Atmospheric oxidation of toluene in a large-volume outdoor photoreactor: In situ determination of ring-retaining product yields

Article abstract of DOI:10.1021/jp982719n

Experiments on the photooxidation of toluene/NO_x/air mixtures were performed in the European Photoreactor (EUPHORE), a large-scale outdoor reaction chamber located in Valencia/Spain. The objective of the study was the in situ determination of the yields of ring-retaining products by differential optical absorption spectroscopy (DOAS) and the elucidation of their formation pathways. The experiments were performed with toluene concentrations between 0.68 and 3.85 ppm and initial NO_x concentrations ranging from 3 to 300 ppb, i.e., down to the range actually observed in the lower atmosphere. The ring-retaining product yields were found to be 5.8 ± 0.8%, 12.0 ± 1.4%, 2.7 ± 0.7%, and 3.2 ± 0.6% for benzaldehyde, o-cresol, m-cresol, and p-cresol, respectively. Under the experimental conditions, no dependency of the yields on the NO_x concentration or the toluene/NO_x ratio could be found. The formation kinetics of the cresols are in line with a "prompt" formation mechanism, i.e., abstraction of a hydrogen atom from the toluene-OH adduct (toluene-OH + O₂ → cresols + HO₂). In addition, substantial evidence was found that reaction with NO₃ radicals represents an important sink for cresols in smog chamber studies conducted under conditions of NO_x concentrations above the range observed in the troposphere, possibly also under tropospheric conditions.

Full text of DOI:10.1021/jp982719n

J. Phys. Chem. A 1998, 102, 10289-10299
10289
Atmospheric Oxidation of Toluene in a Large-Volume Outdoor Photoreactor: In Situ
Determination of Ring-Retaining Product Yields
Bjo1rn Klotz, Søren Sørensen, Ian Barnes,* and Karl H. Becker
Bergische UniVersita¨t, Physikalische Chemie - FB 9, Gaussstrasse 20, D-42097 Wuppertal, Germany
Thomas Etzkorn, Rainer Volkamer, and Ulrich Platt
UniVersita¨t Heidelberg, Institut fu¨r Umweltphysik, Im Neuenheimer Feld 366, D-69120 Heidelberg, Germany
Klaus Wirtz and Montserrat Mart´ın-Reviejo
Fundacio´n Centro de Estudios Ambientales del Mediterraneo (CEAM), Parque Tecnolo´gico, Calle 4,
Sector Oeste, E-46980 Paterna (Valencia), Spain
ReceiVed: June 22, 1998; In Final Form: September 22, 1998
Experiments on the photooxidation of toluene/NOx/air mixtures were performed in the European Photoreactor
(EUPHORE), a large-scale outdoor reaction chamber located in Valencia/Spain. The objective of the study
was the in situ determination of the yields of ring-retaining products by differential optical absorption
spectroscopy (DOAS) and the elucidation of their formation pathways. The experiments were performed
with toluene concentrations between 0.68 and 3.85 ppm and initial NOx concentrations ranging from 3 to 300
ppb, i.e., down to the range actually observed in the lower atmosphere. The ring-retaining product yields
were found to be 5.8 ( 0.8%, 12.0 ( 1.4%, 2.7 ( 0.7%, and 3.2 ( 0.6% for benzaldehyde, o-cresol, m-cresol,
and p-cresol, respectively. Under the experimental conditions, no dependency of the yields on the NOx
concentration or the toluene/NOx ratio could be found. The formation kinetics of the cresols are in line with
a “prompt” formation mechanism, i.e., abstraction of a hydrogen atom from the toluene-OH adduct (toluene-
OH + O2 f cresols + HO2). In addition, substantial evidence was found that reaction with NO3 radicals
represents an important sink for cresols in smog chamber studies conducted under conditions of NOx
concentrations above the range observed in the troposphere, possibly also under tropospheric conditions.
Introduction
formation of a methyl-hydroxy-cyclohexadienyl radical,6,7 here-
after termed toluene-OH adduct. Under atmospheric condi-
In urban and industrial areas large amounts of aromatic
hydrocarbons are emitted into the atmosphere, predominantly
from anthropogenic sources such as fossil fuel burning and
solvent use.1 In the presence of NOx (NO + NO2), the
degradation of these compounds in the troposphere ultimately
leads to the formation of ozone and other photooxidants in these
areas.2,3 Because of their reactivity and the amounts emitted,
aromatic hydrocarbons are regarded as the most important class
of hydrocarbons with regard to photooxidant formation in urban
air.4
Among the aromatic hydrocarbons, toluene is the most
abundant species found in the urban atmosphere, making it the
most important single compound with regard to photooxidant
formation.4 Predictions such as this, however, are based on
model calculations which rely heavily on the correctness of the
input data, i.e., the chemical mechanisms describing the
degradation of the various hydrocarbons and emission inven-
tories. The chemical degradation of toluene and other aromatic
hydrocarbons is known to proceed almost exclusively via
reaction with OH radicals.3,5 For toluene, two pathways are
operative. Abstraction of a hydrogen atom from the methyl
group accounts for less than 10% of the total, with the remainder
being addition of OH to the aromatic ring3,5 resulting in the
tions, the toluene-OH adduct will react predominantly with
molecular oxygen.8 In this reaction, an important and particu-
larly debated uncertainty exists in the exact yield of hydroxylated
aromatics, i.e., cresols in the case of toluene. Various recent
studies have come to significantly different conclusions, ranging
from negligible yields9 to extremely high values,10 with two
studies obtaining values intermediate between these ex-
tremes.11,12 A detailed overview of cresol yields published in
the literature is given below. One drawback with these studies
is the fact that the compounds were not detected in situ but
indirectly by GC-FID or HPLC. Although these indirect
methods are very sensitive, they necessitate the employment of
sampling and sample preparation techniques which may lead
to unintentional errors being made such as the possible formation
or destruction of highly reactive sample compounds such as
cresols. These difficulties may help to explain the large
discrepancies between the various studies. Such errors can be
avoided when absolute measurement techniques such as DOAS
(differential optical absorption spectroscopy)13,14 or FT-IR
(Fourier transform infrared) spectroscopy15 are employed whose
accuracy solely depends on the correctness of the absorption
cross sections used for the evaluation of the spectra. With these
methods, no calibration of apparatus is needed. They both have
been successfully employed in previous studies performed in
the European Photoreactor (EUPHORE).16,17 DOAS is a
* Author to whom correspondence should be addressed. Fax: xx49-202/
439-2505. E-mail: barnes@physchem.uni-wuppertal.de.
