with NO potentially occurs under the
conditions of the present experiments,
RCH(OH)O2 - RCHO þ HO2
RCH(OH)O2 þ NO - RCH(OH)O þ NO2
(17)
(18)
leading to NO-to-NO2 conversion. The corresponding chem-
istry assigned to the stabilised a-hydroxyperoxy radicals, (VI)
and (VII), formed in the present system, is illustrated in Fig. 14.
In the case of (VII), the oxy radical formed from reaction with
NO decomposes to yield HCOOH and a C5 organic radical, the
subsequent chemistry of which results in additional NO-to-
NO2 conversions. In the present appraisal, the oxy radical
formed from the reaction of (VI) with NO is assumed to react
with O2, because the alternative decomposition would yield a
vinyl-type radical.
The formation of HCOOH by the above mechanism is
analogous to that proposed by Orlando et al.39 to explain its
formation from the OH-initiated oxidation of a-pinene in the
presence of high [NO]. The trends in the HCOOH yields
observed in the present study (Table 1 and Fig. 13) provide
strong support for this mechanism. The yield of HCOOH
shows a clear increase with initial [NO], consistent with the
competition between reactions (17) and (18) for a-hydroxyper-
oxy radical (VII), with the curvature in the yield plots (Fig. 13)
resulting from the depletion of NO during the course of each
experiment. Fig. 13 also shows the results of simulations of
experiments A–E (Table 1), with the chemistry illustrated in
Fig. 14 included. Based on simulation of the entire dataset (as
shown in Fig. 13), an optimised value of k17 ¼ 680 sꢂ1 was
obtained, with a typical value of k18 ¼ 8.6 ꢁ 10ꢂ12 cm3
moleculeꢂ1 sꢂ1 being assigned to the competing reaction.
Optimisation of each experiment individually provided values
of k17 lying in the approximate range 500–1000 sꢂ1, which is
regarded as indicative of the level of uncertainty in the deter-
mination. This value may be compared with the value of 1250
sꢂ1 used recently by Capouet et al.41 for decomposition of the
analogous C10 radical formed from the OH-initiated oxidation
of a-pinene, based partially on theoretical estimates.
The results of a simulation with the base mechanism are
shown in Fig. 11 (grey lines). The simulated time profiles for
NO, NO2 and cyclohexa-1,3-diene were optimised to the
experimental data by varying the values of a and b, using
non-linear least squares methodology, with the reasonable
assumption that a ¼ b (as discussed in detail and justified
previously for the hexa-2,4-diene system12). Clearly, however,
the simulated time profiles do not fully describe the system: the
simulated NO-to-NO2 conversion is insufficient, even though
the simulated removal of the diene is greater than that ob-
served. It was not possible to resolve this discrepancy by
varying parameter values, indicating that a fundamental
change to the mechanism was required.
Because cyclohexa-1,3-diene is mainly removed by reaction
with OH when NO is present, the performance of the mechan-
ism could be improved significantly by including additional
NO-to-NO2 conversions per cyclohexa-1,3-diene removed by
the OH-initiated oxidation cycle (Fig. 10), which could occur if
RO is converted to HO2 by a mechanism involving additional
peroxy radicals. Support for this was provided by the observa-
tion of HCOOH formation during the ‘NO present’ period of
experiments A–E (Table 1 and Fig. 13), indicating that the OH-
initiated chemistry must lead to chemical processes other than
those illustrated in Fig. 3. To account for HCOOH formation,
in conjunction with additional NO-to-NO2 conversion, the
mechanism shown in Fig. 14 was adopted for the reactions
of the a-hydroxy organic radicals (II) and (IV) with O2. The
mechanism of this class of reaction is generally accepted to
involve the isomerisation and decomposition of an energy rich
36–40
a-hydroxyperoxy radical intermediate,
RCHOH þ O2 - [RCH(OH)O2]z - RCHO þ HO2 (16)
The above conclusion is determined by the relative variation
of the HCOOH yield, both between and within the experi-
ments. The magnitude of the HCOOH yields required ca. 50%
of the initially-formed activated a-hydroxyperoxy radicals to
be stabilised to form RCH(OH)O2 in the optimised mechan-
ism. It is recognised that this value is also influenced by a
number of other assumptions in the mechanism, such as the
For small radicals, such as [HOCH2O2]z, rapid decomposi-
tion to HCHO and HO2 occurs exclusively.40 It is becoming
clear, however, that the possibility of formation of a stabilised
a-hydroxyperoxy radical, RCH(OH)O2, increases with the size
of the organic group.39,41 Although the stabilised RCH(OH)O2
also decomposes (more slowly) to form RCHO and HO2,
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
1201

View Article Online
Fig. 14 Mechanism assigned to the reactions of the a-hydroxy organic radicals, (II) and (IV), with O2, as part of the OH-initiated oxidation of
cyclohexa-1,3-diene in the presence of NO. The reaction sequence leads to the generation of HCOOH, and additional NO-to-NO2 conversion.
