Formation of phenol and carbonyls from the atmospheric reaction of OH radicals with benzene

Article abstract of DOI:10.1039/b514148f

The gas-phase reaction of OH radicals with benzene has been studied in a flow tube operated at 295 ± 2 K and 950 mbar of synthetic air or O ₂. Ozonolysis of tetramethylethylene (dark reaction) with a measured OH radical yield of 0.92 ± 0.08 or photolysis of methyl nitrite in the presence of NO served as the OH sources. For investigations in the presence of NO_x, the conditions were chosen so that more than 95% of the OH/benzene adduct reacted with O₂ even for the highest NO₂ concentration occurring in the experiment. In the absence of NO_x, a phenol yield from the reaction of OH radicals with benzene of 0.61 ± 0.07 was measured by means of long-path FT-IR and UV spectroscopy over a wide range of experimental conditions. This yield was confirmed by measurements performed in the presence of NO_x. Detected carbonyls were glyoxal, cis-butenedial and trans-butenedial with formation yields of 0.29 ± 0.10, 0.08 ± 0.03 and 0.023 ± 0.007, respectively, measured in synthetic air and in the presence of NO_x. There was no significant difference in the product yields applying both experimental approaches for OH generation (dark reaction or photolysis). Nitrobenzene and o-nitrophenol were detected in traces. The yield of nitrobenzene increased with increasing NO _x resulting in a maximum formation yield of 0.007. The detected products in the presence of NO_x account for ～78% of the reacted carbon. Butenedial yields from benzene degradation are reported for the first time. In the absence of NO_x, glyoxal, cis-butenedial and trans-butenedial were also detected, but with distinctly lower yields compared to the experiments with NO_x. the Owner Societies 2006.

Full text of DOI:10.1039/b514148f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Formation of phenol and carbonyls from the atmospheric reaction of OH
radicals with benzene
Torsten Berndt* and Olaf Boge
¨
Received 5th October 2005, Accepted 19th January 2006
First published as an Advance Article on the web 8th February 2006
DOI: 10.1039/b514148f
The gas-phase reaction of OH radicals with benzene has been studied in a flow tube operated at
2
95 ꢀ 2 K and 950 mbar of synthetic air or O . Ozonolysis of tetramethylethylene (dark reaction)
2
with a measured OH radical yield of 0.92 ꢀ 0.08 or photolysis of methyl nitrite in the presence of
NO served as the OH sources. For investigations in the presence of NO , the conditions were
chosen so that more than 95% of the OH/benzene adduct reacted with O even for the highest
NO concentration occurring in the experiment. In the absence of NO , a phenol yield from the
x
2
2
x
reaction of OH radicals with benzene of 0.61 ꢀ 0.07 was measured by means of long-path FT-IR
and UV spectroscopy over a wide range of experimental conditions. This yield was confirmed by
measurements performed in the presence of NO . Detected carbonyls were glyoxal, cis-butenedial
x
and trans-butenedial with formation yields of 0.29 ꢀ 0.10, 0.08 ꢀ 0.03 and 0.023 ꢀ 0.007,
x
respectively, measured in synthetic air and in the presence of NO . There was no significant
difference in the product yields applying both experimental approaches for OH generation (dark
reaction or photolysis). Nitrobenzene and o-nitrophenol were detected in traces. The yield of
nitrobenzene increased with increasing NO
x
resulting in a maximum formation yield of 0.007. The
detected products in the presence of NO account for B78% of the reacted carbon. Butenedial
x
yields from benzene degradation are reported for the first time. In the absence of NO , glyoxal,
x
cis-butenedial and trans-butenedial were also detected, but with distinctly lower yields compared
to the experiments with NOx.
1
. Introduction
2
benzene adduct with NO can become important only in the
presence of high NO concentrations being used, as demon-
x
4,11,12
strated in a few cases under laboratory conditions.
Benzene represents the simplest aromatic compound and is
mainly released into the atmosphere by human activities, such
as emissions from burning oil and coal, motor vehicle exhaust,
and evaporation of solvents and from gasoline stations.
Stimulated by increasing incidence of leukaemia, the U.S.
Environmental Protection Agency has classified benzene as a
OH/benzene þ NO
Experimental studies with product analyses
2
- products
(2)
8–13
supported by
6,14–18
investigations on a theoretical level
were undertaken for
understanding the atmospheric degradation process. Because
of the relatively slow but atmospherically important reaction
1
Group-A human carcinogen. Besides the toxicity to humans,
conversion of benzene in the atmosphere can play a significant
role in tropospheric ozone formation as well as in the forma-
tion of particulate matter as demonstrated experimentally for
of the OH/benzene adduct with O
ꢂ1 ꢂ1
2
, k1,overall = (1.9–2.5) ꢁ
ꢂ16
3
cm molecule
4–6
at room-temperature, it is difficult
1
0
s
to suppress unwanted reactions in laboratory studies such as
the combination of the OH/benzene adduct with HO /RO or
the reaction with NO . The impact of NO concentrations on
2
the derivative 1,3,5-trimethylbenzene. The degradation of
2
2
benzene in the atmosphere is exclusively initiated by the attack
2
x
of OH radicals. Here, addition forming the OH/benzene
3
adduct dominates over the H-abstraction step. The fate of
product yields was shown in a previous report from this
11
12
laboratory as well as from a chamber study. Different
experimental conditions are probably the reason for the large
discrepancies in the formation yields of first generation pro-
ducts given in the literature so far.
the OH/benzene adduct is governed by its reaction with O
2
.
OH/benzene þ O - products (1)
2
The first elementary steps of this path (equilibrium with the
For the ring-retaining product phenol, a yield of 0.236 ꢀ
corresponding RO radical, irreversible overall reaction) were
2
0
.044 was reported as a result of a chamber study performed
the subject of intense kinetic studies and seem to be well
4
9
under NO
concentrations of NO
laboratory in a discharge flow tube resulting in a phenol yield
x
-rich conditions. A model considering different
–7
understood.
On the other hand, the reaction of the OH/
x
was applied to the data obtained in this
1
1
of 0.38 for atmospheric conditions. The highest yield of
Leibniz-Institut fu¨r Tropospha¨renforschung e.V., Permoserstr. 15,
4318 Leipzig, Germany. E-mail: berndt@tropos.de;
0
.531 ꢀ 0.066 comes from investigations performed in a large
0
1
chamber using mixing ratios for NO over a wide range. For
2
Fax: þ49 341 235-2325
x
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 1205–1214 | 1205

View Article Online
1
1
ꢂ3
the ring-cleavage product glyoxal, the yields span a range of
8
stances were about 10 molecule cm (averaging per data
point: 10 s).
