Reaction of phenylperoxy radicals with NO2 at 298 K

Article abstract of DOI:10.1039/b705193j

In the present work, phenylperoxy radicals were generated by stationary 254 nm photolysis of iodobenzene and nitrosobenzene in the presence of O ₂ and NO₂ at 298 K and a total pressure of 1 bar (M = N₂). Experiments were performed on time scales of seconds or minutes in a temperature controlled photoreactor made of quartz (v = 209 L). Major gas phase products identified and quantified in situ by long-path IR absorption include N₂O₅, NO, HONO, HNO₃, CO, and o-nitrophenol. In addition, evidence is presented for the formation of an aerosol consisting of p-nitrophenol. The occurrence of N₂O ₅ as a major product in both reaction systems, the strong loss of NO₂ in the iodobenzene system and the comparison of measured product distributions with the results of numerical model calculations suggest that the reaction C₆H₅O₂ + NO₂ → C ₆H₅O + NO₃; k₅ occurs in both photolysis systems, a major part of the NO₃ being scavenged as N ₂O₅. The results of ab initio calculations imply that reaction (5) proceeds via a short-lived peroxynitrate intermediate. In the photolysis of nitrosobenzene-NO₂-O₂-N₂ mixtures, NO and NO₂ compete for C₆H₅O ₂ radicals. Comparison of measured and modelled product distributions allows to set a lower limit of k₅ > 1 × 10^-12 cm³ molecule^-1 s^-1 at 298 K. This lower limit is consistent with the assumption that k₅ is equal to the high pressure recombination rate constant of RO₂ + NO₂ → RO₂NO₂ reactions, i.e. with k₅ ≈ 7 × 10^-12 cm³ molecule^-1 s^-1 at 298K; 1 bar. the Owner Societies.

Full text of DOI:10.1039/b705193j

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Reaction of phenylperoxy radicals with NO2 at 298 K
Stefan Jagiella and Friedhelm Zabel*
Received 5th April 2007, Accepted 20th June 2007
First published as an Advance Article on the web 20th July 2007
DOI: 10.1039/b705193j
In the present work, phenylperoxy radicals were generated by stationary 254 nm photolysis of
iodobenzene and nitrosobenzene in the presence of O2 and NO2 at 298 K and a total pressure of
1 bar (M = N2). Experiments were performed on time scales of seconds or minutes in a
temperature controlled photoreactor made of quartz (v = 209 L). Major gas phase products
identified and quantified in situ by long-path IR absorption include N2O5, NO, HONO, HNO3,
CO, and o-nitrophenol. In addition, evidence is presented for the formation of an aerosol
consisting of p-nitrophenol. The occurrence of N2O5 as a major product in both reaction systems,
the strong loss of NO2 in the iodobenzene system and the comparison of measured product
distributions with the results of numerical model calculations suggest that the reaction
C6H5O2 þ NO2 ! C6H5O þ NO3; k5
ð5Þ
occurs in both photolysis systems, a major part of the NO3 being scavenged as N2O5. The results
of ab initio calculations imply that reaction (5) proceeds via a short-lived peroxynitrate
intermediate. In the photolysis of nitrosobenzene–NO2–O2–N2 mixtures, NO and NO2 compete
for C6H5O2 radicals. Comparison of measured and modelled product distributions allows to set a
lower limit of k5 4 1 ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1 at 298 K. This lower limit is consistent
with the assumption that k5 is equal to the high pressure recombination rate constant of RO2 +
NO2 - RO2NO2 reactions, i.e. with
k5 ꢂ 7 ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1 at 298 K; 1 bar:
tion in reaction (2a). The most stable peroxynitrates known to
Introduction
date are those with R = R0–C(O). The simplest member of
this group is acetyl peroxynitrate (CH3C(O)O2NO2 = PAN)
which is an important component of photochemical smog and
contributes to the long-range transport of NOx (=NO +
NO2) in the troposphere. However, in our attempts to estab-
lish structure–activity relationships for the thermal lifetimes of
peroxynitrates3–5 it proved to be impossible to identify
RO2NO2 as a product of the generation of R radicals in the
presence of O2 and NO2 for R = HC(O)CO,4 CH3C(O)CO,4
CH3SO,4 and C6H5.5 For R = HC(O)CO, Orlando and
Tyndall6 suggested that HC(O)C(O)O2NO2 is possibly not
formed from HC(O)C(O)O2 and NO2 because of the weak
O–O bond in HC(O)C(O)OONO2, leading to the decomposi-
tion products HC(O)C(O)O and NO3 or, by simultaneous
fission of the O–O and C(O)–C bonds, directly to HCO, CO2,
and NO3. Independently, we detected large yields of N2O5 in
reaction systems generating CH3C(O)C(O)7 and C6H58 radicals
in the presence of O2 and NO2, suggesting that reaction (4),
Peroxy (RO2) radicals are important intermediates in the
atmospheric degradation of hydrocarbons.1,2 Major loss reac-
tions in the troposphere are:
RO2 þ NO ! RO þ NO2
ð1aÞ
ð1bÞ
RO2 þ NO þ M ! RONO2 þ M
!
RO þ NO þ M RO NO þ M
ð2a; ꢁ2bÞ
2
2
2
2
RO2 þ R0O2 ! products
ðR; R0 ¼ H or organic sgroupÞ
ð3Þ
The relative importance of reactions (1)–(3) mainly depends
on the relative and absolute concentrations of NO and NO2 as
well as on the chemical nature of R. Under atmospheric
conditions, reaction (2) is reversible. In general, peroxynitrates
RO2NO2 are formed when organic radicals R are generated in
the presence of O2 and NO2. At 298 K and atmospheric
pressure, the thermal lifetimes of peroxynitrates are between
about 0.5 and 10 000 s;3 however, their actual lifetime in a
specific air mass can be considerably longer due to re-forma-
RO2 þ NO2 ! RO þ NO3
ð4Þ
should also be considered as a possible reaction path for the
interaction of certain RO2 radicals with NO2. Investigation of
reaction (4) for R = C6H5, i.e. of reaction (5),
C6H5O2 þ NO2 ! C6H5O þ NO3;
ð5Þ
Institut fur Physikalische Chemie, Universitat Stuttgart,
¨
¨
Pfaffenwaldring 55, D-70569 Stuttgart, Germany.
E-mail: f.zabel@ipc.uni-stuttgart.de
is the subject of the present publication.
ꢃc
This journal is the Owner Societies 2007
5036 | Phys. Chem. Chem. Phys., 2007, 9, 5036–5051