10.1021/jp982719n CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/20/1998

10290 J. Phys. Chem. A, Vol. 102, No. 50, 1998
Klotz et al.
particularly sensitive technique for the detection of aromatic
hydrocarbons.18 In previous studies using DOAS for the
detection of aromatic hydrocarbons reliable quantification was
difficult due to absorptions of O2 which significantly interfere
with those of aromatic hydrocarbons. This difficulty has
recently been resolved by Volkamer et al.,19 who measured high-
resolution reference spectra of molecular oxygen. The difficul-
ties with the determination of accurate gas-phase absorption
cross sections, especially for the highly reactive and nonvolatile
cresols, were also recently overcome by calibrations performed
at low concentrations in a large-volume cell.20
One problem in some previous smog-chamber studies11,21 was
the influence of the very high NO2 concentrations of several
ppm usually employed. It is known that the toluene-OH adduct
initially formed from the addition of the OH radical to the
aromatic ring rapidly reacts with NO2, compared to a very slow
reaction with O2.8 Under smog-chamber conditions when NO2
concentrations reach ppm levels, reaction with NO2 becomes a
significant removal process for the toluene-OH adduct. Though
product data for this reaction are not available, it is assumed
that high NO2 concentrations lead to higher yields of the cresol
isomers.9,11,22
In this study, the yields of the ring-retaining products
benzaldehyde and o-, m-, and p-cresol have been determined
by an in situ method at NOx concentrations close to those
observed in the troposphere of urban areas. This, in conjunction
with the facts that real sunlight could be used for the photo-
oxidation and that the experiments were conducted in a large-
volume chamber, has resulted in a reliable determination of the
ring-retaining products from the oxidation of toluene which can
be used in chemical models used to describe tropospheric
photooxidation formation.
Figure 1. Example of the chemical mechanism used in the simulations
of the concentration-time profiles of ring-retaining products from
toluene, for explanation see text. TOL, toluene; BALD, benzaldehyde;
OCRE, o-cresol; MCRE, m-cresol; PCRE, p-cresol; PROD, products;
CON, initial concentrations.
coated, while the DOAS mirrors are aluminum coated and have
an additional coating (Allflex-UV, Balzers, Liechtenstein) to
enhance their reflectivity in the UV spectral range. The DOAS
consists of a f/6.9 Czerny-Turner spectrograph (Acton Spectra
Pro 500) with a focal length of 0.5 m. For the experiments
presented here, a 1200 gr mm-1 grating was used, corresponding
to a dispersion of 0.038 nm pixel-1. The spectral resolution,
defined as the fwhm (full width at half-maximum) of an atomic
emission line, was 0.22 nm. For most experiments, the detector
was a photodiode array (Hoffmann, Rauenberg), which was
cooled to -20 °C to reduce the dark current of the diodes. In
the first experiment conducted, see Figure 2, a photodiode array
developed at the University of Heidelberg was used which was
cooled to -30 °C. A detailed description of the DOAS system
used is given in the literature.23 The absorption cross sections
used for the evaluation of the DOAS and FT-IR spectra were
determined in a separate study.20
In addition to these instruments, two J(NO2) filter radiometers
(one measuring direct sunlight, one facing downward to measure
reflected light from the floor panels) and sampling ports for an
ozone monitor (Monitor Labs), a Monitor Labs ML9841A NOx
monitor with a catalytic converter and an Eco Physics CLD
770 AL NOx monitor with a photolytic converter PLC 760 are
installed. The temperature in the chamber was measured with
two thermocouples PT-100, one measuring the floor temperature
and one measuring the temperature of the chamber air. The
ozone, NOx, temperature, and radiation measurement data were
collected and saved by a data acquisition system. A detailed
description of the European Photoreactor EUPHORE can be
found in the literature.23
Experimental Section
All experiments were performed in the European Photoreactor
(EUPHORE), a large-volume outdoor smog chamber integrated
into the Centro de Estudios Ambientales del Mediterraneo
(CEAM) in Valencia/Spain. EUPHORE consists of two half-
spherical FEP (fluorine ethene propene) chambers mounted on
aluminum floor panels covered with FEP foil. This foil is highly
transparent even to short-wavelength sunlight, with transmis-
sions ranging from 85 to 90% in the range 500-320 nm to
>75% at 290 nm. To avoid unwanted heating of the chamber,
the aluminum floor panels are fitted with an active cooling
system designed to keep their temperature constant even during
long-term irradiations in summertime, allowing realistic atmo-
spheric temperature conditions to be maintained. The chamber
used for the experiment described here has a volume of ca. 170
m3, two ventilators with a throughput of 67 m3 min-1 each are
installed to ensure homogeneous reaction mixtures. To supply
the chambers with hydrocarbon and NOy-free dry air, an air-
drying and purification system is installed which is able to
reduce the nonmethane hydrocarbon content to levels below 0.3
µg m-3. When not in use, the chambers are protected by
hydraulically operated steel housings.
In the chamber where the experiments were performed, two
White mirror systems are installed on the aluminum floor panels,
one for the FT-IR and one for the DOAS system. The FT-IR,
a NICOLET Magna 550 with an MCT detector and a resolution
of 1 cm-1, was operated at an optical path length of 326.8 m.
The DOAS, constructed and operated by the group from the
University of Heidelberg, was operated at a short optical path
of 128 m to allow the simultaneous detection of reactants and
products. The mirrors of the FT-IR White system are gold
Toluene was added to the chamber by gently heating the
required amount in a stream of nitrogen entering the chamber,

Oxidation of Toluene in an Outdoor Photoreactor
J. Phys. Chem. A, Vol. 102, No. 50, 1998 10291
Figure 2. Concentration-time profiles of a toluene/NO photooxidation experiment (1050 ppb toluene, 230 ppb NO). (a) Experimental and simulated
concentration-time profiles of toluene (left scale), benzaldehyde, the cresols, and OH (all right scale). Lines denote the simulated profiles, dotted
lines represent the error limits of the simulations, and symbols are experimental data. The long dashed line represents the calculated OH-radical
concentrations. (b) Comparison of the c/t-profiles of benzaldehyde and o-cresol to those of NO (long dashed line), NO2 (dotted line), and ozone
(dashed line). Symbols: (O) toluene, ([) benzaldehyde, (b) o-cresol, (1) m-cresol, (2) p-cresol.
Figure 3. Concentration-time profiles of a toluene/NO photooxidation experiment (730 ppb toluene, 115 ppb NO). Graph (b) also includes the
calculated concentration-time profile of NO3 (solid line), legend otherwise same as for Figure 2.
and NO was injected into the same stream through a septum.