,10
.2–0.35.
Further information on other reaction products is sparse.
0
For FT-IR analysis a Magna 750 spectrometer with MCT
detector (Nicolet) connected to a 10 m gas cell was utilised.
The formation of butenedials, being the corresponding co-
products of glyoxal, was proposed from theoretical investiga-
ꢂ1
250–2000 spectra (range: 4000–800 cm ) with a resolution of 1
ꢂ1
1
5
tions. Up to now, product studies do not confirm the
occurrence of these substances although an attempt to observe
or 8 cm were recorded for averaging. Phenol was detectable
1
2
ꢂ3
for concentrations Z 2 ꢁ 10 molecule cm
.
1
3
them failed. Isomers of hexa-2,4-dienedial (here predomi-
nantly the trans,trans species) were observed experimentally
under conditions of enhanced NO concentrations starting
Calibrated reference spectra (UV and FT-IR) of volatile
substances were measured after dilution of definite gas mix-
tures from a mixing device. For the solids (phenol and trans-
butenedial), a carrier gas stream was flushed over a sample and
after dilution spectra were recorded several times during a
definite time period. Under conditions of a nearly constant
sublimation during the whole time period (Dm/m is small,
nearly constant surface area and measured spectral intensity
are almost identical), spectra can be set on an absolute scale
after determination of the lost sample mass Dm, see also
ref. 19.
1
1
from the peroxyl radical of the OH/benzene adduct. This
pathway, however, can be neglected for atmospheric condi-
5
tions. Furthermore, the formation of epoxides (epoxy hexa-
1
,4-dienedial and epoxy butenedial) has been considered, but
7
2
experimental evidence has not been provided to date.
The aim of this work is to determine formation yields of first
generation products under experimental conditions that are
close to atmospheric ones. Special attention is given to phenol
and glyoxal, resolving the gap of existing formation yields
reported in literature. A reliable glyoxal yield is needed for an
assessment of benzene’s possible role in the formation of
secondary organic aerosol comparable to the process observed
In the GC-MS analysis, the concentrations of detected
substances were determined relative to the concentration of
phenol. For calibration, gas mixtures with phenol were simul-
taneously measured by means of FT-IR spectroscopy and
GC-MS. The determination of the absolute concentrations
from the FT-IR spectra allowed the calculation of the relative
response factors for the GC-MS analysis. In the MS, the
2
for 1,3,5-trimethylbenzene. The identification of further pro-
ducts is required for completion of the carbon balance allow-
ing a better insight into the reaction mechanism.
1
In contrast to the previous study from this laboratory using
a discharge technique as the OH radical source, here experi-
mental approaches with in situ formation of OH radicals were
chosen allowing investigations of OH radical levels close to
atmospheric conditions. Improvement of analytical techniques
including lowering of detection limits for oxidation products
permitted experiments over a wide range of conditions.
substances were measured using the [M] ion traces to im-
prove the sensitivity. Detection limits (signal/noise Z 3) were
as follows (flow-tube concentrations in molecule cm ); gly-
ꢂ3
1
0
9
9
oxal: 10 , cis- and trans-butenedial: 3 ꢁ 10 , phenol: 10 ,
9
10
9
furan: 10 , furfural: 10 and nitrobenzene: 2 ꢁ 10 .
3
The total gas flow-rate was set at 5000 standard cm min
ꢂ1
resulting in a bulk residence time of 10.4 min. Calibrated mass
flow controllers (MKS 1259 and 1179) were used and the
pressure was measured by a capacitive manometer (Baratron).
2
. Experimental
The gases had stated purities as follows: O
2 2
(99.999%), N
The experiments have been conducted in a flow tube (15.0 cm
(99.999%) (Linde) and NO (0.5 vol% mixture in N ) (Messer
2
i.d., length: 310 cm) operated with synthetic air (20 vol% O in
2
Griesheim).
Tetramethylethylene
(>99%),
benzene
N
2
) or O
2
as carrier gas at 295 ꢀ 2 K and 950 mbar. For
(>99.9%), phenol (>99.99%), nitrobenzene (99%), furan
(>99%) (Aldrich) and o-nitrophenol (>99%), furfural
(>99%) (Fluka) were used as purchased. Glyoxal monomer
photolysis, the tube was surrounded with four Silvania BLB
lamps (nominal power 36 W, lighted section: 125 cm) emitting
in the range 310–410 nm. Detection of gas-phase substances
has been achieved by means of long-path UV and FT-IR
spectroscopy pumping continuously a gas stream from the
flow tube through the corresponding gas cells. A smaller flow
was pumped through a heated GC-loop coupled with a pre-
focusing device allowing on-line GC-MS analysis (HP 6890
with separation column: HP 5MS and HP MSD 5973) for low
reactor concentrations. In the pre-focusing device the gas
2
0
was prepared as described by Schweitzer et al. cis-Butene-
2
dial was synthesized by reacting furan with dimethyldi-
1
2
2
oxirane. trans-Butenedial was made by hydrolysis of its
bis(dimethylacetal). For this, a solution of the bis(dimethyl-
acetal) in water–acetone was stirred at room temperature in
presence of a sulfonic acid–based cation exchange resin (Am-
berlyst 15). After filtration and concentration, the residue was
extracted with methylene chloride. The organic layer was dried
over sodium sulfate and concentrated yielding trans-butene-
dial as yellow crystals. The desired bis(dimethylacetal) of
trans-butenedial was synthesized by reacting furan with bro-
3
sample from the GC-loop (10 cm ) was flushed through an
uncoated, deactivated capillary (0.32 mm i.d.) cooled to liquid
nitrogen temperature at a distance of B5 cm. The collected
substances were injected directly at the separation column
after flash heating.
2
3
mine in methanol as described by Gree et al. The synthesis of
methyl nitrite was carried out by the method given in ref. 24.
3
ꢂ1
Measurements in the UV range were performed with a
Lambda 800 spectrometer (PerkinElmer) using path lengths
of 2.11 or 5.28 m. Phenol was measured between 260–280 nm
in a high resolution mode (slit: 0.22 nm, data spacing: 0.05 nm)
and benzene between 190–260 nm with lower resolution (slit:
In the case of ozonolysis reactions, 100 standard cm min
were flush through an ozone generator (UVP OG-2) and the
resulting O concentration was monitored by UV absorption
O
2
3
measurements at 255 nm (PerkinElmer, Lambda 2). The
addition of this O –O mixture to the main gas stream
3
2
1
nm; data spacing: 0.2 nm). Detection limits for both sub-
proceeded directly at the entrance to the flow tube.