Nitrate-forming channels of the peroxy radical þ NO reactions are not shown, but were included in the mechanism where appropriate, with
branching ratios based on the Master chemical mechanism (MCM) protocol.42
branching ratios assigned to the relevant oxidation routes in Fig.
3, and the details of the mechanisms following the reactions of
the stabilised a-hydroxyperoxy radicals, (VI) and (VII), with
NO (Fig. 14). Nevertheless, the estimate of 90% stabilisation
applied recently to the activated C10 a-hydroxyperoxy radical by
Capouet et al.,41 based on a detailed RRKM analysis, suggests
that the current value for an analogous C6 species is not
unreasonable. It should also be noted that the failure to detect
HCOOH in our previous study of hexa-2,4-diene12 is consistent
with insignificant stabilisation of the C2 and C4 activated a-
hydroxyperoxy radicals, [CH3CH(O2)OH]z and [CH3CHQ
CHCH(OH)O2]z, likely to be formed in that system.
basis of k(OH þ cyclohexa-1,3-diene) ¼ (1.68 ꢀ 0.42) ꢁ 10ꢂ10
cm3 moleculeꢂ1 sꢂ1 reported here, and a typical ambient OH
radical concentration of 106 molecule cmꢂ3, a lifetime of
cyclohexa-1,3-diene with respect to reaction with OH radicals
of ca. 1.7 h can be estimated. Reaction with ozone also occurs
very rapidly. Using a rate constant of 1.2 ꢁ 10ꢂ15 cm3
moleculeꢂ1 sꢂ1 (Calvert et al.6) and a typical background
ozone level of 35 ppb, leads to a lifetime of cyclohexa-1,3-
diene with respect to reaction with ozone of ca. 15 min.
Polluted urban air typically contains NO2 concentrations of
ca. 50–100 ppb, although concentrations of several hundred
ppb have been reported during wintertime pollution
episodes.43 Using k(NO2 þ cyclohexa-1,3-diene) ¼ (1.75 ꢀ
Inclusion of the chemistry illustrated in Fig. 14 was also
found to lead to a much improved description of the time
dependence of NO, NO2 and cyclohexa-1,3-diene (e.g., see Fig.
11). With the assumption that a ¼ b, simultaneous simulation
of experiments A–E, using least squares methodology, resulted
in an optimised value of 0.21 for a and b. Optimisation of each
experiment individually provided consistent values of a and b
lying in the range 0.19–0.22 (Table 1). We therefore assign
recommended values of (0.21 ꢀ 0.05) to both a and b.
0.15) ꢁ 10ꢂ18 cm3 moleculeꢂ1
s
ꢂ1, reported here, a lifetime of
cyclohexa-1,3-diene with respect to reaction with NO2 of ca.
2.5–5 d can be estimated for the typical polluted conditions.
Reaction with NO2 is therefore a negligible sink for cyclohexa-
1,3-diene under the vast majority of ambient conditions, and
only potentially makes a minor contribution under extreme
wintertime episode conditions when a shallow inversion layer
can lead to a combination of high NOX levels and stagnant air
for periods of a day or more.43,44 Under such conditions, ozone
and OH are suppressed to very low levels. Under more typical
conditions, however, the atmospheric lifetime of cyclohexa-1,3-
diene is determined by the reactions with OH radicals and
ozone, and is likely to be less than 1 h in most environments.
