Gas-phase reaction of OH radicals with phenol

Article abstract of DOI:10.1039/b208187c

Atmospheric degradation of benzene and its alkylated derivatives (toluene, xylenes, etc.) in the gas phase is exclusively initiated by the reaction with OH radicals. The gas-phase reaction of OH radicals with phenol was studied in a flow tube at 266-364 K and 100 mbar. In the presence of O₂, OH radicals were generated via the reaction sequence H + O₂ + M → HO₂ + M, HO₂ + NO → OH + NO₂ and in the absence of O₂ via H + NO₂ → OH + NO. The H-atom reaction did not compete with the reaction of OH radicals with phenol. At 295 K, the product distribution was studied for different O₂, NO, and NO₂ concentrations. Identified products were catechol, o-nitrophenol, and p-benzoquinone. Under atmospheric conditions, the OH/phenol adduct reacted exclusively with O₂. The observed formation of o-nitrophenol depended on the NO₂ concentration but independent on O₂.

Full text of DOI:10.1039/b208187c

View Article Online / Journal Homepage / Table of Contents for this issue
Gas-phase reaction of OH radicals with phenol
¨
Torsten Berndt and Olaf Boge
Institut fu¨r Tropospha¨renforschung e.V., Permoserstraße 15, 04318, Leipzig, Germany
Received 21st August 2002, Accepted 15th November 2002
First published as an Advance Article on the web 4th December 2002
The gas-phase reaction of OH radicals with phenol was investigated in a flow tube in the temperature range of
266–364 K and a pressure of 100 mbar. The product formation was followed by on-line FT-IR spectroscopy
and GC-MS measurements. Newly formed particles were detected by means of a low-pressure CPC
(condensation particle counter). In the presence of O2 , OH radicals were generated via the reaction sequence
H + O2 + M ! HO2 + M, HO2 + NO ! OH + NO2 and in the absence of O2 via H + NO2 ! OH + NO.
For evaluation of a possible competing process, the rate constant for H + phenol was measured,
k(H + phenol) ¼ (2.5 ꢀ 1.5) ꢁ 10ꢂ13 cm3 moleculeꢂ1 sꢂ1 (295 ꢀ 2 K, 25 mbar He). Under the experimental
conditions used the H-atom reaction does not compete with the reaction of OH radicals with phenol. At 295 K,
the product distribution was studied for different O2 , NO and NO2 concentrations. Identified products were
catechol, o-nitrophenol and p-benzoquinone. Under all experimental conditions catechol represented the main
product. The measured dependence of the catechol yield on NO and NO2 for constant O2 concentrations
allowed an estimate of the reactivity of the OH/phenol adduct towards O2 , NO and NO2 , k(adduct + O2)/
k(adduct + NO) > 10ꢂ3 and k(adduct + O2)/k(adduct + NO2) ¼ (1.4 ꢀ 0.5) ꢁ 10ꢂ4. For constant gas
composition, in the absence of additional NO2 , the product distribution was measured for different
temperatures. With increasing temperature the catechol yield increased from 0.37 ꢀ 0.06 (266 K) to 0.87 ꢀ 0.04
(364 K). The yields of o-nitrophenol and p-benzoquinone were nearly constant. Below 295 K, with decreasing
temperature enhanced formation of newly formed particles was observed. For realistic atmospheric conditions,
a catechol yield of 0.73–0.78 (295 K) can be recommended from this study.
Kinetic data for pathways (1), (ꢂ1) and (2) exist in the litera-
Introduction
ture.9,10 At 298 K the addition pathway (1) dominates over
the abstraction pathway (2) and the ratio k1/(k1 + k2) is in
the range 0.76–0.87.9,10 Kinetic investigations concerning the
subsequent reaction of the OH/phenol adduct with O2 , NO2
Atmospheric degradation of benzene and its alkylated deriva-
tives (toluene, xylenes, etc.) in the gas phase is exclusively
initiated by the reaction with OH radicals.1 Ring-retaining
products of this reaction are phenol-type compounds (phenol,
cresols, dimethylphenols, etc.).2 For benzene, experimentally
obtained phenol yields under nearly atmospheric conditions
are in the range 0.24–0.53.3–6 The fate of the formed phenol-
type compounds is governed by the reaction with OH and
NO3 radicals.7 For phenol, the reaction with OH radicals pro-
ceeds via a reversible addition step forming the OH/phenol
adduct (for simplification only the attack in ortho position is
shown).
and NO (rate constants in units of cm3 moleculeꢂ1 sꢂ1
:
k3 ¼ (3.0 ꢀ 0.7) ꢁ 10ꢂ14 (323 K), k4 ¼ (3.4 ꢀ 0.6) ꢁ 10ꢂ11 (331
K) and k5 < 7 ꢁ 10ꢂ14 (316–332 K), respectively)9,10 clearly
favour the reaction with O2 under atmospheric conditions.
Analogous studies for the reaction of phenoxy radicals with
O2 , NO2 and NO (k6 < 2 ꢁ 10ꢂ18 (280–500 K)11 or
k6 < 5 ꢁ 10ꢂ21 (296 K),12 k7 ¼ (2.08 ꢀ 0.15) ꢁ 10ꢂ12 (296 K)12
and k8 ¼ (1.88 ꢀ 0.16) ꢁ 10ꢂ12 (296 K),12 respectively) indicate
that the fate of phenoxy radicals in the atmosphere is not gov-
erned by the reaction with O2 , but by the reaction with NOx or
possibly other species such as O3 .
ð1= ꢂ 1Þ
and via H-atom abstraction predominantly from the OH sub-
To the authors’ knowledge, two product studies are reported
so far concerning the gas-phase reaction of OH radicals with
phenol.7,13 These experiments were performed in batch reac-
tors (room temperature, 1 bar synthetic air) using photolysis
of methyl nitrite for OH generation. The yields of the identi-
fied products were: catechol (0.804 ꢀ 0.121),13 o-nitrophenol
(0.067 ꢀ 0.015)7 and (0.058 ꢀ 0.010)13 as well as p-benzoqui-
none (0.037 ꢀ 0.012).13 An aqueous-phase study of the reac-
tion of OH radicals with phenol in the presence of NO2 at
pH ¼ 3 yielded o- and p-nitrophenol with formation yields
of 0.18 ꢀ 0.08 and 0.12 ꢀ 0.08, respectively, as well as traces
stituent group forming phenoxy radicals.2,8
ð2Þ
342
Phys. Chem. Chem. Phys., 2003, 5, 342–350
This journal is # The Owner Societies 2003
DOI: 10.1039/b208187c

View Article Online
of catechol.14 Generally, catechol formation can be described
starting from the OH/phenol adduct (ortho-position) assum-
ing H-abstraction by O2 :13
zoquinone) were obtained by flushing out the gas phase over
the solid sample by the carrier gas over a time period of 1.5–
23 h. During this period, FT-IR spectra of the resulting gas
mixture were measured in different intervals and in the end
of that the mass differences of the solid sample was deter-
mined. Under conditions of nearly constant sublimation in
the whole time period (Dm/m is small, surface of the solid
stays nearly unchanged, constant temperature), from the
mass difference and the total volume of the gas stream an
averaged concentration of the sublimated substance can be
calculated setting the FT-IR spectra on an absolute scale.