1
206 | Phys. Chem. Chem. Phys., 2006, 8, 1205–1214
This journal is ꢃc the Owner Societies 2006

View Article Online
3
. Results and discussion
3
.1. OH radical yield from the ozonolysis of
tetramethylethylene (TME)
The ozonolysis of tetramethylethylene (TME) was chosen as a
‘
‘dark source’’ of OH radicals.
O
3
þ TME - x OH þ products
(3)
The OH radical yields reported in the literature span a range of
2
5
0
.65–1.0. Caused by the relatively large uncertainty of this
yield a reinvestigation was necessary. The experiments were
conducted in pure O or with an O
concentration of 4.5 ꢁ
molecule cm diluted in N with initial TME concentra-
2
2
1
7
ꢂ3
1
0
2
1
3
ꢂ3
12
tions of (3.5–3.6) ꢁ 10 molecule cm and (1.6–11.1) ꢁ 10
ꢂ3
molecule cm for O
3
. Under these conditions the consump-
Fig. 1 Extra TME consumption regarding pathway (3), D[TME]/
3
tion of O was complete (>99.9%) after the bulk residence
[O ]
3
0
ꢂ 1, as a function of total TME conversion for different O
2
time of 10.4 min even for an assumed OH radical yield of 1.0
concentrations in the carrier gas. The dashed line represents simple
data averaging. The dotted line shows the result from a linear fit and
the full result from fitting using a quadratic equation.
and conversion of OH radicals via the consecutive reaction
ꢂ15
3 ꢂ1 ꢂ1 26
cm molecule s ,
with TME, k = 1.0 ꢁ 10
ꢂ1 ꢂ1 27
s .
k4 = 1.1 ꢁ
3
ꢂ10
3
1
0
cm molecule
deduced from the axis intercept (extrapolation to ‘‘zero’’
TME conversion) is hardly affected by the unwanted reactions
and an OH radical yield of 0.92 ꢀ 0.08 can be recommended as
a result of this study. It is to be noted, that the quadratic
equation was not derived from resulting differential equations
describing an assumed mechanism including the unwanted
OH þ TME - products
(4)
The consumption of OH radicals via the reaction with acetone,
the main product from
OH þ acetone - products
(5)
pathways (3) and (4), is of less importance caused by the low
ꢂ13
reactions of OH radicals or O
3
with products.
3 ꢂ1 ꢂ1 27
cm molecule s .
reactivity of acetone, k
5
= 2.3 ꢁ 10
3
.2. Determination of benzene conversion in ozonolysis
2
28
An estimate (kwall B 4D/r , D: diffusion coefficient; r: tube
radius) yields first order rate coefficients of 0.02 s for the
experiments (absence of NOx)
ꢂ1
ꢂ5
diffusion controlled wall loss of OH radicals and 3.2 ꢁ 10
The majority of the experiments was conducted with a very
large excess of benzene making (i) the consecutive reactions of
products negligible and (ii) pathway (6) dominant over path-
way (4) in the case of ozonolysis for OH radical formation,
ꢂ1
s
3
for the wall loss of O applying the experimentally
obtained collision efficiency of 1/425 taken from a former
2
study. Both k-values point at negligible wall loss for OH
6
= 1.26 ꢁ 10ꢂ12 cm molecule
3
ꢂ1 ꢂ1 27
s .
radicals and O as long as the TME concentrations do not fall
3
k
6
1
below 10 molecule cm . Other OH radical or O consuming
2
ꢂ3
3
OH þ benzene - products
The large excess of benzene prevented the direct spectroscopic
determination of the amount of reacted benzene. Considering
pathway (3) with an OH radical yield of 0.92 and pathways (4)
(6)
steps with any products or intermediates can not fully be ruled
out. However, they should be repressed with lowering the
TME conversion. Fig. 1 shows the extra TME consumption
regarding pathway (3), D[TME]/[O
3
]
0
ꢂ 1, as a function of
total TME conversion using FT-IR measurements. The extra
TME consumption is attributed to pathway (4) and represents
the OH radical yield from the ozonolysis of TME. Simple
averaging of all data leads to an OH radical yield of pathway
and (6), eqn (I) follows for total O
3
conversion from the
, TME and C
balance of the amount of substances for O
3
6 6
H .
D[C
H
6 6
] = 1.92 ꢁ [O
3
]
0
ꢂ D[TME]
(I)
(
3) of 0.89 ꢀ 0.07 (two standard deviations), cf. the dashed line
The validity of eqn (I) was controlled in a series of experiments
and compared with results from a modelling study considering
in a simple version pathways (3), (4) and (6) only. Initial
in Fig. 1. An averaging, however, does not take into account
unwanted steps being noticeable for higher TME conversions.
Using a linear fit (axis intercept: 0.93 ꢀ 0.04), in principle, a
contribution of these unwanted steps is considered, but with a
linear term only, cf. the dotted line in Fig. 1. The unwanted
ꢂ3
concentrations were (unit: molecule cm ): [TME] = 3.5 ꢁ
1
3
15
12
10 , [C H ] = 8.9 ꢁ 10 and [O ] = (4.4–16.7) ꢁ 10 . FT-IR
6
6
3
measurements were chosen for the determination of the TME
model
conversion, D[TME], needed in eqn (I). The plot D[C H ]
6 6
versus D[C
0.04 (intercept: (2.8 ꢀ 5.6) ꢁ 10 molecule cm ). The
excellent agreement between the results from the modelling
with those from the measurements using eqn (I) indicates the
absence of further, noticeable OH radical consuming steps.