In studies of the chemical composition of vehicle exhaust or
polluted urban air, it is often convenient to collect ‘bag’ or
‘bottle’ samples, which are subsequently analysed using GC
techniques. The initial concentrations of NO2 in diluted ex-
haust samples are typically in the range 1–10 ppm. Under such
conditions cyclohexa-1,3-diene has a lifetime with respect to
reaction with NO2 of 40–400 min. As illustrated in Fig. 10, the
presence of NO leads to formation of OH radicals and addi-
tional loss of cyclohexa-1,3-diene, such that this range should
be considered as upper limit lifetimes. Accordingly, a number
of studies have reported that conjugated dienes are unstable in
exhaust samples.45–48 For example, Hoekman48 reported that
buta-1,3-diene was moderately unstable (half life E 4 h) in
diluted petrol engine exhaust, and that larger conjugated dienes
( Z C5) were highly unstable (e.g., the half life of cyclopenta-
1,3-diene was {1 h), consistent with removal initiated by
reaction with NO2. Given that cyclohexa-1,3-diene is three
orders of magnitude more reactive than buta-1,3-diene and a
factor of five more reactive than cyclopenta-1,3-diene, it is clear
that its contribution (and those of other reactive dienes) to
Experiments A–E were carried out with [O2]/[NOX] Z 7 ꢁ
104. To provide further confirmation of the generation of OH
in the NO2/NO/cyclohexa-1,3-diene/O2 system under those
conditions, and the associated observations described above,
an additional experiment was carried out at reduced [O2] and
elevated [NOX], such that [O2]/[NOX] E 100 (experiment F,
Table 1). Under these conditions, the NO2-initiated chemistry
is expected to be fully terminated by reactions (13), (15b) and
(19), such that generation of HOX radicals does not occur:
NRO þ NO - products
(19)
As indicated in Table 1, the decay of cyclohexa-1,3-diene
was fully accounted for by reaction with NO2 at its observed
concentrations, confirming that insignificant formation of OH
occurred. In addition to this, the yield of benzene was compar-
able to that observed in the NO2/cyclohexa-1,3-diene/N2 ex-
periments (Section 3.4.2), and only trace concentrations of
HCOOH were observed, indicating that their formation in
experiments A–E was either predominantly or exclusively due
to the occurrence of OH-initiated chemistry.
4. Conclusions and atmospheric implications
A substantial body of experimental data is presented concern-
ing the atmospheric chemistry of cyclohexa-1,3-diene. On the
1202
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5

View Article Online
exhaust emissions may be underestimated. Similarly, the upper
limit lifetime of cyclohexa-1,3-diene in bottled samples of
polluted urban air initially containing 100 ppb NO2 is ca. 2.5
d. Consequently, analysis of such samples following storage
will underestimate its ambient concentration.
References
1
A. Guenther, C. N. Hewitt, D. Erickson, R. Fall, C. Geron, T.
Graedel, P. Harley, L. Klinger, M. Lerdau, W. A. McKay, T.
Pierce, B. Scholes, R. Steinbrecher, R. Tallamraju, J. Taylor and
P. Zimmerman, J. Geophys. Res., 1995, 100, 8873.
C. A. Jemma, P. A. Shore and K. A. Widdicombe, J. Chromatogr.
Sci., 1995, 33, 34.
The observation of benzene as a product of the photolysis of
cyclohexa-1,3-diene/CH3ONO/NO/O2 mixtures confirms that
the OH-initiated oxidation of cyclohexa-1,3-diene possesses a
minor, but significant H atom abstraction channel forming
cyclohexadienyl radicals. The observed formation of benzene
in the NO2-initiated studies indicates that the reaction of NO2
with cyclohexa-1,3-diene also possesses a minor H atom ab-
straction channel. Assuming the reaction of cyclohexadienyl
radicals with O2 produces benzene as the sole organic product,
the results are consistent with abstraction channel branching
ratios of (8.1 ꢀ 0.2)% and (1.5 ꢀ 0.4)% for the OH and NO2
reactions, respectively. The dependence of the benzene yield on
[NO2]/[O2] in the NO2-initiated experiments could be inter-
preted in terms of a competitive reaction of cyclohexadienyl
radicals with NO2, which partially forms benzene (27 ꢀ 7)%
and possesses a rate coefficient greater than the O2 reaction by
a factor of (1.8 ꢀ 0.5) ꢁ 105.
2
3
4
J. P. Shi and R. M. Harrison, Atmos. Environ., 1997, 31, 3857.
D. Smith, P. Cheng and P. Spanel, Rapid Commun., Mass
Spectrom., 2002, 16, 1124.
5
P. Ciccioli, A. Cecinato, E. Brancaleoni, A. Brachetti, M. Frattoni
and R. Sparapani, Composition and distribution of polar and non-
polar VOCs in urban, rural, forest and remote areas. Proceedings
of the 6th European Symposium on the Physico-chemical Behaviour
of Atmospheric Pollutants, Varese, 18–22 October 1993, pp.
549–568, ed. G. Angeletti and G. Restelli, European Commission,
Report EUR 15609/1 EN, ISBN 92-826-7922-5, 1994.