Using this approach, the reproducibility of the results was
found to be reasonable and determined absorption cross sec-
tions for the strongest absorption bands were (base 10, unit:
10ꢂ19 cm2 moleculeꢂ1); catechol: 4.0 ꢀ 0.3 at 1273 cmꢂ1, o-
nitrophenol: 3.9 ꢀ 0.4 at 1342 cmꢂ1 and p-benzoquinone:
9.1 ꢀ 3.2 at 1680 cmꢂ1. Error limits represent two standard
deviations. Because of too low sublimation vapour pressure,
for p-nitrophenol this procedure did not work. Appropriate
cross sections given in the literature13 were measured with
an instrumental resolution of 1 cmꢂ1 making an reasonable
comparison with the data of this study (instrumental resolu-
tion: 8 cmꢂ1) impossible.
For on-line GC-MS analysis (HP 5890 with HP MSD 5971),
a smaller part of the gas stream was pumped continuously
through a GC loop connected with the flow tube by means
of a heated transfer line. For product separation, a 30 m,
0.25 mm id column (HP 5MS) was chosen. Using the corre-
sponding ion traces, o-nitrophenol (m/z ¼ 139 u) and p-ben-
zoquinone (m/z ¼ 108 u) were detected efficiently allowing
the analysis of both substances down to flow-tube concentra-
tions of approximately 1010 molecule cmꢂ3. For signal calibra-
tion, GC-MS chromatograms and FT-IR spectra from the
same sample were measured simultaneously using the deter-
mined concentration from the FT-IR measurement as the
reference value (see the paragraph above). Because of probably
effective wall loss in the sampling device, the determination of
catechol was not successful using this GC method. Also for
the GC-MS analysis of o-nitrophenol and p-benzoquinone,
wall processes affecting the determined concentrations under
reaction conditions (presence of NOx) can not fully be ruled
out.
ð3aÞ
The formation of o-nitrophenol is attributed to the reaction of
phenoxy radicals with NO2 .7
ð7aÞ
Note that the given pathway represents an overall reaction
only. At least an intramolecular H-atom transfer step is needed
to produce a phenol-type substance starting from phenoxy
radicals.
The aim of the present study was to determine the products
of the reaction of OH radicals with phenol in dependence on
gas composition (O2 , NO and NO2 concentration) and tem-
perature. The experimental finding that k1 , kꢂ1 , k2 and k3
are pressure independent above 30–50 mbar (Ar as dilution
gas)10 justifies the use of a total pressure of 100 mbar and
allows the application of the results to atmospheric pressure.
The measurement of newly formed particles should clarify
the possible formation of condensable substances not detect-
able in the gas phase. From the sum of all experimental find-
ings under different conditions more insight into the
mechanism of the reaction of OH radicals with phenol is
expected. Additionally, for the evaluation of a possible com-
peting process under the applied reaction conditions, the rate
constant for the reaction of H-atoms with phenol was
measured.
Experimental
Furthermore, few runs were performed with particle mea-
surements or with cryo-trapping of the reaction products.
For particle measurements, a modified CPC (condensation
particle counter, TSI 7610) was used operating with FC 43
(perfluorotributylamine) as working fluid.15 The CPC was
directly attached to the flow tube approximately 45 cm down-
stream the mixing point of the reactants resulting in a resi-
The experiments were carried out at a pressure of 100 mbar
in a quartz glass flow tube (4.0 cm id, total length: 100 cm)
surrounded with a thermo-jacket allowing operation in the
temperature range of 266–364 K. In the presence of sufficient
O2 in the reaction gas ([O2] ¼ 1.68 ꢁ 1018 molecule cmꢂ3
)
OH radicals were produced via the reaction sequence:.
dence time of 0.63
s before particle counting. Before
H þ O2 þ M ! HO2 þ M
HO2 þ NO ! OH þ NO2
In these experiments NO additions were in the range (9.8–
ð9Þ
ð10Þ
measurements, the counting efficiency of the CPC at 100 mbar
was determined using an aerosol electrometer in a low-pressure
calibration set-up.16,17 For cryo-trapping, for a time period of
1 h the whole gas stream was pumped through a gas trap held
at liquid argon temperature. After disconnecting the trap from
the flow tube, the frozen material was immediately dissolved in
methanol and analysed by GC-MS (HP 6890 with HP MSD
5973). For product separation, an identical column as used
for the on-line GC-MS analysis was chosen.
244) ꢁ 1012 molecule cmꢂ3. In the absence of O2 ([O2] < 2 ꢁ
1013 molecule cmꢂ3, [NO2] ¼ (1.8–17.2) ꢁ 1013 molecule cmꢂ3
OH radicals were produced via:
)
H þ NO2 ! OH þ NO
ð11Þ
Approximately 50 cm downstream of the point for the
entrance of phenol and the additions diluted in the carrier
gas, in a side-arm H-atoms were generated in a commercial
microwave discharge (SAIREM GMP 03 K/SM) using a mix-
ture of 0.012–0.023 vol% H2 in He.
The main part of the reaction gas was pumped continu-
ously through a White cell (volume: 2050 cm3, optical path
length: 10 m) for FT-IR analysis (Nicolet Magna 750) using
an instrumental resolution of 8 cmꢂ1 averaging 200–2000
scans. At 100 mbar and 295 K, calibrated reference spectra
of the reaction products (catechol, o-nitrophenol and p-ben-
For the metering of phenol into the flow tube the same
approach was used as described for the signal calibration of
the reaction products. The resulting phenol concentration in
the tube was permanently controlled by FT-IR measure-
ments. The phenol source was found to be stable over a time
period of days.