After allowing some time for mixing of the compounds, the
photooxidation was initiated by opening the protective housings
and exposing the chamber to sunlight. Due to small cracks in
the FEP foil and the pressure necessary to keep the chamber
inflated, most experiments had a leak rate of ca. 3% h-1. To
accurately determine this dilution rate, ca. 20 ppb of SF6 (1
ppm ) 2.46 × 1013 cm-3 at 1013 mbar and 298 K) was added
as an inert tracer substance in most experiments. The decay of
SF6 was monitored using its very intense IR absorption band
centered around 950 cm-1. In the experiments where no SF6
was added, the dilution rate was calculated from the loss of
toluene in darkness prior to exposing the chamber to sunlight.
Method of EValuation. The concentration-time profiles of
toluene and its oxidation products were evaluated by computer-
aided simulation of the profiles using a greatly simplified
chemical mechanism. In these simulations, the OH radical
concentration was adjusted to describe the measured toluene
decay in the experiments. Reaction with OH radicals was
assumed to be the sole chemical loss process of toluene and its
ring-retaining oxidation products. In addition, losses through
chamber leaks were taken into account. The limitations of these
assumptions will be discussed in detail below. The chemical
simulation program ChemSimul developed at the Danish
national research facility at Risø was used for the calculations.
As an example for the mechanism used, Figure 1 shows the
relevant parts of the input file for the experiment depicted in
Figure 4. The OH radical concentrations are controlled through
reactions RE15-RE20. The initial formation of OH results
from RE17, after two intermediate steps, RE15 and RE16. After
two additional delaying steps, RE18 and RE19, OH is consumed
by RE20. The rate constants for RE15-RE20 were adjusted
to yield a best fit of the calculated concentration-time profile
of toluene to the experimental data. As an additional loss
process, dilution through chamber leaks was taken into account
by RE10-RE14. Reactions RE6-RE9 represent losses of the

10292 J. Phys. Chem. A, Vol. 102, No. 50, 1998
Klotz et al.
Figure 4. Concentration-time profiles of a toluene/NO photooxidation experiment (725 ppb toluene, 55 ppb NO). Legend same as for Figure 3.
ring-retaining products through reaction with OH radicals, the
rate constants are those recommended by Atkinson.5 The
reactions do not consume OH, as its concentration is controlled
by reactions RE15-RE20.
Reactions RE1-RE4 are the formation channels of the ring-
retaining products observed in the photooxidation of toluene,
RE5 is needed to achieve the total OH reaction rate constant of
higher toluene concentrations of ca. 3.8 ppm, several toluene
absorption bands were outside of the applicability of the
Lambert-Beer law in the DOAS spectra, necessitating an
evaluation at less intense bands, which resulted in a higher error
in the concentrations determined. Therefore, and because only
a minor ozone interference was observed in the FT-IR spectra
at these high toluene concentrations, the FT-IR data were used
for toluene in these experiments.
To determine the influence of nitrate radicals on the observed
concentration-time profiles, estimates of their concentrations
in the various runs were made. For these estimations, a steady
state was assumed for NO3, i.e., d[NO3]/dt ) 0. From eq 1
below it can be calculated that the formation and degradation
rates of NO3 are significantly higher than their quotient, making
the assumption of a steady state valid. The following reactions
were taken into account:
k
OH(toluene) ) 5.96 × 10-12 cm3 s-1 recommended by
Atkinson.5 As is evident from RE1-RE4, the formation of
benzaldehyde and the cresols was assumed to result directly
from the reaction of OH with toluene, i.e., not through further
reaction of a stable intermediate. The delay resulting from the
initial formation of a toluene-OH adduct and its subsequent
reaction with O2 is negligible at atmospheric partial pressures
of O2.
To determine the individual formation yields of the ring-
retaining products, the rate constants of their formation channels
were adjusted so the simulated concentration-time profiles of
the products showed the best possible agreement with the
experimental data. Formation yields are then given by the ratios
of the individual bimolecular rate constants to the total rate
constant of the reaction of toluene with OH radicals. In many
experiments, the simplified model was unable to describe the
experimental data toward the end of the experiment due to an
apparent additional loss of the cresols through reaction with
NO3 radicals (see below). In such cases, only the initial
formation of the cresols was fitted, before the additional loss
process became important.
NO2 + O3 f NO3 + O2
NO3 + NO f 2 NO2
(a)
(b)
(c)
NO3 + hν f NO + O2
f NO2 + O
NO3 + o-cresol f products
NO3 + m-cresol f products
NO3 + p-cresol f products
(d)
(e)
(f)
The error limits given for the product yields calculated in
this way represent the range in which the yields could be
changed without the simulated concentration-time profiles
significantly deviating from the error limits of the experimental
values.
Since the total decomposition rate of NO3 is only insignifi-
cantly different from its formation rate, the NO3 concentration
can be calculated by eq 1:
In most experiments, the DOAS was used for the determi-
nation of both the concentration-time profiles of toluene and
its oxidation products. The FT-IR spectrometer available in
addition to the DOAS was used to verify the concentration-
time profiles of toluene determined by DOAS. In all experi-
ments, both instruments showed identical values within their
experimental uncertainties. In the experiments conducted with
starting concentrations of toluene around 700 ppb, the toluene
concentrations obtained by DOAS were used as ozone absorp-
tions significantly interfered with the FT-IR analysis at these
concentrations. In the two experiments conducted at much
[NO3] )
ka[O3][NO2]
kb[NO] + kc + kd[o-cresol] + ke[m-cresol] + kf[p-cresol]
(1)
The equilibrium NO3 + NO2 h N2O5 need not be considered
in this calculation, since under the steady-state assumption the
rates of the forward and reverse reactions are equal and cancel
each other out. The photolysis frequencies of NO3 radicals

Oxidation of Toluene in an Outdoor Photoreactor
J. Phys. Chem. A, Vol. 102, No. 50, 1998 10293
Figure 5. Concentration-time profiles of a toluene/NO photooxidation experiment (690 ppb toluene, 24 ppb NO). Legend same as for Figure 3.
Figure 6. Concentration-time profiles of a toluene/NO photooxidation experiment (3850 ppb toluene, 120 ppb NO). Legend same as for Figure
3.
(reaction c) could not be measured directly in the EUPHORE
chamber, therefore an approximation was used for their deter-
mination. They were calculated relative to the photolysis
frequencies of NO2, which were measured with a filter
radiometer. A conversion factor was estimated based on the
highest NO2 photolysis frequency measured in the experiments
performed, which was at 11:40 h in the experiment displayed
in Figure 6. For that time and day, a corresponding NO3
photolysis frequency was calculated based on the data of
Orlando et al.24 Additional correction factors were incorporated
to account for absorption of sunlight by the FEP foil and for
the albedo of the aluminum floor panels of the chamber.