This fact was confirmed using an enhanced model considering
also intermediates and products of pathways (3), (4) and (6)
reactions of OH radicals or O
processes and for higher TME conversions a curvature of the
plot (D[TME]/[O
more adequate description of the measurements is given using
a quadratic equation with a non-linear term (axis intercept:
3
with products are non-linear
eqn (I)
6
H
6
]
gave a straight line with the slope: 0.99 ꢀ
11 ꢂ3
3
]
0
ꢂ 1) vs D[TME] is expected. Therefore, a
0
.92 ꢀ 0.08), cf. the full line in Fig. 1. The latter analysis
describes well the slight curvature of the data points for
relatively large TME conversions. The OH radical yield
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 1205–1214 | 1207

View Article Online
Table 1 Reaction mechanism used for estimation of OH radical concentrations
3
ꢂ1 ꢂ1
ꢂ1
or s
Step
k
295K / cm molecule
s
Source
ꢂ15
ꢂ10
O
3
þ TME - 0.92 OH þ products
1.0 ꢁ 10
1.1 ꢁ 10
26
27
a
OH þ TME þ (O
OH/TME–O
þ NO - products þ NO
OH þ benzene - OH/benzene
OH/benzene þ O - 0.61 phenol þ 0.61 HO
OH/benzene–O
þ NO - products þ NO
- products
- OH þ 2O
þ HO - H þ O
- NO
þ OH
- NO
þ O
Photolysis: CH
ONO þ (O
þ (O ) - NO þ O
2
) - OH/TME–O
2
ꢂ11
2
2
þ HO
2
10
1.26 ꢁ 10
Estimated
27
4
Estimated
4
31
31
31
ꢂ12
ꢂ16
2
2
þ 0.39 OH/benzene–O
2
2 ꢁ 10
ꢂ11
2
2
þ HO
2
10
ꢂ11
ꢂ15
ꢂ12
ꢂ12
ꢂ14
ꢂ4
OH/benzene þ NO
2
2.7 ꢁ 10
1.8 ꢁ 10
1.7 ꢁ 10
8.5 ꢁ 10
1.7 ꢁ 10
7.8 ꢁ 10
4.7 ꢁ 10
HO
HO
2
þ O
3
2
2
2
2
O
2
2
NO þ HO
NO þ O
2
2
3
2
2
31
Measured
Measured
a
3
2
) - NO þ HO
2
þ CH
2
O
a
ꢂ3
Photolysis: NO
2
2
3
a
2 2
(O ) denotes that O is not involved in the rate limiting step.
1
2
with the corresponding rate coefficients available in literature.
The differences of calculated D[C ] values from the ‘‘sim-
ple’’ and the ‘‘enhanced’’ model were small. In the following,
modelling including pathways (3), (4) and (6) was used for
determining the amount of converted benzene for ozonolysis
experiments a fixed initial O
ꢂ3
3
concentration of 7.0 ꢁ 10
H
6 6
molecule cm was chosen and the benzene concentration was
1
4
ꢂ3
changed from 8.9 to 118 ꢁ 10 molecule cm . The initial
13 ꢂ3
TME concentration was 3.5 ꢁ 10 molecule cm in each case
ensuring total O conversion. The calculated maximal OH
3
6
reactions in the absence of NO . The usage of this model
x
radical concentrations were in the range (0.94–45) ꢁ 10
ꢂ3
permits also a simple estimate of OH radical concentrations to
ꢂ3
molecule cm . Phenol yields from the different sets did not
6
be in the range (9.4–35) ꢁ 10 molecule cm for the initial
show a dependence on any experimental condition, see Fig. 2.
Therefore, effects of unwanted reactions of the OH/benzene
adduct leading to a decrease or increase of the phenol yield can
be ruled out. The lifetime of the OH/benzene adduct regarding
conditions chosen in this set of experiments.
3
.3. Reaction mechanism used for the estimation of OH
ꢂ4
radical concentrations
the overall O reaction (pathway (1)) in pure O is B2 ꢁ 10
2
2
4
s. For a supposed k-value of 10
–6
ꢂ10
3
cm molecule s , the
ꢂ1 ꢂ1
Modelling calculations have been performed in order to
estimate the OH radical concentrations present in the flow
tube for the different conditions (ozonolysis in the absence or
lifetime of the OH/benzene adduct regarding a reaction with a
ꢂ2
supposed ‘‘reactive species’’ is >2 ꢁ 10
s assuming a
maximal formation yield of unity for the ‘‘reactive species’’
1
x
presence of NO as well as photolysis), cf. Table 1. Attention
1
from ozonolysis (initial O concentration of 4.4 ꢁ 10 mole-
3
was concentrated on making the description of the main HOx-
ꢂ3
cule cm ). With increasing initial O concentration, the plot
3
and NO -steps as accurate as possible. This mechanism is not
x
of the phenol yield vs reacted benzene would show a curvature
if the unwanted reaction contributes significantly. The range
of reacted benzene in this study covers approximately two
orders of magnitude and such a curvature is not visible, cf.
Fig. 2. Therefore, averaging of all data was justified resulting
appropriate for the modelling of product formation. The given
photolysis coefficients were determined under commonly used
reaction conditions using FT-IR spectroscopy for reactant or
2
product measurements. In the case of methyl nitrite, CH O
formation was followed. In order to prevent any interference,
these experiments were conducted in the presence of a large
excess of NO initiating HO
2
þ NO - OH þ NO
2
and
subsequently OH þ NO - HONO. NO
2
photolysis experi-
ments have been performed in pure nitrogen using TME as the
O-atom scavenger. Here the photolysis coefficient was deter-
mined from the measured NO decay.
2
For each set of experiments, the OH radical profiles in the
flow tube were calculated for the different initial concentra-
tions chosen. The range of maximal OH radical values found
for the several profiles of each set of experiments is given in the
beginning of the corresponding paragraph.
3
.4. Formation of phenol
.4.1. Absence of NO . Four sets of experiments have been
conducted using synthetic air or O as the carrier gas. Phenol
was measured either by means of UV or FT-IR spectroscopy.
In the first three sets the initial O was varied in the range (4.4–
3
x
2
3
Fig. 2 Phenol yields in the absence of NO measured in a wide range
x
1
1
ꢂ3
1
95) ꢁ 10 molecule cm keeping the benzene concentration
of experimental conditions; OH radical generation by ozonolysis of
1
5
ꢂ3
constant at 8.9 ꢁ 10 molecule cm . In the 4th set of
TME.
1
208 | Phys. Chem. Chem. Phys., 2006, 8, 1205–1214
This journal is ꢃc the Owner Societies 2006

View Article Online
11
ꢂ3
Fig. 3 (a) Measured difference UV spectrum (photolysis on/off) with a benzene conversion of 6.7 ꢁ 10 molecule cm . A: raw spectrum, B: after
subtraction of phenol, C: after addition of reacted benzene. (b) Reference spectra of benzene (1.12 ꢁ 10 molecule cmꢂ3) and phenol (1.07 ꢁ 10
1
3
13
ꢂ3
molecule cm ).
in a phenol yield of 0.61 ꢀ 0.07 (two standard deviations,
n = 67; n is the number of experimental points).