The gas flows were set by calibrated mass flow controllers
(MKS 1259) and the pressure in the tube and in the gas cell
was measured using capacitive manometers (Baratron). The
total gas flow was set at 5000 standard cm3 minꢂ1 resulting
in a bulk residence time in the reaction zone of the flow tube
Phys. Chem. Chem. Phys., 2003, 5, 342–350
343

View Article Online
of 0.7 s (295 K). The temperature of the gas stream was mea-
sured by means of a thermocouple located in the middle of the
flow tube approximately 3 cm downstream the point for the
entrance of the H-atoms.
by changing the H-atom concentration using variable H2 con-
tents in the gas stream for the microwave discharge.
The gases used had stated purities as follows: He (99.999%),
O2 (99.9996%), Ar (99.999%) (Linde), H2 (99.999%), NO (0.5
vol% mixture in N2) (Messer Griesheim) and NO2 (99.5%)
(Merck). Phenol ( > 99.99%), catechol ( > 99%), mesitylene
(99.8%), o-nitrophenol ( > 99%), p-nitrophenol ( > 99%), 4-
nitrocatechol ( > 98%) and p-benzoquinone ( > 99.5%) (Fluka)
were used as purchased. 3-Nitrocatechol was synthesized by
treatment of catechol with acetic anhydride forming the
catechol monoacetate, followed by nitration in glacial acetic
acid, separation of the desired isomer and saponification as
described by Berti.23
The initial concentration of phenol was (5.9–6.7) ꢁ 1013
molecule cmꢂ3 and the amount of reacted phenol was in the
range (4.9–8.9) ꢁ 1012 molecule cmꢂ3. Consecutive reactions
of products with OH radicals were taken into consideration
in the following way. Assuming the simplified reaction scheme,
OH þ phenol ! Yi producti þ :::
OH þ producti ! :::
ð12Þ
ð13Þ
by dividing of the rate laws for the producti and phenol, eqn.
(I) follows.
d½productiꢃ
d½phenolꢃ
k13 ½productiꢃ
k12 ½phenolꢃ
Results and discussion
¼ ꢂYi þ
ðIÞ
Kinetics of the reaction H + phenol
Eqn. (I) was solved numerically by an explicit extrapolation
method18 connected with a Newton-technique18 for the deter-
Because OH radicals were generated from H-atoms via the
sequence of pathways (9) and (10) or via pathway (11), a pos-
sible contribution of the reaction of H-atom with phenol for
the total conversion of phenol had to be clarified. In the litera-
ture, rate constants for H + phenol are only available for high
temperature conditions.24 Here, the rate constant was obtained
at 295 ꢀ 2 K and a total pressure of 25 mbar He using the rela-
tive rate technique. Initial concentrations of 2.8 ꢁ 1013 mole-
cule cmꢂ3 for phenol and 1.1 ꢁ 1013 molecule cmꢂ3 for
mesitylene were applied. For the MS detection of phenol the
ion trace at m/z ¼ 94 u and for mesitylene at m/z ¼ 105 u
was measured. First, the reaction products of the reaction of
H-atoms with the organic substances were analysed to be sure
that the chosen ion traces were not afflicted by the products.
Identified products were 2-cyclohexenol, cyclohexanol, 2-
cyclohexenone and cyclohexanone arising from phenol as well
as the 1,3,5-trimethyl derivatives of cyclohexadiene, cyclohex-
ene and cyclohexane arising from mesitylene.22 In the 70 eV EI
spectra of all these products there are no appreciable signals at
m/z ¼ 94 u and m/z ¼ 105 u. In Fig. 1 the experimental find-
ings are plotted according to eqn. (II). The slope k14/k15 was
½productiꢃ
k13
mination of the formation yield Yi . Neglecting
,
k12 ½phenolꢃ
eqn. (I) represents the definition of the formation yield for pro-
ducti in the absence of consecutive processes and by integra-
tion follows Yi ¼ [producti]/D[phenol]. Here D[phenol]
stands for the reacted phenol. For an evaluation of the impor-
½productiꢃ
k13
k12
tance of the correction,
½phenolꢃ , the needed rate constants
were taken from literature (in units of cm3 moleculeꢂ1 ꢂ1):
s
k12 ¼ (2.8 ꢀ 0.6) ꢁ 10ꢂ11 (296 K),19 k13,catechol ¼ (1.04 ꢀ
0.21) ꢁ 10ꢂ10 (300 K),20 k13,o-nitrophenol ¼ 9 ꢁ 10ꢂ13 (room tem-
perature),8 k13,p-benzoquinone ¼ (4.6 ꢀ 0.9) ꢁ 10ꢂ11 (300 K).20
Because of the low reactivity of o-nitrophenol, the correction
for the consecutive reaction via pathway (13) was not neces-
sary. For p-benzoquinone, approximately 5–10% of this sub-
stance reacted with OH radicals via pathway (13). Because of
a total uncertainty of 50% in the determination of p-benzoqui-
none, this effect was neglected. Therefore, the product yields of
o-nitrophenol and p-benzoquinone were determined according
to Yi ¼ [producti]/D[phenol]. For catechol, up to 25% of this
substance reacted with OH radicals via pathway (13) making
the determination of the product yield with the help of eqn.
(I) necessary. For the temperature-dependent measure-
ments, Arrhenius expressions for k12 and k13,catechol were
needed. Because only for k12 this information is available in
the literature,21 k13,catechol/k12 was assumed to be tempera-
ture-independent.
Additionally, the rate constant for the reaction of H-atoms
with phenol was measured. The experimental approach used
has been described elsewhere,22 here only a brief discussion
is given. The experiments were performed in a quartz glass
flow tube (2.0 cm id, length: 50 cm) at 295 ꢀ 2 K and a total
pressure of 25 mbar He. H atoms were generated by means
of a microwave discharge (SAIREM GMP 03 K/SM). For
the determination of the rate constant the relative rate tech-
nique was employed and mesitylene served as the reference
substance.
H þ phenol ! products
H þ mesitylene ! products
ð14Þ
ð15Þ
From the measured signals without conversion, index ‘‘0’’,
and in the presence of H-atoms, index ‘‘t’’, eqn. (II) follows:.
k14
lnð½phenolꢃ0=½phenolꢃtÞ
¼
ðIIÞ
k15 lnð½mesityleneꢃ0=½mesityleneꢃtÞ
From the ratio k14/k15 , the desired rate constant k14 followed
using the literature value for k15
22 The disappearance of both
substances was observed using a quadrupole mass spectro-
meter (Balzers QMA 200) operating in the 70 eV EI mode.