Another requirement of this approach is that the scattering of
light in the region around 630 nm, where NO3 radicals
photolyze, must be very similar to the scattering of light around
400 nm, where NO2 absorbs, which is not necessarily true in
all cases. The errors introduced into the calculated NO3
concentrations by these uncertainties are, however, very small,
as the major degradation pathway of the NO3 radicals is the
reaction with NO, not photolysis. Typically, the reaction with
NO was found to be 10-50 times faster than photolysis.
The NO3 concentrations calculated in this way may, however,
only be regarded as upper limits. The true NO3 concentration
will be lower, as some sinks could not be included in the
calculation. These are notably reaction with peroxyl radicals
and other (nonradical) intermediates in the photooxidation of
aromatic hydrocarbons.
Results and Discussion
A total of nine experiments were performed. The initial
conditions of these experiments are listed in Table 1, together
with the formation yields of the ring-retaining products ben-
zaldehyde and o-, m-, and p-cresol obtained from simulations
as described above. The individual experiments are not ordered
chronologically, but by their initial toluene/NO ratios, with
experiments where NOx concentrations were kept constant added
at the end.
The averaged yields from Table 1 are 5.8 ( 0.8% for
benzaldehyde and 12.0 ( 1.4%, 2.7 ( 0.7%, and 3.2 ( 0.6%
for o-, m-, and p-cresol, respectively. Within their experimental
uncertainties, the yields are dependent neither on the NOx
concentration nor on the initial toluene/NO ratio. This behavior
might have been expected as the NO2 concentration is suf-

10294 J. Phys. Chem. A, Vol. 102, No. 50, 1998
Klotz et al.
TABLE 1: Overview of the Initial Conditions and Results of the Experiments Presented in This Study
initial conditions
product yields
experiment
in figure
c0(toluene)
[ppb]
c0(NO)
[ppb]
benzaldehyde
[%]
o-cresol
[%]
m-cresol
[%]
p-cresol
[%]
all cresols
[%]
2
3
4
5
6
7
8
9
10
1050
730
725
230
115
55
24
120
28
5.9 ( 1.0
6.6 ( 1.0
6.9 ( 1.0
5.4 ( 1.0
5.6 ( 0.5
4.9 ( 0.7
4.7 ( 1.0
5.7 ( 1.0
6.9 ( 1.0
5.8 ( 0.8
10.7 ( 2.0
10.6 ( 2.5
11.7 ( 2.0
13.9 ( 2.0
12.1 ( 2.0
10.2 ( 1.5
13.4 ( 2.0
13.4 ( 2.0
14.8 ( 2.0
12.0 ( 1.4
2.7 ( 1.5
1.6 ( 1.5
2.5 ( 1.5
3.2 ( 1.5
2.5 ( 1.5
2.2 ( 1.5
4.0 ( 1.5
2.7 ( 1.5
2.3 ( 1.5
2.7 ( 0.7
2.9 ( 1.0
2.3 ( 1.0
2.9 ( 1.0
3.9 ( 1.0
3.4 ( 1.0
2.8 ( 1.0
3.5 ( 1.0
3.9 ( 1.0
3.5 ( 1.0
3.2 ( 0.6
16.3 ( 2.7
14.5 ( 3.1
17.1 ( 2.7
21.0 ( 2.7
18.0 ( 2.7
15.2 ( 2.3
20.9 ( 2.7
20.0 ( 2.7
20.6 ( 2.7
17.9 ( 2.5
690
3850
3800
680
725
710
3-3.5 consta
8.5-12 consta
300 constb
average yields
a In these experiments, the total concentration of NOx was kept at constant levels of 3-3.5 and 8.5-12 ppb, respectively. b In this experiment,
a constant NO concentration of ca. 300 ppb was maintained. The cresol yields in this experiment might have been influenced by the high NO2
concentrations which accumulated; they were therefore not included in the calculation of the average yields.
TABLE 2: Comparison of the Formation Yields of o-Cresol and Benzaldehyde Obtained in This Study to Literature Values
yields
literature
experimental conditions, methods
NOx, solar simulator, GC/FID
CH3ONO/NO, solar simulator, GC/FID
CH3ONO/NO, blacklamps, HPLC
NOx, blacklamps, FTIR
benzaldehyde [%]
o-cresol [%]
21
13.1 ( 7.2
Atkinson et al. 1980 (25)
Atkinson et al. 1983 (26)
Shepson et al. 1984 (27)
Bandow et al. 1985 (28)
Gery et al. 1985 (29)
12
7.3 ( 2.2
5.4
11 ( 1
10.4 ( 2.9
7.1
HONO, UV fluorescent lamps, GC/FID
NOx, sunlight, GC/FID
22
16
Leone et al. 1985 (30)
Atkinson et al. 1989 (22)
Seuwen and Warneck 1996 (10)
NOx, blacklamps, GC/FID
2 + 10 l cells, H2O2/UV lamps, 1000 ppm H2O2 and toluene,
GC/FID and HPLC
6.5 ( 0.8
5.3 ( 1.1
20.4 ( 2.7
38.5 ( 8.3
Bierbach et al. 1994 (9)
400 + 3 l cells, H2O2/UV lamps, 1000 ppm H2O2 and toluene,
GC/FID
1080 l reactor, H2O2/UV lamps, GC/FID and FTIR
alkenes + O3, GC/FID
CH3ONO/NO, 1.4 ppm NO2, blacklamps, GC/FID
CH3ONO/NO, 7 ppm NO2, blacklamps, GC/FID
CH3ONO/NO, < 1 ppm NOx, fluorescent bulbs, GC/FID
toluene/NO, sunlight, DOAS/FTIR
7.0 ( 1.5
7.0 ( 1.5
up to 23
not quantifiable
12.3 ( 2.2
14.5 ( 0.7
16.0 ( 0.8
12.3 ( 1.2
12.0 ( 1.4
Atkinson and Aschmann 1994 (11)
Smith et al. 1998 (12)
this work
6.0 ( 0.6
5.8 ( 0.8
ficiently low in all the experiments for its reaction with the
toluene-OH adduct formed from the initial attack of OH to
toluene to be negligible, with the possible exception of the last
experiment, see below.
This conjecture is supported by a comprehensive study of
Bierbach et al.,9 in which the yield was investigated over a wide
range of experimental conditions.