J. G. Calvert, R. Atkinson, J. A. Kerr, S. Madronich, G. K.
Moortgat, T. J. Wallington and G. Yarwood, The mechanisms of
atmospheric oxidation of alkenes, Oxford University Press, New
York, ISBN 0-19-513177-0, 2000.
6
7
8
9
R. Atkinson, S. M. Aschmann, A. M. Winer and J. N. Pitts, Jr.,
Int. J. Chem. Kinet., 1984, 16, 697.
T. Ohta, H. Nagura and S. Suzuki, Int. J. Chem. Kinet., 1986, 18,
1.
E. C. Tuazon and R. Atkinson, Int. J. Chem. Kinet., 1990, 22,
1221.
The results of the NO2-initiated study in the absence of O2
are consistent with dominant sequential addition of two NO2
molecules to form one or more dinitrocyclohexene isomers. In
the presence of O2, the results suggest that formation of
nitrocyclohexenyl peroxy radicals occurs. On the basis of
kinetics considerations, and the observed high yield of HOX
radicals in the presence of NO, it is postulated that the
chemistry proceeds predominantly via formation of the 4-
nitrocyclohex-2-enylperoxy isomer, which undergoes conven-
tional peroxy radical reactions (i.e. with NO, NO2 and peroxy
radicals).
When NO is present, the NO2-initiated oxidation leads to
NO-to-NO2 conversion, and the formation of HOX radicals in
significant yield. The present results are consistent with an
HOX yield of (0.79 ꢀ 0.05) at typical urban levels of NOX, and
therefore confirm the inferred generation of OH, as reported by
Atkinson et al.7 The present work thus also supports the
general features of the mechanism postulated by Shi and
Harrison3 for cyclohexa-1,3-diene and 1-methylcyclopenta-
1,3-diene, and the potential of conjugated diene/NO2 reactions
to promote NO to NO2 oxidation in samples of diluted vehicle
exhaust and fuel vapour. However, it is not possible to assess
the relative importance of these reactions for the generation of
HOX radicals in the polluted urban environment without
further quantitative measurements of the ambient concentra-
tions of conjugated dienes.
The observation of HCOOH as a product in the NO2/
cyclohexa-1,3-diene/NO/NO2 experiments provides evidence
for significant formation of stabilised C6 a-hydroxyperoxy
radicals from the OH-initiated chemistry, and their subsequent
reaction with NO. On the basis of the NO concentration
dependence of the HCOOH yield, an estimate of ca. 500–
1000 sꢂ1 is made for the decomposition rate of the stabilised
a-hydroxyperoxy radical, indicating that decomposition and
reaction with NO occur equally at an NO concentration of ca.
2.5–5 ppm. Consequently, formation of HCOOH (and possibly
other acids) from the oxidation of cyclohexa-1,3-diene by this
mechanism can only contribute under highly polluted atmo-
spheric conditions (i.e. [NO] Z 1 ppm), but could be a
significant source of carboxylic acids in exhaust samples.
10 R. M. Harrison, J. P. Shi and J. L. Grenfell, Atmos. Environ.,
1998, 32, 2769.
11 M. D. King, C. E. Canosa-Mas and R. P. Wayne, Phys. Chem.
Chem. Phys., 2000, 4, 295.
12 M. E. Jenkin, M. Sørensen, M. D. Hurley and T. J. Wallington, J.
Phys. Chem. A, 2003, 107, 5743.
13 E. C. Tuazon, A. Alvarado, S. M. Aschmann, R. Atkinson and J.
Arey, Environ. Sci. Technol., 1999, 33, 3586.
14 X. Liu, H. E. Jeffries and K. G. Sexton, Atmos. Environ., 1999, 33,
3005.
15 T. Ohta, Bull. Chem. Soc. Jpn., 1984, 57, 960.
16 R. Atkinson, S. M. Aschmann, E. C. Tuazon, J. Arey and B.
Zelinska, Int. J. Chem. Kinet., 1989, 21, 593.
17 S. E. Paulson, R. C. Flagan and J. H. Seinfeld, Int. J. Chem.
Kinet., 1992, 24, 79.
18 A. Miyoshi, S. Hatakeyama and N. Washida, J. Geophys. Res.,
1994, 99, 18779.
19 E. S. C. Kwok, R. Atkinson and J. Arey, Environ. Sci. Technol.,
1995, 29, 2467.