Changing consumption of the organic substances was achieved
.
Fig. 1 Experimental data for the determination of the relative rate
constant of the reaction of H-atoms with phenol using mesitylene as
the reference substance; T ¼ 295 ꢀ 2 K, p ¼ 25 mbar He. Data are
plotted according to eqn. (II).
344
Phys. Chem. Chem. Phys., 2003, 5, 342–350

View Article Online
found to be 0.540 ꢀ 0.008, the intercept was ꢂ0.0015 ꢀ 0.006,
and the correlation coefficient was 0.989. Uncertainties
represent 2s. Using a value k15 ¼ (4.6 ꢀ 2.7) ꢁ 10ꢂ13 cm3
moleculeꢂ1 sꢂ1 from literature,22 k14 ¼ (2.5 ꢀ 1.5) ꢁ 10ꢂ13
cm3 moleculeꢂ1 sꢂ1 follows.
As a result of kinetic investigations of the reaction of H-
atoms with benzene at 298 K no pressure dependence in the
range 16–133 mbar Ar was observed.25 It was concluded that
the rate constant was in the high pressure limit even at 16
mbar. By analogy for H + phenol, it can be assumed that k14
represents a high pressure value also. Therefore, the measured
rate constant k14 is applicable for the conditions at 100 mbar
of this study.
catechol, 3-nitrocatechol and p-nitrophenol. The identified
butylated hydroxytoluene came from the solvents itself. The
unambiguous assignment of the several products was achieved
using the mass spectra of the reference substances. For experi-
ments performed in the absence of O2 (T ¼ 295 K, [NO2] ¼
cmꢂ3, D[phenol] ꢄ 8 ꢁ 1012 molecule cmꢂ3), the same product
pattern was visible as found for conditions in the presence of
O2 . The signal intensity for catechol, however, was lowered
compared to that for the nitrophenols. Repeated injections
of the same sample of dissolved reaction products showed a
clear decrease of the catechol signal and an increase especially
of the p-nitrophenol signal with time. Obviously, in the con-
densed phase further reactions took place probably initiated
by NO2 . This fact and the unknown collecting efficiency for
the different products in the cryo-trap prevented a determina-
tion of product yields.
4.1 ꢁ 1013 molecule cmꢂ3
,
[phenol] ¼ 6.5 ꢁ 1013 molecule
Mechanistic informations
From the detected products it can be speculated that the reac-
tion of H-atoms with phenol proceeds, on the one hand, via
ring addition producing 2-cyclohexenol and cyclohexanol after
consecutive H-atom additions. This would be a similar beha-
viour as described for methyl substituted benzenes.22 On the
other hand, also H-abstraction can be postulated producing
primary phenoxy radicals and H2 . The detected 2-cyclohexe-
none and cyclohexanone can be regarded as stable products
of phenoxy radicals from further reactions with H-atoms.
Despite the speculative nature and the uncertainty of the pri-
mary steps, the attack of H-atoms can be described tentatively
in the same manner as the attack of OH radicals, namely by
ring addition and by H-abstraction from the OH substituent
group, cf. the introductory paragraph.
Gas-phase products in the presence of O2
In these experiments the O2 concentration was 1.68 ꢁ 1018
molecule cmꢂ3 and the initial phenol concentration 6.5 ꢁ 1013
molecule cmꢂ3. A comparison of the H-atoms lifetimes with
respect to the reaction with O2 via pathway (9) and phenol
via pathway (14) of 4 ꢁ 10ꢂ6 s and 0.06 s, respectively, shows
that the reaction of H-atoms with phenol was negligible under
these conditions. The rate constant needed for pathway (9) was
taken from the literature, k9(M) ¼ 6.1 ꢁ 10ꢂ32 cm6 moleculeꢂ2
sꢂ1 at 298 K (M ¼ N2 here applied for the O2/He mixture),26
and that for pathway (14) from the present study. The amount
of reacted phenol was in the range (4.9–7.9) ꢁ 1012 molecule
cmꢂ3
.
Cryo-trapping
Fig. 3 demonstrates the determination of product yields by
means of FT-IR measurements (T ¼ 295 K, [O2] ¼
The intention of experiments with cryo-trapping was to get a
qualitative overview of the products. Experimental conditions
were similar to those with FT-IR and on-line GC-MS analysis,
see later. Fig. 2 shows a typical GC-MS chromatogram from a
cryo-trapping experiment (T ¼ 295 K, [O2] ¼ 1.68 ꢁ 1018
molecule cmꢂ3, [NO] ¼ 4.1 ꢁ 1013 molecule cmꢂ3, [phenol] ¼
6.5 ꢁ 1013 molecule cmꢂ3, D[phenol] ꢄ 5 ꢁ 1012 molecule cmꢂ3).
Detected products were p-benzoquinone, o-nitrophenol,
1.68 ꢁ 1018 molecule cmꢂ3, [NO] ¼ 4.9 ꢁ 1013 molecule cmꢂ3
,
D[phenol] ¼ 6.2 ꢁ 1012 molecule cmꢂ3). In the original differ-
ence spectrum (upper part) absorptions from consumed sub-
stances (phenol, NO) appear as negative bands and product
absorptions as positive bands. After addition of phenol bands
and subtraction of catechol bands, only the bands arising from
Fig. 3 Upper part: Typical FT-IR spectrum from the reaction of
OH radicals with phenol in the presence of O2 ; T ¼ 295 K,
[NO] ¼ 4.9 ꢁ 1013 molecule cmꢂ3, D[phenol] ¼ 6.2 ꢁ 1012 molecule
cmꢂ3. Lower part: Residual spectrum after addition of phenol bands
and subtraction of catechol bands using reference spectra each.
Fig. 2 GC-MS analysis of a cryo-trapping experiment from the reac-
tion of OH radicals with phenol in the presence of O2 . Products were
assigned using the mass spectra of the reference substances. ‘‘n.i.’’
stands for ‘‘not identified’’.
Phys. Chem. Chem. Phys., 2003, 5, 342–350
345

View Article Online
background CO2 , NO2 and from reacted NO remained (lower
part). There were no clear indications for further products,
such as p-benzoquinone and o-nitrophenol. The weak absorp-
tions in the range 1500–1000 cmꢂ1 were unspecific and prob-
ably influenced by the procedure of spectral subtraction.