For the formation yields of o-cresol, a different picture
emerges. In the eighties, most studies reported yields of around
20% for this compound, albeit with some variations, see for
example Atkinson et al.25 and Atkinson et al.26 The extremely
high value of 38.5 ( 8.3% recently published by Seuwen and
Warneck10 can be attributed to the experimental conditions of
that study, under which a significant influence of heterogeneous
reactions is expected. Bierbach et al.9 obtained similar results
under analogous conditions but could show that the o-cresol
yield is reduced dramatically when the starting concentrations
are reduced. Therefore, it seems reasonable to conclude that
the results of Seuwen and Warneck10 are not representative of
the troposphere. Bierbach et al.9 estimated an upper limit of
3% for the yield of o-cresol under atmospheric conditions, which
is in stark contrast to the value of 12.0 ( 1.4% obtained in the
present study. The discrepancy is almost certainly due to the
sampling procedure employed by Bierbach et al.,9 which
consisted of taking gas-phase samples with a glass syringe and
injecting them directly into a GC-FID. This procedure probably
leads to losses of polar and nonvolatile compounds such as
o-cresol; these losses could not be quantified. In addition,
photolysis of o-cresol might have been an additional loss process
in their study.
Some of the experiments listed in Table 1 have been
independently evaluated by Mart´ın-Reviejo et al.31 using a
different method. The yields obtained by Mart´ın-Reviejo et al.
are generally in very good agreement with those obtained in
the present study. Since the method of Mart´ın-Reviejo et al.
does not take into account losses of reactants and products
through dilution of the chamber air, it could only be employed
reliably in experiments 8-10. In those experiments, losses of
NOx were compensated by the addition of NO, leading to high
OH radical concentrations throughout the experiments (see
below). Under these conditions, losses through dilution are
small compared to those through OH radical reaction.
A comparison of the formation yields of ring-retaining
products obtained in the present study to literature values is
given in Table 2. In this table, only the data for o-cresol and
benzaldehyde are listed, as the database for m- and p-cresol is
very limited. For benzaldehyde, all reported yields are around
7%, with the exception of a few studies from the early
eighties.25,28,29 The result obtained in this work, 5.8 ( 0.8%,
is somewhat below this average; however, the values can be
regarded as equal within their experimental errors. A very
recent study by Smith et al.12 resulted in a benzaldehyde yield
of 6.0 ( 0.6%, almost identical to our value. From the data
presented in Table 1, it appears that the benzaldehyde yield is
relatively constant regardless of the experimental conditions.
The results of the present study are in excellent agreement
with a recent study of Atkinson and Aschmann11 who found
o-cresol yields of 12.3 ( 2.2% under NOx free conditions and

Oxidation of Toluene in an Outdoor Photoreactor
J. Phys. Chem. A, Vol. 102, No. 50, 1998 10295
Figure 7. Concentration-time profiles of a toluene/NO photooxidation experiment (3800 ppb toluene, 28 ppb NO). Legend same as for Figure 3.
Figure 8. Concentration-time profiles of a toluene/NO photooxidation experiment (680 ppb toluene, 3-3.5 ppb NOx held constant). Legend same
as for Figure 3.
to the results of Smith et al.,12 who found 12.3 ( 1.2% for
o-cresol at sub-ppm levels of NOx. Atkinson and Aschmann11
found that the o-cresol yield rose by several percent in the
presence of ppm levels of NOx. In the present study, only a
single experiment (Figure 10) was conducted at NOx concentra-
tions approaching those employed by Atkinson and Aschmann,11
this experiment gave the highest measured yield of o-cresol with
14.8 ( 2.0%. At an NO2 concentration of 1.4 ppm, Atkinson
and Aschmann found an o-cresol yield of 14.5 ( 0.7%, which
is again in excellent agreement with our value.
For m- and p-cresol, which could only be measured as a sum
by the chromatographic methods previously employed, yields
of 5%29 and 4.8 ( 0.9%22 are reported in the literature. This
is in reasonable agreement with the value of 5.9 ( 0.9% which
can be calculated from the results presented in Table 1. The
only study in which m- and p-cresol were be separated is that
of Smith et al.12 They obtained yields of 2.6 ( 0.3% for
m-cresol and 3.0 ( 0.3% for p-cresol, in good agreement with
our data.
as (a), left-hand side of each figure, show experimental and
simulated concentration-time profiles of toluene, benzaldehyde,
and the cresol isomers as well as the OH radical concentrations
calculated from the decay of toluene. Error limits of the
simulations are presented only for benzaldehyde and o-cresol
and not for m- and p-cresol since these are large and would
only complicate the graphs. The plots marked as (b), right-
hand side of each figure, show a comparison of the experimental
concentration-time profiles of o-cresol and benzaldehyde to
those of ozone, NO, NO2, and the estimated NO3 concentration.
The individual experiments will be discussed in the following
paragraphs.
In the experiment displayed in Figure 2, a mixture of 1050
ppb toluene and 230 ppb NO was photolyzed. With 4.6, the
initial toluene/NO ratio is the lowest of all experiments
performed. A simulation as described above resulted in the
yields given in Table 1. The concentration-time profile of
benzaldehyde is well described over the entire range of Figure
2a. In contrast, the observed concentrations of the cresols
become significantly lower than predicted by the simulation
toward longer reaction times. This indicates the presence of
an additional chemical loss process for the cresols at this time,
The concentration-time profiles obtained in the individual
experiments conducted in this study are plotted in Figures 2-10.
Each figure consists of two separate plots. The plots marked

10296 J. Phys. Chem. A, Vol. 102, No. 50, 1998
Klotz et al.
Figure 9. Concentration-time profiles of a toluene/NO photooxidation experiment (725 ppb toluene, 8.5-12 ppb NOx held constant). Legend
same as for Figure 3.
Figure 10. Concentration-time profiles of a toluene/NO photooxidation experiment (710 ppb toluene, 300 ppb NO held constant). Legend same
as for Figure 3.
which does not seem to affect benzaldehyde. From Figure 2b
it is evident that the losses of cresols occur at the time when
NO2 as well as ozone have accumulated in high concentrations.
Under these conditions, NO3 radicals can be formed. NO3
radicals are known to react rapidly with cresols,5 while their
reaction with benzaldehyde is almost 4 orders of magnitude
slower.32 The assumption that NO3 radicals are responsible for
the observed losses of the cresols is therefore consistent with
the observations. In contrast to the other experiments per-
formed, no estimation of the NO3 radical concentration could
be made here, as only the NOx analyzer with a catalytic
converter was available. The catalytic converter converts not
only NO2 but also several NOy, notably acetyl peroxynitrate
(PAN) and analogous compounds, to NO and measures them
as NO2. An exact knowledge of the NO2 concentration is,
however, a prerequisite for the calculation of NO3 concentrations
(see eq 1).