Dividing the resulting rate law for phenol by the rate law for
benzene, eqn (II) follows:
3
x
.4.2. Presence of NO . The photolysis of methyl nitrite
d[phenol]/d[benzene]=ꢂy þ k /k6a ꢁ [phenol]/[benzene] (II)
7
2
was applied for OH radical generation using synthetic air as
4
the carrier gas. Measured conversion of the precursor was
1
Eqn (II) was solved numerically by an explicit extrapolation
method connected with a Newton-technique for determining
1
ꢂ3
about 15–20% inserting at most 2 ꢁ 10 molecule cm of
3
0
NO into the reaction system for an initial methyl nitrite
ꢂ3
the formation yield of phenol y. The comparison of phenol
yields determined by means of eqn (II) and otherwise from the
measured concentrations, [phenol]/D[benzene], showed that
7–20% of phenol formed via pathway (6a) was consumed by
pathway (7). Phenol yields corrected due to the consecutive
reaction (7) were found to be independent on initial NO
concentration as can be seen in Fig. 4. Data averaging gives
a phenol yield of 0.61 ꢀ 0.06 (two standard deviations, n =
18). This value is in excellent agreement with the finding from
measurements using ozonolysis for OH radical formation in
absence of NOx (dark reaction), see the paragraph before.
Therefore, a phenol yield of 0.61 ꢀ 0.07 can be recommended
as a result of all data.
1
2
concentration of 1.0 ꢁ 10 molecule cm . Additionally
1
1
ꢂ3
added NO was in the range (3.3–13.2) ꢁ 10 molecule cm
1
2
giving rise to a maximal NO concentration of 1.5 ꢁ 10
x
ꢂ3
molecule cm . Under these conditions more than 95% of the
OH/benzene adduct reacted with O
competition with pathway (2) even for the case in which all
NO was converted to NO . Here, the calculated maximal OH
2
via pathway (1) in
2
8
ꢂ3
radical concentrations were (1.7–5.5) ꢁ 10 molecule cm
.
1
3
ꢂ3
The benzene concentration (5.4 ꢁ 10 molecule cm ) was
kept relatively low making the UV-spectroscopic determina-
tion of converted benzene possible. For demonstration, in Fig.
3
a UV spectra from a measurement with a benzene conversion
1
1
ꢂ3
of 6.7 ꢁ 10 molecule cm are depicted. Spectral regions
used for the subtraction procedure were 260–280 nm for
phenol and 195–210 nm for benzene. Fig. 3b shows the spectra
of the reference substances for comparison. Residual absorp-
tions in the region below 230 nm of spectrum C in Fig. 3a can
be explained partly by the occurrence of butenedial as a
product, see later.
The phenol yield from this study is the highest value
reported so far. Table 2 summarizes literature data for near-
atmospheric conditions. The result of this study is roughly in
line with findings from a chamber reporting a phenol yield of
1
2
0.531 ꢀ 0.066. It has to be noted that in ref. 12 the majority
of the measurement points for mixing ratios of NO and NO
o20 ppb are above the modelling curve describing the forma-
tion yield stated, cf. Fig. 6 in ref. 12. In another chamber study
reporting a phenol yield from atmospheric benzene degrada-
2
On the other hand, under these experimental conditions the
consecutive reaction of phenol with OH radicals can not be
3
neglected and corrections are necessary, k
ꢂ1 ꢂ1 29
7
= 2.8 ꢁ 10ꢂ11 cm
tion for the first time, the NO concentrations were obviously
x
9
too high. Under these conditions the reaction of the OH/
molecule
s
.
benzene adduct with NO via pathway (2) has to be taken into
2
OH þ benzene - y phenol þ products
OH þ phenol - products
(6a)
(7)
account leading to a decrease in the phenol yield. The model-
ling of NO -dependent data from experiments with relatively
x
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 1205–1214 | 1209

View Article Online
Fig. 4 Phenol yield dependence on initial NO concentration for
measurements performed in synthetic air; OH radical generation by
photolysis of methyl nitrite.
x
high NO concentrations from this laboratory resulted in a
1
1
phenol yield of 0.38 for atmospheric conditions. Beside the
uncertainty arising from modelling itself, unwanted reactions
caused by high radical concentrations (discharge technique)
can not fully be ruled out.
3
.5. Formation of carbonyls and nitro compounds
.5.1. Absence of NO . Formation yields of carbonyls were
determined under NO -free conditions in synthetic air using
3
x
Fig. 5 Relative product yields of detected carbonyls (regarding
x
phenol) as a function of converted benzene measured in synthetic air
ozonolysis of TME for OH radical formation analogous to the
approach for phenol, see paragraph 3.4.1. Initial concentra-
and in the absence of NO ; OH radical generation by ozonolysis of
x
TME.
ꢂ3
13
tions were (unit: molecule cm ): [TME] = 3.5 ꢁ 10 , [C H ]
6
6
1
5
11
=
8.9 ꢁ 10 and [O ] = (1.2–257) ꢁ 10 . The calculations
3
yielded maximal OH radical concentrations for these condi-
concentrations and, therefore, relatively high radical levels) an
efficient growth of the concentrations for furfural and furan
was measured. The relative concentrations of furan regarding
phenol remained below 0.01. The occurrence of furfural and
furan points at a modification of the pathways being impor-
6
ꢂ3
tions of (0.25–55) ꢁ 10 molecule cm . Product analysis was
performed by means of on-line GC-MS. The presence of large
amounts of acetone from TME conversion prevented a reliable
FT-IR analysis of the carbonyls originating from benzene. In
Fig. 5 the experimental findings are given as a function of the
amount of reacted benzene. The relative product yields of
glyoxal and cis-butenedial showed a slight decrease with
tant for the fate of the OH/benzene–O
2
radical under these
circumstances. Relative product yields regarding phenol can
1
be given for a benzene conversion below 10 molecule cm ;
2
ꢂ3
1
2
increasing benzene conversion for D[C H ] > 2 ꢁ 10 mole-
glyoxal: 0.261 ꢀ 0.091, cis-butenedial: 0.082 ꢀ 0.027 and trans-
butenedial: o0.023. Using the phenol yield from this study,
6
6
ꢂ3
cule cm . For the highest benzene conversions (high initial O
3
Table 2 Reported phenol yields at room temperature for near-atmospheric conditions
Experimental
approach
Initial NO /
molecule cm
Initial NO
2
/
molecule cm
Phenol
yield
ꢂ3
ꢂ3
T/K
Pressure
Ref.