Indeed, p-benzoquinone and o-nitrophenol were unambigu-
ously identified by on-line GC-MS analysis. The FT-IR
detection limits were estimated to be 3 ꢁ 1011 molecule cmꢂ3
for p-benzoquinone and 7 ꢁ 1011 molecule cmꢂ3 for o-nitro-
phenol corresponding to minimal formation yields of 0.04–
0.06 and 0.09–0.14, respectively, for the range of reacted
phenol in this study. Only in the presence of large NO2 concen-
trations, absorptions were visibly assigned to o-nitrophenol.
Because of the low p-benzoquinone and o-nitrophenol concen-
trations, on-line GC-MS measurements using ion traces had to
be chosen for quantitative detection. Note, as a result of on-
line GC-MS measurements, there was no experimental evi-
dence for the formation of 3-nitrocatechol and p-nitrophenol
as detected in cryo-trapping experiments. Either both sub-
stances were formed below the detection limits or effective wall
losses prevented an identification using this technique.
dimension as the amount of reacted phenol. Experimentally
obtained ratios [NO2]/D[phenol] from FT-IR measurements
increased from 1.11 to 1.77 with increasing NO concentration
indicating the occurrence of other NO2 producing steps than
pathway (10). The highest NO2 concentration obtained for
the maximum NO addition was 1.2 ꢁ 1013 molecule cmꢂ3
.
Influence of the NO2 concentration
In the next set of experiments the influence of the NO2 concen-
tration on the product yields was investigated (T ¼ 295 K,
[O2] ¼ 1.68 ꢁ 1018 molecule cmꢂ3, [NO] ¼ 1.2 ꢁ 1013 molecule
cmꢂ3, [NO2] ¼ (7.0–141) ꢁ 1012 molecule cmꢂ3). Even for the
highest NO2 concentration the reaction with H-atoms was of
minor importance; lifetime of H-atoms with respect to the
reaction with NO2 via pathway (11) 5.5 ꢁ 10ꢂ5 s (for the O2
reaction 4 ꢁ 10ꢂ6 s). The rate constant k11 ¼ 1.3 ꢁ 10ꢂ10 cm3
moleculeꢂ1 sꢂ1 was taken from the literature.27 Fig. 5 shows
the product yields obtained for varying initial NO2 concentra-
tion. Up to NO2 concentrations of (3–4) ꢁ 1013 molecule cmꢂ3
the catechol yield was nearly constant. For further increasing
NO2 concentrations, a small but significant decrease of the
catechol yield was detectable. The yield of o-nitrophenol, how-
ever, showed nearly an inverse behaviour. The p-benzoquinone
yield remained unaffected by NO2 with an averaged value of
0.010 ꢀ 0.005.
Influence of the NO concentration
In Fig. 4 the obtained product yields for different initial NO
concentration are given (T ¼ 295 K, [O2] ¼ 1.68 ꢁ 1018 mole-
cule cmꢂ3, [NO] ¼ (9.8–244) ꢁ 1012 molecule cmꢂ3). The cate-
chol yield of 0.73 ꢀ 0.04 (2s limits) was found to be
independent of NO concentration. Assuming for the highest
NO concentration a contribution of the NO reaction to the
conversion of the OH/phenol adduct via pathway (5) of less
than 10% in competition to the O2 reaction via pathway (3),
k3/k5 > 10ꢂ3 at 295 K can be estimated. Using the literature
data, k3 ¼ (3.0 ꢀ 0.7) ꢁ 10ꢂ14 cm3 moleculeꢂ1 sꢂ1 (323 K)10
and k5 < 7 ꢁ 10ꢂ14 cm3 moleculeꢂ1 sꢂ1 (316–332 K),9,10 k3/
k5 > 0.4 follows consistent with the ratio obtained in this
study. For o-nitrophenol and p-benzoquinone, because of rela-
tively large scatter no dependency on the NO concentration
was detectable, cf. Fig. 4. The averaged yields were
0.036 ꢀ 0.017 and 0.012 ꢀ 0.006, respectively. A NO2 source
under these experimental conditions was the OH generation
via pathway (10) and the NO2 levels were at least in the same
An attempt was undertaken to estimate the rate constant
ratio k3/k4 from the measured catechol yields in dependence
on NO2 concentrations for constant O2 concentration assum-
ing that catechol only came from the OH/phenol adduct.
First, it was investigated whether catechol was also formed
in the reaction of the OH/phenol adduct with NO2 or not.
ð4aÞ
Fig. 5 Product yields in the presence of O2 obtained by varying the
initial NO2 concentration; T ¼ 295 K, [NO] ¼ 1.2 ꢁ 1013 molecule
cmꢂ3. The dotted line depicts the modeled catechol yield according
to eqn. (III), see text. For comparison, the dashed line shows the mod-
eled catechol yield using the rate constants k3 and k4 from the Zetzsch
group.9,10
Fig. 4 Product yields in the presence of O2 determined for different
initial NO concentrations; T ¼ 295 K.
346
Phys. Chem. Chem. Phys., 2003, 5, 342–350

View Article Online
Under conditions of a preferred reaction of the OH/phenol
adduct with NO2 , a catechol yield of 0.35 ꢀ 0.03 was mea-
k1 k4a
sured, see later. With the help of pathways (1)–(4) including
(3a) and (4a), this yield corresponds to and the total
catechol yield can be described in the following way:
k1þk2 k4
ꢀ
ꢀ
ꢁ
.
k1
k3a
k3½O2ꢃ
Ycatechol
¼
1 þ
k1 þ k2 k3
k4½NO2ꢃ
ꢀ
ꢁꢁ
.
k4a
k4
k4½NO2ꢃ
k3½O2ꢃ
þ
1 þ
ðIIIÞ
By means of a procedure for non-linear parameter fitting,18
k1 k4a
k1þk2 k4
k1 k3a
k1þk2 k3
setting
¼ 0.35, the free parameters
and k3/k4
from eqn. (III) were estimated to be 0.78 ꢀ 0.02 and
(1.4 ꢀ 0.5) ꢁ 10ꢂ4, respectively (2s limits). The first free para-
meter corresponds to the catechol yield under conditions of
a sole catechol formation via pathway (3a) and is in good
agreement with the obtained catechol yield from experiments
with NO additions, cf. Fig. 4. In Fig. 5 the dotted line depicts
the modeled catechol yield according to eqn. (III). For
comparison, the literature data k3 ¼ (3.0 ꢀ 0.7) ꢁ 10ꢂ14 cm3
moleculeꢂ1 sꢂ1 (323 K)10 and k4 ¼ (3.4 ꢀ 0.6) ꢁ 10ꢂ11 cm3
moleculeꢂ1 sꢂ1 (331 K)9,10 result in a ratio k3/k4 ¼ (8.8 ꢀ
2.6) ꢁ 10ꢂ4. This is significantly higher compared to value of
this study, k3/k4 ¼ (1.4 ꢀ 0.5) ꢁ 10ꢂ4
. The dashed line
Fig. 6 Product yields in the absence of O2 obtained by varying the
initial NO2 concentration; T ¼ 295 K.