20 L. Ruppert and K.-H. Becker, Atmos. Environ., 2000, 34, 1529.
21 T. Ohta, Int. J. Chem. Kinet., 1984, 16, 1945.
22 H. Niki, P. D. Maker, C. M. Savage, L. P. Breitenbach and M. D.
Hurley, Int. J. Chem. Kinet., 1986, 18, 1235.
23 T. J. Wallington and S. M. Japar, J. Atmos. Chem., 1989, 9, 399.
24 R. Atkinson, J. Phys. Chem. Ref. Data, 1997, 26(2), 215.
25 J. Peeters, S. Vandenberk, E. Piessens and V. Pultau, Chemo-
sphere, 1999, 38(6), 1189.
26 L. Vereecken and J. Peeters, Chem. Phys. Lett., 2001, 333, 162.
27 R. Atkinson, Int. J. Chem. Kinet., 1997, 29, 99.
28 A. G. Lewin, D. Johnson, A. W. Price and G. Marston, Phys.
Chem. Chem. Phys., 2001, 3, 1253.
29 R. Lesclaux, Combination of peroxyl radicals in the gas phase, in
‘Peroxyl Radicals’, ed. Z. B. Alfassi, John Wiley and Sons, 1997.
30 R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F.
Hampson, R. G. Hynes, M. E. Jenkin, M. J. Rossi and J. Troe,
Atmos. Chem. Phys., 2004, 4, 1461.
31 P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau, G. D.
Hayman, M. E. Jenkin, G. K. Moortgat and F. Zabel, Atmos.
Environ., 1992, 26A, 1805.
32 M. E. Jenkin, A. A. Boyd and R. Lesclaux, J. Atmos. Chem., 1998,
29, 267.
33 R. Knispel, R. Koch, M. Siese and C. Zetzsch, Ber. Bunsen-Ges.
Phys. Chem., 1990, 94, 1375.
34 C. Zetzsch, R. Koch, B. Bohn, R. Knispel, M. Siese and F. Witte,
Transp. Chem. Transform. Pollut. Troposphere, 1997, 3, 247.
35 J. G. Calvert, R. Atkinson, K.-H. Becker, R. M. Kamens, J. H.
Seinfeld, T. J. Wallington and G. Yarwood, The mechanisms of
atmospheric oxidation of aromatic hydrocarbons, Oxford Univer-
sity Press, New York, ISBN 0-19-514628-X, 2002.
36 H. H. Grotheer, G. Riekert, D. Walter and T. Just, J. Phys.
Chem., 1988, 92, 4028.
Acknowledgements
MEJ acknowledges the UK Natural Environment Research
Council, NERC, for support via a Senior Research Fellowship
(NER/K/S/2000/00870). MPSA thanks the Danish Research
Agency for funding. We thank Dick Chase (Ford) for helpful
discussions.
37 F. L. Nesbitt, W. A. Payne and L. J. Stief, J. Phys. Chem., 1988,
92, 4030.
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
1203

View Article Online
38 P. Pagsberg, J. Munk, C. Anastasi and V. J. Simpson, J. Phys.
Chem., 1989, 93, 5162.
44 QUARG. Urban air quality in the United Kingdom. First
report of the Quality of Urban Air Review Group. Prepared at
the request of the Department of the Environment, London,
1993.
45 F. Lipari, J. Chromatogr., 1990, 503, 51.
46 N. Pelz, N. M. Dempster and P. R. Shore, J. Chromatogr. Sci.,
1990, 28, 230.
39 J. J. Orlando, B. Noziere, G. S. Tyndall, G. E. Orzechowska, S. E.
Paulson and Y. Rudich, J. Geophys. Res., 2000, 105, 11561.
40 T. S. Dibble, Chem. Phys. Lett., 2002, 335, 193.
41 M. Capouet, J. Peeters, B. Noziere and J.-F. Muller, Atmos.
Chem. Phys., 2004, 4, 2285.
42 M. E. Jenkin, S. M. Saunders and M. J. Pilling, Atmos. Environ.,
1997, 31, 81.
43 J. S. Bower, G. F. J. Broughton, J. R. Stedman and M. L.
Williams, Atmos. Environ., 1994, 28(3), 461.
47 E. W. Kaiser, W. O. Siegl, D. F. Cotton and R. W. Anderson,
Environ. Sci. Technol., 1992, 26, 1581.
48 S. K. Hoekman, J. Chromatogr., 1993, 639, 239.
1204
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 1 1 9 4 – 1 2 0 4
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5