1. As expected, the concentration-time profiles of the products
show a behavior qualitatively similar to that observed in Figure
2, only the OH radical concentration drops markedly toward
the end of the experiment, probably due to the considerably
longer reaction time in this experiment. Here, too, the model
can only describe the cresol concentrations adequately at the
beginning of the experiment, while benzaldehyde is reproduced
well even toward the end. Again, the loss of the cresols occurs
when conditions exist for the production of NO3 radicals, namely
absence of NO and simultaneous high concentrations of NO2
and ozone. The NO3 concentration calculated according to eq
1 starts to become significant at the time the cresol losses start,
see Figure 3b. At the end of the experiment, the estimated NO3
concentration has reached a maximum of almost 5 ppt.
In Figure 4, an experiment with starting concentrations of
725 ppb toluene and 55 ppb NO is shown, the initial toluene/
NO ratio has been approximately doubled, it is about 13 here.
It is evident in Figure 4a that the concentration-time behavior
of benzaldehyde is reproduced well by the simplified chemical
mechanism used in the simulation, while for the cresols, a “dip”
is observed in their concentrations which cannot be explained
The experiment shown in Figure 3 was conducted at a slightly
higher initial toluene/NO ratio of 6.3, but the starting concentra-
tions were lower, with 730 ppb toluene and 115 ppb NO. The
yields calculated from the simulation are again given in Table

Oxidation of Toluene in an Outdoor Photoreactor
J. Phys. Chem. A, Vol. 102, No. 50, 1998 10297
by OH reaction alone. In contrast to the previous experiments,
the cresol concentrations do not drop to near zero. This effect
is best visible for o-cresol, the isomer formed in highest yields.
Its concentration stabilizes at ca. 5 ppb. As can be seen in
Figure 4b, the maximum decay rate of o-cresol corresponds to
the maximum of the calculated NO3 radical concentration. The
maximum concentration of NO3 is about 1 order of magnitude
lower than in the previous experiment. The rapid drop in the
NO3 concentration after its maximum is due to the limited
availability of NO2 and also its rapid reaction with the cresols,
the latter reactions are dominating eq 1 at this point. The almost
quantitative loss of NOx at the end of the experiment is also
the probable reason for the drop in the OH radical concentration
toward the end of the experiment (see Figure 4a).
concentration-time profiles of benzaldehyde and the cresols
well, no “dip” is evident in the cresol profiles. In accordance
with this observation, a very low maximum NO3 concentration
of about 0.12 ppt is calculated, see Figure 7b. As in the
preceding experiments, the concentration-time profile of OH
exhibits a pronounced maximum and drops parallel to the NOx
concentration.
In the experiment shown in Figure 8, the sum of the
concentrations of NO and NO2 was maintained at a constant
level of 3-3.5 ppb. This was achieved by an automated system
which added more NO when the NOx concentration reached a
given minimum value. Initially, 680 ppb toluene and ap-
proximately 2.3 ppb of NO were added. The compensation for
losses of NOx has the important consequence that the OH radical
concentration remained at a constantly high level throughout
the experiment, with no distinctive maximum observed (see
Figure 8a). The simulation described the concentration-time
behavior of benzaldehyde and the cresols very well. This could
be expected as the NOx concentrations are even lower here than
in the experiment displayed in Figure 5, the amount of NO2
present is insufficient to give rise to a significant formation of
NO3, as evidenced by the very low maximum NO3 concentration
of about 0.14 ppt reached toward the end of the experiment.
Figure 5 shows an experiment with a toluene concentration
of 690 ppb, the initial NO concentration was 24 ppb, resulting
in an initial toluene/NO ratio of 29, twice as high as in the
previous experiment. In this experiment, the concentration-
time behavior of both benzaldehyde and the cresols is described
well by the simulation, no unexplained loss of cresols is
apparent. Such an effect would be expected to occur at the
time the NO3 concentration reaches its maximum. The maxi-
mum concentration is, however, more than a factor of 2 lower
than in the previous experiment. This is apparently too low to
have an observable influence on the concentration-time profile
of the cresols.
The experiment shown in Figure 9 is of the same type as
that shown in Figure 8, but here a somewhat higher NOx
concentration of between 8.5 and 12 ppb was kept constant.
As before, this leads to the OH radical concentration remaining
relatively constant at around 107 cm-3 throughout the experiment
(see Figure 9a). In contrast to the previous experiment, the
simulation is only able to describe the concentration-time
profiles of the cresols until shortly after their maxima. After
this time, the decay of the cresols is significantly faster than
predicted by the simplified mechanism. This discrepancy occurs
at the same time when the calculated NO3 radical concentration
sharply rises (see Figure 9b). Toward the end of the experiment,
it reaches a maximum of almost 0.9 ppt, and is thus at a level
high enough to have an influence on the cresols.
The next two experiments, shown in Figures 6 and 7, have
been conducted at significantly higher initial toluene concentra-
tions. This was done mainly to minimize losses of the cresols
due to OH radical reaction. In this way, the unknown additional
loss process should be more easily distinguishable from the OH
reaction.
The first experiment of this type, shown in Figure 6, was
performed with initial concentrations of 3850 ppb for toluene
and 120 ppb for NO, resulting in an initial toluene/NO ratio of
32. The concentration-time profile of benzaldehyde is de-
scribed well throughout the experiment, while a “dip” is evident
in the profile of the cresols which is not described by the
simulation. In Figure 6b, a correlation of these cresol losses
with the maximum of the calculated NO3 concentration can be
seen. The calculated maximum NO3 concentration is about 1
ppt in this experiment, and thus significantly higher than in the
experiment shown in Figure 5. The higher calculated NO3
concentrations can be attributed to the higher NOx concentrations
employed. In contrast to the experiment presented in Figure 4,
the cresol concentration not only stabilizes after the NO3 has
been consumed, but starts to rise again. This is probably due
to the higher concentration of toluene, which acts as an OH
radical scavenger. The concentration-time profile of OH
exhibits an even sharper maximum as before, probably due to
the almost complete loss of NOx halfway through the experi-
ment. At this time, still existing losses of NOx appear to be
compensated by decomposition of acetyl peroxynitrate (PAN)
or analogous compounds. A low but constant NOx concentration
of a few ppb is maintained, leading to a low but constant OH
radical concentration of ca. 2 × 105 cm-3. It should be noted
that the apparent odd behavior of NOx and consequently NO3
between 11:30 and 11:45 h is an artifact of maintenance
operations at the NOx analyzer, which was disconnected from
the chamber during that period.
The experiment displayed in Figure 10 was conducted under
conditions aimed at suppressing the formation of NO3 radicals
as much as possible while at the same time having very high
concentrations of NOx. Therefore, in contrast to the experiments
presented in Figures 8 and 9, where the total concentration of
NOx (NO + NO2) was kept constant, in this experiment the
NO concentration was maintained at a constant level of 300
ppb. This extremely high NO concentration scavenged most
of the ozone and virtually any NO3 formed.