(Zetzsch group) in Fig. 5 shows the modeling curve of the cate-
chol yield using k3/k4 from the literature9,10 and the other
parameters as before. A reason for the discrepancy can not
be given. Determinations of k3 and k4 for different tempera-
tures yielded a small or negligible temperature dependence.9,10
So, the higher literature value for k3/k4 can not be explained
by the higher temperature used in the experiments from the
Zetzsch group.
with the reaction of phenoxy radicals with other intermediates
including the self reaction. The p-benzoquinone yield showed
an apparent, slight increase with increasing NO2 concen-
trations. Because of the insignificance of this effect, only the
averaged value of 0.010 ꢀ 0.006 can be given.
Gas-phase products in the absence of O2
Temperature dependence of the product yields
As a result of the experiments given before, with increasing
NO2 there was a decrease of the catechol yield and at the same
time an increase of the of o-nitrophenol yield, cf. Fig. 5.
Experiments in the absence of O2 were conducted to find out
the product distribution under conditions of a preferred reac-
tion of the OH/phenol adduct with NO2 (T ¼ 295 K,
[NO2] ¼ (1.8–17.2) ꢁ 1013 molecule cmꢂ3, Ar as the carrier
gas). The initial phenol concentration was 5.9 ꢁ 1013 molecule
cmꢂ3 and the amount of reacted phenol (5.9–8.9) ꢁ 1012 mole-
cule cmꢂ3. OH radicals were produced from the reaction of H-
atoms with NO2 via pathway (11). Using the rate constants as
mentioned above, the H-atom lifetimes with respect to the
reaction with NO2 via pathway (11) and with phenol via path-
way (14) were p4ꢁ10ꢂ4 s and 0.07 s, respectively, making the
latter step negligible. Fig. 6 shows the product yields for vary-
ing initial NO2 concentration. With increasing NO2 the cate-
chol yield increased and for [NO2] > 4 ꢁ 1013 molecule cmꢂ3
a constant yield of 0.35 ꢀ 0.03 was measured. This behaviour
can be tentatively explained by the formation of catechol via
pathway (4a) to be in competition mainly with the self reaction
of the OH/phenol adduct. It is to be noted, that the HONO
formation pointed out in pathway (4a) was not experimentally
proved. For o-nitrophenol, the formation yield increased con-
tinuously with increasing NO2 in the whole NO2 range. A com-
parison of the shape of the curves describing the yields for
catechol and o-nitrophenol shows that both substances were
not produced in a simple parallel process starting from the
OH/phenol adduct, cf. Fig. 6. On the other hand, the NO2
dependent yields of o-nitrophenol in the absence of O2 (Fig.
6) and in the presence of O2 (Fig. 5) are in reasonable agree-
ment. From these facts it can be speculated that o-nitrophenol
was formed from phenoxy radicals via pathway (7a). The
influence of the NO2 concentration on the o-nitrophenol
yield can be described by pathway (7a) to be in competition
In the temperature range of 266–364 K, product yields were
measured for constant initial mole fractions of the reactants
(concentrations calculated for T ¼ 295 K, [O2] ¼ 1.68 ꢁ 1018
molecule cmꢂ3, [NO] ¼ 4.9 ꢁ 1013 molecule cmꢂ3, [phenol] ¼
(6.6–6.7) ꢁ 1013 molecule cmꢂ3). The amount of reacted phenol
was in the range (5.4–7.2) ꢁ 1012 molecule cmꢂ3. Because k3
and k4 are weakly temperature-dependent,9,10 in the whole
temperature range the fate of the OH/phenol adduct was gov-
erned by the reaction with O2 via pathway (3). Fig. 7 depicts
the experimental findings. The catechol yield showed a distinct
increase with increasing temperature, 0.37 ꢀ 0.06 for 266 K
and 0.87 ꢀ 0.04 for 364 K. In principle, a similar behaviour
was found for the phenol yield from the reaction of OH radi-
cals with benzene.6 Because in eqn. (I) k13,catechol/k12 was
assumed to be temperature-independent, the determination
of the catechol yield via eqn. (I) was probably incorrect. In
Fig. 7 the averaged ‘‘raw data’’ of the catechol yield deter-
mined by neglecting pathway (13), Ycatechol ¼ [catechol]/
D[phenol], are also given as lower limit. From kinetic data,10
the formation yield of the OH/phenol adduct, k1/(k1 + k2),
can be expected to be 0.7 at 364 K and by extrapolation 0.8
at 266 K. Other kinetic data from the same group9 yield some-
what different estimates for the OH/phenol adduct yields (0.74
at 364 K and 0.91 at 266 K). From the observed catechol yields
above 320 K of approximately 0.85 (raw data: 0.72), cf. Fig. 7,
it can be speculated that under these conditions the OH/phe-
nol adduct is mainly converted to catechol via pathway (3a).
Starting from 320 K, with decreasing temperature other steps
than pathway (3a) become more important for the fate of
the OH/phenol adduct. In contrast to catechol, for the
yields of o-nitrophenol and p-benzoquinone there was no tem-
perature effect detectable. The averaged yields were 0.043 ꢀ
0.017 and 0.012 ꢀ 0.006, respectively. Summarising the forma-
tion yields, the gas-phase products accounted for 92 ꢀ 6% of
Phys. Chem. Chem. Phys., 2003, 5, 342–350
347

View Article Online
Fig. 8 Measured particle numbers (d > 15 nm) for constant initial
mole fractions in dependence on temperature and the amount of
reacted phenol; [NO] ¼ 1.2 ꢁ 1013 molecule cmꢂ3, [O2] ¼ 1.68 ꢁ 1018
molecule cmꢂ3 (calculated for 295 K).