9
1
4
14
Chamber, air, GC-FID analysis,
photolysis
Flow tube, O , FT-IR analysis,
(0–2.5) ꢁ 10
(0–2.4) ꢁ 10
10a
o10
298 ꢀ 2
B740 torr
0.236 ꢀ 0.044
>0.30 ꢀ 0.02 (0.38)
0.531 ꢀ 0.066
0.61 ꢀ 0.07
1
2
b
2
(8.2–12.3) ꢁ 10
293 ꢀ 2
287–307
295 ꢀ 2
500 mbar
Ambient
950 mbar
950 mbar
11
discharge techn.
c
c
Chamber, air, FT-IR and DOAS
analysis, photolysis
Flow tube, O or air, UV and FT-IR
analysis, ozonolysis
Flow tube, air, UV analysis, photolysis
12
10a
o10
10a
o10
2
This study
This study
11
10a
(3.3–13.2) ꢁ 10
o10
analyser. From modelling of NO
295 ꢀ 2
0.61 ꢀ 0.06
with a mixing ratio o0.2 ppb–2.0 ppm
a
b
c
dependent data. Only total NO
Below the detection limit of a NO
is stated.
x
x
x
1
210 | Phys. Chem. Chem. Phys., 2006, 8, 1205–1214
This journal is ꢃc the Owner Societies 2006

View Article Online
3
2
presence of NO is 1.7 ꢀ 0.1 and in the absence of NO it can
not be higher than 2.0.) Consequently, under these conditions
the reaction of NO with HO represents the predominant OH
2
radical source. Nevertheless, the ozonolysis is needed for
initialization of HO in the system. The calculated maximal
x
OH radical concentrations for these experiments were in the
6
ꢂ3
range (0.79–5.9) ꢁ 10 molecule cm
.
2
As HO is converted by NO, peroxyl radicals have also to
react in this way. The resulting, relative product yields of the
carbonyls increased with respect to NO -free conditions, cf.
Fig. 6 and Fig. 5. Relative product yields regarding phenol
x
1
1
ꢂ3
were as follows ([NO] > 2.5 ꢁ 10 molecule cm ); glyoxal:
0
.516 ꢀ 0.151, cis-butenedial: 0.137 ꢀ 0.038 and trans-bute-
nedial: o0.038. Using again the phenol yield from this study,
formation yields (regarding converted benzene) of 0.31 ꢀ 0.10,
0
.08 ꢀ 0.03 and o0.023, respectively, can be derived. For
trans-butenedial, the same behaviour (decreasing yield with
decreasing benzene conversion) was observed as in the absence
of NO
Furthermore, the produced NO
FT-IR spectroscopy with an average yield of 1.02 ꢀ 0.05
regarding initial NO. Total NO conversion forming NO is
x
. A mechanistic explanation of this remains speculative.
2
was measured by means of
2
in accordance with a radical chain process recycling OH
radicals via pathway (8). This fact indicates also that
N-containing organics such as nitro-compounds and organic
nitrates were not formed in substantial amounts.
Fig. 6 Relative product yields of carbonyls (regarding phenol) as well
as the ratio [NO ]/[NO]initial as a function of initial NO concentration
2
3
x
.5.3. Presence of NO (photolysis). Photolysis experi-
for measurements in synthetic air; OH radical generation by ozono-
lysis of TME. For explanation of the normalized acetone data see text.
1
2
ꢂ3
ments with methyl nitrite (1.0 ꢁ 10 molecule cm ) in
synthetic air were undertaken for an assessment of the possible
influence of intermediates from ozonolysis or the OH reaction
with TME on the pathways yielding the products of benzene;
formation yields (regarding converted benzene) of 0.16 ꢀ 0.06,
1
5
ꢂ3
0
.05 ꢀ 0.02 and o0.014, respectively, follow. Errors given
[C H ] = (2.9–4.5) ꢁ 10 molecule cm and initial [NO] =
6
6
11 ꢂ3
represent two standard deviations. No clear explanation can
be given for the decreasing yield of trans-butenedial with
decreasing benzene conversion visible in the whole measure-
ment range.
(4.5–14.2) ꢁ 10 molecule cm . Calculated maximal OH
6
radical concentrations were in the range (2.8–11) ꢁ 10
ꢂ3
molecule cm . Fig. 7 shows the findings from these experi-
ments. The relative product yields of the carbonyls (incl. trans-
1
1
x
butenedial!) did not show a dependence on the NO concen-
3
.5.2. Presence of NO (ozonolysis). NO ((1.4–14.9) ꢁ 10
x
ꢂ3
tration; glyoxal: 0.443 ꢀ 0.082, cis-butenedial: 0.131 ꢀ 0.020
and trans-butenedial: 0.038 ꢀ 0.011. The resulting formation
yields regarding converted benzene are 0.27 ꢀ 0.06, 0.08 ꢀ
0.02 and 0.023 ꢀ 0.007, respectively. A more comprehensive
GC-MS analysis allowed the detection of nitrobenzene and
o-nitrophenol additionally. The latter compound was measured
qualitatively only. There was a considerable scatter of the
product yield, obviously caused by strong wall interactions.
The product yield of nitrobenzene increased with increasing
molecule cm ) was added to the carrier gas starting from the
same conditions as the experiments in the absence of NOx
(
cf. paragraph 3.5.1) for the lowest initial O
1
3
concentration of
1
ꢂ3
1
.2 ꢁ 10 molecule cm . Besides the desired reaction of NO
with peroxyl radicals, NO recycles OH radicals in the reaction
ꢂ12
3
with HO
molecule
2
as well as consuming O
3
, k
8
= 8.5 ꢁ 10
ꢂ1 ꢂ1 31
s .
cm
ꢂ1 ꢂ1 31
= 1.7 ꢁ 10ꢂ14 cm molecule
3
s
and k
9
NO þ HO - OH þ NO
(8)
(9)
2
2
NO resulting in a maximum value of 0.011 for the relative
x
NO þ O
3
- NO
2
þ O
2
yield regarding phenol and, consequently, 0.007 regarding
converted benzene. Fig. 8 shows two successively performed
GC-MS analyses (with and without photolysis) demonstrating
the signal strength of products using this analytical approach.
In Fig. 6, the normalized yield of acetone regarding phenol is
plotted using GC-MS measurements of the corresponding
1
M] ion traces. The ratio of the ion intensity obtained in
[
the absence of NO was set to 1.0. Acetone is a main product of
the TME conversion via pathways (3) and (4). With increasing
NO concentrations, the relative acetone yield drops clearly
indicating that ozonolysis via pathway (3) is repressed and
pathway (4) is mainly responsible for the detected acetone.