The concentration-time profile of OH showed a broad
maximum, followed by a slow decay (see Figure 10a). This
behavior can probably be attributed to increasing scavenging
of OH through reaction with NO2 toward the end of the
experiment. As expected, only very low NO3 concentrations
were calculated under these conditions, the maximum is about
0.02 ppt at the end of the experiment. Consequently, the
concentration-time behavior of both benzaldehyde and the
cresols is described very well by the simplified simulation (see
Figure 10a).
Conclusions and Implications for the Atmosphere. Ring-
retaining products are formed in the photooxidation of toluene
in significant yields, with o-cresol being the most important
single compound. The individual yields determined were 5.8
( 0.8% for benzaldehyde, 12.0 ( 1.4% for o-cresol, 2.7 ( 0.7%
for m-cresol, and 3.2 ( 0.6% for p-cresol. The total yield of
the cresol isomers is 17.9 ( 2.5%, thus they comprise a
significant fraction of the overall carbon balance.
Figure 7 shows an experiment conducted at an extremely high
initial toluene/NO ratio of 136, 3800 ppb toluene and 28 ppb
NO were used. As in the low-NOx experiment presented in
Figure 5, the simplified simulation is able to describe the

10298 J. Phys. Chem. A, Vol. 102, No. 50, 1998
Klotz et al.
The formation of cresols in significant yields has implications
for the photooxidant formation from aromatic hydrocarbons, as
the cresols have a comparably low reactivity with regard to
oxidant formation.33 This is in contrast to ring-fragmentation
products, some of which represent radical sources which leads
to increased photooxidant formation. The formation of signifi-
cant amounts of cresols is therefore expected to lessen the
photooxidant formation from aromatic hydrocarbons.
molecular oxygen. Stockwell et al. incorporated these pathways
into their Regional Atmospheric Chemistry Model (RACM) in
an attempt to reconcile the data from Bierbach et al.9 with those
of Atkinson and Aschmann.11
This proposed mechanism is not supported by the results of
the present study, as a rapid cresol formation was observed in
the early phases of the experiment, when the ozone concentration
was very low. Reaction of the toluene-OH adduct with ozone
can therefore not be the main source of cresols in the
experiments presented in this paper. As mentioned above, the
data suggest the direct formation of cresols from the reaction
of the toluene-OH adduct with molecular oxygen, with no
major contribution from other reactions.
Another important result of this study is the observed loss of
the cresols toward the end of many experiments, which could
not be explained by OH reaction alone. No such losses of
benzaldehyde occurred. This and the good correlation of the
losses to calculated concentration-time profiles and maximum
concentrations of NO3 radicals leads to the conclusion that
reaction of the cresols with NO3 radicals is most probably
responsible for this effect. The reaction of the cresols with NO3
is known to be considerably faster than that of benzaldehyde.5,32
In their kinetic study of the NO3 radical reactions with cresols,
Carter et al.34 had already suggested that this reaction can be
an important atmospheric sink for cresols. The question of
whether this is the case even under daylight conditions is,
however, difficult to answer. For o-cresol, the OH reaction rate
constant is about 3 times higher than the NO3 reaction rate
constant,5,32 and it can be calculated that if the OH concentration
is in the range 106-107 cm-3 or 0.04-0.4 ppt under typical
photosmog conditions, NO3 radical concentrations of 0.12-1.2
ppt are required for the two reactions to be competitive.
Atmospheric daytime NO3 radical concentrations are currently
not known, the detection limit of even the most sensitive
measurement technique currently employed is around 1 ppt at
night.35 Daytime NO3 concentrations above the detection limit
have not been observed, and estimates give upper limits lower
than 1 ppt.14 They might therefore very well be in the range
where reaction of NO3 can be a nonnegligible daytime sink for
the cresols and other fast-reacting organic compounds. This
may be the case in polluted situations with dense cloud cover
during which photolysis rates of O3, NO2, and NO3 will be low
thus allowing higher steady-state concentrations of NO3 radicals.
Therefore, further investigations seem highly desirable to
elucidate daytime NO3 radical concentrations. This might be
accomplished by simultaneously monitoring humidity, NOx,
ozone, RO2 radicals, the spectral light distribution and intensity,
and hydrocarbons which might constitute additional sinks of
NO3, downwind of a large city. With these data, daytime NO3
concentrations may be estimated using the pseudo steady-state
approximation for NO3.
In an earlier paper by Klotz et al.,36 the possible formation
of phenol through photolysis of an intermediate, benzene oxide/
oxepin, was proposed for the case of benzene. Such a pathway
cannot be operative in the case of toluene investigated here, as
the formation yield of cresols in the photolysis of the arene
oxide expected to be formed from toluene, toluene-1,2-oxide/
2-methyloxepin, was found to be negligible.17
Apart from their formation in the photooxidation of aromatic
hydrocarbons, emissions of phenol and cresols have recently
been observed in the tailpipe exhaust gases from automobiles
in tunnel measurements using gas chromatography37 and the
DOAS technique.38 Therefore, to elucidate the total atmospheric
burden of cresols, it will be necessary to measure such direct
emissions under realistic conditions, e.g., in a traffic tunnel. To
evaluate the formation of cresols from direct emissions and/or
as products of the photooxidation of toluene, measurements of
cresol/toluene ratios in a downtown area are also desirable.
Differential optical absorption spectroscopy (DOAS) is an ideal
tool for such measurements.
Acknowledgment. The authors are grateful to the technical
staff at CEAM/Valencia for their assistance with the measure-
ments. The authors also thank Dr. Palle Pagsberg for supplying
the chemical simulation program ChemSimul developed at the
Danish national research center at Risø. Financial support by
the “Bundesministerium fu¨r Bildung, Wissenschaft, Forschung
und Technologie” (BMBF), the “Comisio´n Interministerial de
Ciencia y Tecnolog´ıa” (CICYT), the “Generalitat Valenciana”,
and the European Commission for this work is gratefully
acknowledged.
References and Notes
(1) Piccot, S. D.; Watson, J. J.; Jones, J. W. J. Geophys. Res. 1992,
97, 9897.
(2) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics:
From Air Pollution to Climate Change; John Wiley & Sons Inc.: New York,
1998.
(3) Transport and Chemical Transformation of Pollutants in the
Troposphere, Vol. 3: Chemical Processes in Atmospheric Oxidation; Le
Bras, G., Ed.; Springer: Berlin, Germany, 1997.