Fig. 7 Temperature-dependent product yields in the presence of O2
measured for constant initial mole fractions; [NO] ¼ 4.9 ꢁ 1013 mole-
cule cmꢂ3 (calculated for 295 K). The ‘‘raw data’’ of the catechol yield
stand for values not corrected for the consecutive reaction with OH,
see text.
6.4 ꢁ 1012 molecule cmꢂ3, respectively) n ¼ 16 was found.
A power of 16 in this nucleation law demonstrates the extre-
mely strong increase of the number of formed particles with
increasing concentration of the condensable species. Because
n was determined just for a single temperature, the found value
can be regarded only as a rough estimate. For the nucleation of
sulfuric acid, n ¼ 7–13 in dependence on the humidity in the
system was experimentally observed.28 For the chemical nature
of the nucleating substances in the phenol system, there exists
no reasoning. To the best of our knowledge, there is no infor-
mation available in the literature describing the formation of
condensable substances from the reaction of OH radicals with
phenol.
the carbon of the reacted phenol at 364 K and 42 ꢀ 8% at 266
K.
Particle formation
As discussed in the paragraph before, with decreasing tempera-
ture there was an increasing lack of carbon in the observed gas-
phase products. A possible explanation for this behaviour is
the enhanced occurrence of condensable products at lower
temperature and the formation of particles. Particle measure-
ments were performed in the temperature range of 275–295
K using again constant initial mole fractions of the reactants
(concentrations calculated for T ¼ 295 K, [O2] ¼ 1.68 ꢁ 1018
molecule cmꢂ3, [NO] ¼ 1.2 ꢁ 1013 molecule cmꢂ3, [phenol] ¼
6.4 ꢁ 1013 molecule cmꢂ3). In Fig. 8 the measured particle
numbers for four different amounts of converted phenol for
different temperatures are plotted. The varying amounts of
converted phenol were achieved changing the H2 contents in
the gas stream for the microwave discharge and so the amount
of OH radicals. The observed number of newly formed parti-
cles in this system was found to be strongly dependent on tem-
perature and the amount of reacted phenol. The temperature
effect can be explained that with decreasing temperature
probably more condensable species are formed and/or the
saturation vapour pressure of the condensable species is going
down supporting the nucleation process. The observed number
of formed particles for different gas-phase concentration of the
condensable species for constant time can be explained tenta-
tively by eqn. (IV).28
The particle measurements showed that with decreasing
temperature the particle number increased. That helps to fulfil
the carbon balance, at least qualitatively.
Summary and application to the atmosphere
In Table 1 experimentally obtained product yields from the
gas-phase reaction of OH radicals with phenol from the litera-
ture along with the results of this study are given. For NO2
concentrations <(3–4) ꢁ 1013 molecule cmꢂ3 and room tem-
perature, the catechol yield of 0.73–0.78 was found to be in
good agreement with the result of
a chamber study,
(0.804 ꢀ 0.121).13 The formation of catechol can be described
by the reaction of O2 with the OH/phenol adduct via pathway
(3a). Because under realistic atmospheric conditions the fate of
the OH/phenol adduct is also governed by pathway (3a) and
no pressure effects can be expected, the catechol yield of
0.73–0.78 (295 K) is applicable for atmospheric conditions.
Considering the formation of phenol from the reaction of
OH radicals with benzene with a yield of 0.24–0.533–6 at room
temperature, the main products of the OH radical initiated
degradation of these simple aromatics in the atmosphere are
the corresponding hydroxylated species each.
N ¼ Pcncondens
ðIVÞ
N stands for the number of formed particles, ccondens for the
gas-phase concentration of the condensable species, n and P
are free parameters. Under conditions of a kinetically con-
trolled nucleation process, n characterises the number of mole-
cules in the critical cluster. Assuming that ccondens represents a
constant fraction of the reacted phenol for a fixed temperature,
from measurements at 283 K (N ¼ 500, 1.1 ꢁ 104 and 5 ꢁ 104
(estimate) cmꢂ3 for D[phenol] ¼ 4.8 ꢁ 1012, 5.8 ꢁ 1012 and
348
Phys. Chem. Chem. Phys., 2003, 5, 342–350

View Article Online
Table 1 Compilation of product yields
Product yields
Catechol
—
Initial concentrations/
molecule cmꢂ3
Pressure/mbar
Temperature/K
o-Nitrophenol
0.067 ꢀ 0.015
p-Benzoquinone
Source
Ref. 7
ꢄ990 synthetic air
296 ꢀ 2
Phenol: (1.5–2.8) ꢁ 1013
NO: 2.4 ꢁ 1014
—
1000 synthetic air
100
298 ꢀ 2
Phenol: (2.6–10.8) ꢁ 1013
NO: (1.2–4.2) ꢁ 1013
Phenol: 6.5 ꢁ 1013
O2 : 1.68 ꢁ 1018
0.804 ꢀ 0.121
0.73 ꢀ 0.04
0.058 ꢀ 0.010
0.036 ꢀ 0.017
0.037 ꢀ 0.012
0.012 ꢀ 0.006
Ref. 13
295
This study
NO: (9.8–244) ꢁ 1012
Phenol: 6.5 ꢁ 1013
O2 : 1.68 ꢁ 1018
100
295
NO2-dependent
(0.78 ꢀ 0.02 for
[NO2] ! 0)
NO2-dependent
0.010 ꢀ 0.005
This study
NO: 1.2 ꢁ 1013
NO2 : (7.0–141) ꢁ 1012
Phenol: 5.9 ꢁ 1013
O2 : <2 ꢁ 1013
100
100
295
NO2-dependent
(0.35 ꢀ 0.03 for
[NO2] > 4 ꢁ 1013
T-dependent
NO2-dependent
0.010 ꢀ 0.006
0.012 ꢀ 0.006
This study
This study
NO2 : (1.8–17.2) ꢁ 1013
a Phenol: (6.6–6.7) ꢁ 1013
a O2 : 1.68 ꢁ 1018
)
266–364
0.043 ꢀ 0.017
a NO: 4.9 ꢁ 1013
a
Calculated for 295 K.
The temperature-dependent measurements showed a clear
decrease of the catechol yield with decreasing temperature.