2
The ratio [NO ]/[NO]initial derived from FT-IR measurements
took a value of B1.4 for the lowest NO concentration and
dropped to a value close to 1.0 for increasing NO content in
the carrier gas. This behaviour reflects the additional NO input
1
1
ꢂ3
of B2 ꢁ 10 molecule cm arising from photolysis of methyl
(
Note, the formation yield of acetone from pathway (4) in the
nitrite. Furthermore, these measurements confirm the total
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 1205–1214 | 1211

View Article Online
Fig. 8 Ion chromatogram (ion traces: 58, 84, 94, 110, 123, 139 amu)
x
of a GC-MS analysis from a run in the presence of NO and with OH
radical generation by photolysis of methyl nitrite; lower part: without
photolysis, upper part photolysis was switched on. Initial concentra-
ꢂ3
12
tions were (unit: molecule cm ); methyl nitrite: 1.0 ꢁ 10 , benzene:
1
5
12
3
.0 ꢁ 10 and [NO] = 1.4 ꢁ 10 . The impurity detected at 7.2 min
was identified as methyl cyclohexane, the other was not identified.
3
.6. Mechanism
In Scheme 1 possible pathways leading to the products ob-
served in the presence of NO are depicted. This reaction
pattern is mainly based on results from theoretical investiga-
x
Fig. 7 Relative product yields of carbonyls and nitrobenzene (re-
garding phenol) as well as the ratio [NO ]/[NO]initial as a function of
6
,14–17
tions.
1
Starting from the OH/benzene adduct (I ), two
2
initial NO concentration for measurements in synthetic air; OH radical
generation by photolysis of methyl nitrite.
conversion of NO to NO as observed in the case of ozonolysis
2
used for initialization of HO , see paragraph 3.5.2.
x
In Table 3 the findings from this study are compared with
data from the literature. The results of this study in the
x
presence of NO using ozonolysis (dark reaction) as well as
photolysis do not show significant differences allowing aver-
aging of the formation yields. The yield of glyoxal is roughly in
1
0
line with the finding from a chamber study using similar NO
conditions. However, the glyoxal yield measured for elevated
x
8
x
NO concentrations was distinctly lower pointing again at a
significant occurrence of the unwanted reaction of the OH/
benzene adduct with NO . For both isomers of butenedial
2
there are no further data available in the literature. Despite the
unexplained drop of the trans-butenedial with decreasing
benzene conversion in the case of ozonolysis, here, the max-
imum yield matches the average yield arising from the photo-
lysis experiments. There were no indications for cis/trans
isomerisation of the butenedials by photolysis. The yield of
the trace product nitrobenzene was clearly dependent on the
NO
ported from a chamber study can be explained by the elevated
NO levels used.
x
concentration. Therefore, the much higher yields re-
9
x
Table 3 includes also the results from experiments of this
study performed in the absence of NO . For comparison, in
x
the literature there are no data describing formation yields of
carbonyls measured under near-atmospheric conditions in the
Scheme 1 Possible mechanism for the reaction of OH radicals with
benzene in the presence of NOx.
1
0
ꢂ3
absence of NO
x
, ([NO
x
] o10 molecule cm ).
1
212 | Phys. Chem. Chem. Phys., 2006, 8, 1205–1214
This journal is ꢃc the Owner Societies 2006

View Article Online
pathways are of importance under atmospheric conditions, (i)
H-abstraction by O
2
forming phenol and (ii) the reaction with
O leading to the corresponding OH/benzene–O radical (I )
2
2
2
5
being in an equilibrium with the reactants. It is to be noted
that theoretical methods describing the phenol channel pre-
6
dicted a formation yield of B0.55 being in excellent agree-
ment with the experimental value from ref. 12 and from this
study. The reaction of NO with the peroxyl radical (I
2
)
forming the isomers of hexa-2,4-dienedial as main products
5
,11
is negligible for atmospheric conditions.
In contrast, as a
result of cyclization a bicyclic endo-peroxyl allyl radical (I
3
)
1
6
can be produced. This allyl-type radical (I
reaction with O forming the corresponding peroxyl radical
I ) or (ii) a further rearrangement resulting in an epoxy oxyl
3
) can undergo (i) a
2
(
5
1
7
radical (I4). This epoxy oxyl radical (I ) is regarded as a
4
3
source for possible products with an epoxide structure. After
3
the reduction of the peroxyl radical (I ) by NO and successive
5
6
b-scissions of the corresponding oxyl radical (I ), the forma-
tion of glyoxal and butenedial in an equimolar ratio can be
explained. The experimental finding from this study, however,
states much higher yields for glyoxal than for the butenedials
making an additional pathway for glyoxal necessary. Starting
from the epoxy oxyl radical (I
4
), a possible channel was
proposed forming glyoxal and epoxy butenedial as the corres-
1
ponding C -hydrocarbon. In competition to that, also the
7
4
1
7
formation of epoxy hexa-2,4-dienedial was discussed.
The formation of o-nitrophenol and nitrobenzene can be
due to the reaction of the OH/benzene adduct (I ) with NO . A
1
2
more detailed description of possible channels yielding nitro-
phenol is pointed out in ref. 9.
Under NO
system react in a self-reaction or with other peroxyl radicals
including HO
x
-free conditions, the peroxyl radicals in the
2
. Products of the reaction peroxyl radical þ
peroxyl radical are the corresponding oxy radicals and in
competition alcohols and carbonyls from the disproportiona-
3
tion. The reaction of peroxyl radicals with HO leads to the
4
2
3
4
formation of hydroperoxides. Consequently, these channels
produce lower oxyl radical yields in comparison with the
corresponding NO reaction resulting in a decrease of the
formation yields of glyoxal and the butenedials.
4. Summary
The formation of phenol and carbonyls from the reaction of
OH radicals with benzene was studied using a flow-tube
experiment. OH radicals were produced in situ either by means
of ozonolysis of tetramethylethylene (dark reaction) with an
observed OH radical yield of 0.92 ꢀ 0.08 or by photolysis of
methyl nitrite. Experimental conditions were chosen in such a
way that the findings should be applicable for atmospheric
conditions.
Phenol yields were measured over a wide range of condi-
tions in the absence and presence of NO
pure O as the carrier gas. An average formation yield of 0.61
0.07 is reported being valid for atmospheric conditions.
In the presence of NO , average formation yields of glyoxal,
x
using synthetic air or
2
ꢀ
x
cis-butenedial and trans-butenedial measured in synthetic air
were 0.29 ꢀ 0.10, 0.08 ꢀ 0.03 and 0.023 ꢀ 0.007, respectively.