In the simulations of the experimental concentration-time
profiles of benzaldehyde and the cresols obtained in this study,
a “prompt” formation mechanism was assumed for these
products. As evident in Figures 2-10, this mechanism was
well suited to describe their formation kinetics. Attempts to
reproduce the concentration-time behavior with a delayed
mechanism, e.g., through further reaction of a stable intermedi-
ate, proved unsuccessful. It can therefore be concluded that
the cresols are formed directly from the reaction of O2 with the
toluene-OH adduct, i.e.
(4) Derwent, R. G.; Jenkin, M. E.; Saunders, S. M. Atmos. EnViron.
1996, 30, 181.
(5) Atkinson, R. J. Phys. Chem. Ref. Data, Monograph No. 2, 1994.
(6) Fritz, B.; Handwerk, V.; Preidel, M.; Zellner, R. Ber. Bunsen-Ges.
Phys. Chem. 1985, 89, 343.
(7) Markert, F.; Pagsberg, P. Chem. Phys. Lett. 1993, 209, 445.
(8) Knispel, R.; Koch, R.; Siese, M.; Zetzsch, C. Ber. Bunsen-Ges.
Phys. Chem. 1990, 94, 1375.
toluene-OH + O2 f cresol + HO2
(9) Bierbach, A.; Barnes, I.; Becker, K. H.; Klotz, B.; Wiesen, E. In
Sixth European Symposium Physico-Chemical BehaViour of Atmospheric
Pollutants, Angeletti, G., Restelli, G., Eds.; EC Air Pollution Research
Report 50, EUR 15609/1, Brussels, Belgium, 1994; Vol. 1, pp 129-136.
(10) Seuwen, R.; Warneck, P. Int. J. Chem. Kinet. 1996, 28, 315.
(11) Atkinson, R.; Aschmann, S. M. Int. J. Chem. Kinet. 1994, 26, 929.
(12) Smith, D. F.; McIver, C. D.; Kleindienst, T. E. J. Atmos. Chem.
1998, 30, 209.
with no stable intermediate involved.
A different formation mechanism for the cresols was recently
proposed by Stockwell et al.35 They suggested the formation
of cresols from the reaction of ozone with the aromatic-OH
adduct, while assuming a low yield of 2% for the cresol
formation from the reaction of aromatic-OH adducts with
(13) Platt, U.; Perner, D.; Pa¨tz, H. W. J. Geophys. Res. 1979, 84, 6329.

Oxidation of Toluene in an Outdoor Photoreactor
J. Phys. Chem. A, Vol. 102, No. 50, 1998 10299
(14) Heintz, F.; Platt, U.; Flentje, H.; Dubois, R. J. Geophys. Res. 1996,
101, 22891.
(28) Bandow, H.; Washida, N.; Akimoto, H. Bull. Chem. Soc. Jpn. 1985,
58, 2531.
(15) Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Atmospheric Chemistry:
Fundamentals and Experimental Techniques; John Wiley & Sons: New
York, 1986; Chapter 5-C.
(16) Wa¨ngberg, I.; Etzkorn, T.; Barnes, I.; Platt, U.; Becker, K. H. J.
Phys. Chem. A 1997, 101, 9694.
(17) Klotz, B.; Barnes, I.; Becker, K. H. Chem. Phys. 1998, 132, 289.
(18) Trost, B.; Stutz, J.; Platt, U. Atmos. EnViron. 1997, 31, 3999.
(19) Volkamer, R.; Etzkorn, T.; Geyer, A.; Platt, U. Atmos. EnViron.
1998, 32, 3731.
(20) Etzkorn, T.; Klotz, B.; Sørensen, S.; Patroescu, I. V.; Barnes, I.;
Becker, K. H.; Platt, U. Atmos. EnViron., in press.
(21) Atkinson, R.; Aschmann, S. M.; Arey, J. Int. J. Chem. Kinet. 1991,
23, 77.
(29) Gery, M. W.; Fox, D. L.; Jeffries, H. E.; Stockburger, L.; Weathers,
W. S. Int. J. Chem. Kinet. 1985, 17, 931.
(30) Leone, J. A.; Flagan, R. C.; Grosjean, D.; Seinfeld, J. H. Int. J.
Chem. Kinet. 1985, 17, 177.
(31) Mart´ın-Reviejo, M.; Pons, M.; Wirtz, K.; Etzkorn, T.; Senzig, J.
In The Oxidizing Capacity of the Troposphere, EC Air Pollution Research
Report 60, EUR 17482 EN, Larsen, B., Versino, B., Angeletti, G., Eds.;
Luxemburg, 1997; pp 350-354.
(32) Wayne, R. P.; Barnes, I.; Biggs, P.; Burrows, J. P.; Canosa-Mas,
C. E.; Hjorth, J.; Le Bras, G.; Moortgat, G. K.; Perner, D.; Poulet, G.;
Restelli, G.; Sidebottom, H. Atmos. EnViron. 1991, 25A, 1.
(33) Carter, W. P. L. J. Air Waste Manage. Assoc. 1994, 44, 881.
(22) Atkinson, R.; Aschmann, S. M.; Arey, J.; Carter, W. P. L. Int. J.
Chem. Kinet. 1989, 21, 801.
(23) The European Photoreactor EUPHORE; Final Report of the EC-
Project, Contract EV5V-CT92-0059; Becker, K. H., Ed.; Wuppertal,
Germany, 1996.
(24) Orlando, J. J.; Tyndall, G. S.; Moortgat, G. K.; Calvert, J. G. J.
Phys. Chem. 1993, 97, 10996.
(25) Atkinson, R.; Carter, W. P. L.; Darnall, K. R.; Winer, A. M.; Pitts,
J. N., Jr. Int. J. Chem. Kinet. 1980, 12, 779.
(34) Carter, W. P. L.; Winer, A. M.; Pitts, J. N., Jr. EnViron. Sci. Technol.
1981, 15, 829.
(35) Stockwell, W. R.; Kirchner, F.; Kuhn, M.; Seefeld, S. J. Geophys.
Res. 1997, 102, 25847.
(36) Klotz, B.; Barnes, I.; Becker, K. H.; Golding, B. T. J. Chem. Soc.,
Faraday Trans. 1997, 93, 1507.
(37) Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. EnViron. Sci. Technol.
1998, 32, 2051.
(26) Atkinson, R.; Carter, W. P. L.; Winer, A. M. J. Phys. Chem. 1983,
87, 1605.
(27) Shepson, P. B.; Edney, E. O.; Corse, E. W. J. Phys. Chem. 1984,
88, 4122.
(38) Platt, U.; Kurtenbach, R. In BMBF-VerbundVorhaben: Tropospha¨-
renschwerpunkt. Leitthema 3: Prozeâstudien zur Oxidantienbildung und
Oxidationskapazita¨t: Zwischenbericht; Becker, K. H., Ed., Wuppertal,
Germany, 1998; pp 101-109.