The observed formation of o-nitrophenol was found to be
dependent on the NO2 concentration but independent on
O2 , cf. Fig. 5 and 6. At least for relatively low NO2 concentra-
tions, [NO2] < 2 ꢁ 1013 molecule cmꢂ3, it seems to be likely
that o-nitrophenol was formed from the reaction of NO2 with
phenoxy radicals via pathway (7a). It is not clear whether the
observed o-nitrophenol yields (0.067 ꢀ 0.015 from ref. 7,
0.058 ꢀ 0.010 from ref. 13 and 0.036 ꢀ 0.017 from this study)
can be applied for realistic atmospheric conditions with NO2
concentrations of approximately 5 ꢁ 1011 molecule cmꢂ3
(atmospheric mixing ratio: 20 ppb). In the real atmosphere,
phenoxy radicals react probably also with O3 or other species
in competition to the reaction with NOx .12 To clarify that,
more experimental work is needed concerning the atmospheric
fate of phenoxy radicals.
The detected yield of p-benzoquinone was not influenced by
experimental conditions (NO, NO2 , O2 , temperature) and
there were no indications for the formation pathways of this
substance. Therefore, as a result of this study, a discussion
of the atmospheric relevance of the p-benzoquinone yield is
highly speculative.
The measured dependence of the catechol yield on NO and
NO2 for constant O2 concentrations allowed an estimate of the
reactivity of the OH/phenol adduct towards O2 , NO and
NO2 , k3/k5 > 10ꢂ3 and k3/k4 ¼ (1.4 ꢀ 0.5) ꢁ 10ꢂ4. As com-
pared with literature data,9,10 the latter ratio is smaller by a
factor of six. Nevertheless, this estimate confirms the fact that
under atmospheric conditions the OH/phenol adduct reacts
exclusively with O2 .
References
1
2
3
B. J. Finlayson-Pitts and J. N. Pitts, Jr., in Chemistry of the Upper
and Lower Atmosphere, Academic Press, 2000.
R. Atkinson, J. Phys. Chem. Ref. Data, 1994, Monograph
No. 2, 47.
R. Atkinson, S. M. Aschmann, J. Arey and W. P. L. Carter, Int. J.
Chem. Kinet., 1989, 21, 801; reassessment in: R. Atkinson and
S. M. Aschmann, Int. J. Chem. Kinet., 1994, 26, 929.
B. Bohn, M. Elend and C. Zetzsch, in CMD Annual Report 98,
ed. EUROTRAC-2 International Scientific Secretariat, Munich,
1999, p. 112.
4
5
R. Volkamer, B. Klotz, I. Barnes, T. Imamura, K. Wirtz, N.
Washida, K. H. Becker and U. Platt, Phys. Chem. Chem. Phys.,
2002, 4, 1598.
6
7
8
9
T. Berndt and O. Bo¨ge, Phys. Chem. Chem. Phys., 2001, 3,
4946.
R. Atkinson, S. M. Aschmann and J. Arey, Environ. Sci. Technol.,
1992, 26, 1397.
R. Atkinson, J. Phys. Chem. Ref. Data, 1989, Monograph No. 1,
231.
R. Knispel, R. Koch, M. Siese and C. Zetzsch, Ber. Bunsen-Ges.
Phys. Chem., 1990, 94, 1375.
10 R. Koch, Ph.D. Thesis, Hannover, 1992; R. Koch, J. Nowack, M.
Siese and C. Zetzsch, in Proc. CEC/EUROTRAC Discussion
Meeting, York, 1991.
11 F. Berho and R. Lesclaux, Chem. Phys. Lett., 1997, 279,
289.
12 J. Platz, O. J. Nielsen, T. J. Wallington, J. C. Ball, M. D. Hurley,
A. M. Straccia, W. F. Schneider and J. Sehested, J. Phys. Chem.
A, 1998, 102, 7964.
13 R. Olariu, B. Klotz, I. Barnes, K. H. Becker and R. Mocanu,
Atmos. Environ., 2002, 36, 3685.
14 P. Barzaghi and H. Herrmann, Phys. Chem. Chem. Phys., 2002,
4, 3669.
15 C. A. Brock, J. C. Wilson, D. W. Gesler, B. G. Lafleur and J. M.
Reeves, in AGU Fall Meet. Suppl. (abstracts), San Francisco,
1998, p. F138.
16 M. Hermann and J. Heintzenberg, private communication.
17 M. Hermann and A. Wiedensohler, J. Aerosol. Sci., 2001, 32,
975.
18 J. Stoer and R. Bulirsch, Einfu¨hrung in die Numerische Mathema-
tik II, Springer Verlag, Berlin, 1978.
For evaluation of a possible competing process under the
experimental conditions used, the rate constant for H + phenol
was measured, k14 ¼ (2.5 ꢀ 1.5) ꢁ 10ꢂ13 cm3 moleculeꢂ1 sꢂ1
(295 ꢀ 2 K, 25 mbar He), representing the first determination
of this value at room temperature.
19 M. Rinke and C. Zetzsch, Ber. Bunsen-Ges. Phys. Chem., 1984,
88, 55.
20 R. I. Olariu, I. Barnes, K. H. Becker and B. Klotz, Int. J. Chem.
Kinet., 2000, 32, 696.
21 M. Semadeni, D. W. Stocker and J. A. Kerr, Int. J. Chem. Kinet.,
1995, 27, 287.
22 T. Berndt and O. Bo¨ge, Int. J. Chem. Kinet., 2001, 33, 124.
23 G. Berti, Gazz. Chim. Ital., 1957, 87, 659.
Acknowledgements
The authors thank Markus Hermann and Jost Heintzenberg
for help during the particle measurements, Frank Stratmann
and Cornelius Zetzsch for useful discussions, and Herwig
Macholeth for technical assistance.
Phys. Chem. Chem. Phys., 2003, 5, 342–350
349

View Article Online
24 D. L. Baulch, C. J. Cobos, R. A. Cox, C. Esser, P. Frank,
Th. Just, J. A. Kerr, M. J. Pilling, J. Troe, R. W. Walker and
J. Warnatz, J. Phys. Chem. Ref. Data, 1992, 21, 411.
25 J. M. Nicovich and A. R. Ravishankara, J. Phys. Chem., 1984, 88,
2534.
27 H. Gg. Wagner, U. Welzbach and R. Zellner, Ber. Bunsen-Ges.
Phys. Chem., 1976, 80, 1023.
28 S. M. Ball, D. R. Hanson, F. L. Eisele and P. H. McMurry, J.
Geophys. Res., 1999, 104, 23 709.
26 K. J. Hsu, J. L. Durant and F. Kaufman, J. Phys. Chem., 1987,
91, 1895.
350
Phys. Chem. Chem. Phys., 2003, 5, 342–350