There was no significant difference in the product yields
This journal is ꢃc the Owner Societies 2006
Phys. Chem. Chem. Phys., 2006, 8, 1205–1214 | 1213

View Article Online
applying both experimental approaches for OH generation
dark reaction or photolysis). In addition, traces of nitroben-
10 R. Volkamer, U. Platt and K. Wirtz, J. Phys. Chem. A, 2001, 105,
7
865.
1 T. Berndt and O. Bo
946.
(
1
¨
ge, Phys. Chem. Chem. Phys., 2001, 3,
zene and o-nitrophenol were detected. The yield of nitroben-
zene increased with increasing NOx and a formation yield
r0.007 can be given. The products detected in the presence of
4
12 R. Volkamer, B. Klotz, I. Barnes, T. Imamura, K. Wirtz, N.
Washida, K. H. Becker and U. Platt, Phys. Chem. Chem. Phys.,
2
002, 4, 1598.
NO under near-atmospheric conditions account for B78% of
x
1
3 J. F. Hamilton, A. C. Lewis, C. Bloss, V. Wagner, A. P. Hender-
son, B. T. Golding, K. Wirtz, M. Martin-Reviejo and M. Pilling,
Atmos. Chem. Phys., 2003, 3, 1999.
the reacted carbon.
x
In the absence of NO , the formation yields of the carbonyls
1
1
4 G. Ghigo and G. Tonachini, J. Am. Chem. Soc., 1998, 120, 6753.
5 T. H. Lay, J. W. Bozzelli and J. H. Seinfeld, J. Phys. Chem., 1996,
were distinctly lower; glyoxal: 0.16 ꢀ 0.06, cis-butenedial:
0
.05 ꢀ 0.02 and trans-butenedial: o0.014.
1
6 G. Ghigo and G. Tonachini, J. Am. Chem. Soc., 1999, 121, 8366.
00, 6543.
Further experimental work is needed in order to identify the
1
missing products. A more complete carbon balance as well as
experiments for different conditions should give a better in-
sight into the reaction mechanism and, therefore, a better
understanding of the atmospheric benzene degradation
process.
17 F. Motta, G. Ghigo and G. Tonachini, J. Phys. Chem. A, 2002,
06, 4411.
8 C.-C. Chen, J. W. Bozzelli and J. T. Farrell, J. Phys. Chem. A,
004, 108, 4632.
19 T. Berndt and O. Bo
1
1
2
¨
ge, Phys. Chem. Chem. Phys., 2003, 5, 342.
20 F. Schweitzer, L. Magi, P. Mirabel and Ch. George, J. Phys. Chem.
A, 1998, 102, 593.
2
2
2
2
2
1 B. J. Adger, C. Barrett, J. Brennan, P. McGuigan, M. A. McKer-
vey and B. Tarbit, J. Chem. Soc., Chem. Commun., 1993, 1220.
2 W. Adam, J. Bialas and L. Hadjiarapoglou, Chem. Ber., 1991, 124,
2377.
Acknowledgements
The authors thank K. Pielok and H. Macholeth for technical
assistance.
3 R. Gree, H. Tourbah and R. Carrie, Tetrahedron Lett., 1986, 27,
983.
4
4 R. Atkinson, S. M. Aschmann, A. M. Winer and J. N. Pitts, Jr, Int.
J. Chem. Kinet., 1982, 14, 507.
5 M. Siese, K. H. Becker, K. J. Brockmann, H. Geiger, A. Hofzu-
mahaus, F. Holland, D. Mihelcic and K. Wirtz, Environ. Sci.
Technol., 2001, 35, 4660 and references therein.
References
1
U.S. Environmental Protection Agency, Integrated Risk Informa-
tion System (IRIS) on Benzene, National Centre for Environmen-
tal Assessment, Office of Research and Development, Washington,
DC, 2002.
¨
26 M. Witter, T. Berndt, O. Boge, F. Stratmann and J. Heintzenberg,
Int. J. Chem. Kinet., 2002, 34, 394.
27 R. Atkinson, Chem. Rev., 1986, 86, 69.
28 R. C. Reid, J. M. Prausnitz and B. E. Poling, The Properties of
Gases and Liquids, McGraw-Hill, New York, 1987.
29 M. Rinke and C. Zetzsch, Ber. Bunsen-Ges. Phys. Chem., 1984, 88,
55.
30 J. Stoer and R. Bulirsch, Einfu¨hrung in die Numerische Mathematik
II, Springer Verlag, Berlin, 1978.
31 R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson Jr, J. A.
Kerr, M. J. Rossi and J. Troe, J. Phys. Chem. Ref. Data, 1997, 26,
1329.
32 H. Niki, P. D. Maker, C. M. Savage, L. P. Breitenbach and M. D.
Hurley, J. Phys. Chem., 1987, 91, 941.
33 L. J. Bartolotti and E. O. Edney, Chem. Phys. Lett., 1995, 245, 119.
34 T. J. Wallington, O. J. Nielsen and J. Sehested, in Peroxy Radicals,
ed. Z. B. Alfassi, John Wiley & Sons, Chichester, 1997, p. 113.
2
M. Kalberer, D. Paulson, M. Sax, M. Steinbacher, J. Dommen, A.
S. H. Prevot, R. Fisseha, E. Weingartner, V. Frankevich, R.
Zenobi and U. Baltensperger, Science, 2004, 303, 1659.
R. Atkinson, J. Phys. Chem. Ref. Data Monogr., 1994, 2, 47.
R. Knispel, R. Koch, M. Siese and C. Zetzsch, Ber. Bunsen-Ges.
Phys. Chem., 1990, 94, 1375.
3
4
5
6
B. Bohn and C. Zetzsch, Phys. Chem. Chem. Phys., 1999, 1, 5097.
S. Raoult, M.-T. Rayez, J.-C. Rayez and R. Lesclaux, Phys. Chem.
Chem. Phys., 2004, 6, 2245.
7
8
9
D. Johnson, S. Raoult, M.-T. Rayez, J.-C. Rayez and R. Lesclaux,
Phys. Chem. Chem. Phys., 2002, 4, 4678.
E. C. Tuazon, H. Mac Leod, R. Atkinson and W. P. L. Carter,
Environ. Sci. Technol., 1986, 20, 383.
R. Atkinson, S. M. Aschmann, J. Arey and W. P. L. Carter, Int. J.
Chem. Kinet., 1989, 21, 801.
1
214 | Phys. Chem. Chem. Phys., 2006, 8, 1205–1214
This journal is ꢃc the Owner Societies 2006