View Article Online
Phenyl radicals are formed in the atmospheric degradation
of toluene via the ring-retaining intermediate benzaldehyde9,10
as well as in combustion systems by thermal decomposition of
benzene and other aromatic compounds.11 Whereas the reac-
tion of phenyl radicals with O2 generates O and H atoms at
high temperatures (41000 K),11–13 addition of O2 to form
phenylperoxy radicals seems to be the only reaction path at
low temperatures (o500 K).11,14 The products of the reaction
of phenylperoxy radicals with NO2 are not known. Phenyl
peroxynitrate might be formed according to reaction (2a) since
the thermal lifetime of phenyl peroxynitrate with respect to
RO2 and NO2 formation was predicted to be quite long at 298
calibrated volumes or via syringes. Mixing of the gaseous
components was supported by two fans.
Phenylperoxy radicals were generated by continuous photo-
lysis of iodobenzene or nitrosobenzene in the presence of O2,
NO2, and N2 at 253.7 nm, using 24 low pressure mercury
lamps (Philips TUV 36W). Standard photolysis time was 30 s
for iodobenzene. During this period, the loss of iodobenzene
was about 20%. Shorter photolysis times were used for
nitrosobenzene. Typical initial concentrations of the trace
gases were in the range 1 ꢀ 1014 – 3 ꢀ 1014 molecule cmꢁ3
for iodobenzene, nitrosobenzene, and NO2. Most experiments
were performed in synthetic air; in one experiment, 1000 mbar
O2 was applied.
3
K and 1 bar (1.6 h ). Kirchner, however, did not succeed in
scavenging phenyl radicals by O2 and NO2 as phenyl perox-
ynitrate in the 254 nm photolysis of C6H5I–O2–NO2–N2
mixtures.5 Rather, he detected o-nitrophenol as a major
product. A possible reason for this failure is the high reso-
nance stabilization energy of the phenoxy radical of approxi-
Concentrations were determined by spectral subtraction of
reference IR spectra. Research grade chemicals used in this
work include iodobenzene (Aldrich, 499%), nitrosobenzene
(Aldrich, 497%), o-nitrophenol (Riedel-de Haen, 499%),
¨
nitrobenzene (Acros, 499%), iodoethane (Merck Schuchardt,
499%), NO, NO2, and CO. Glyoxal was prepared by drying
an aqueous solution (40%, Merck Schuchardt) with a mix-
ture of P2O5 and sea sand. IR absorption cross sections
s (ꢇ ln(I0/I)/(cd)) were determined from IR spectra of pure
purchased samples based on pressure measurements whenever
possible. IR absorption coefficients of N2O5 were determined
from experiments reacting O3 with an excess of NO2 in N2
as buffer gas, assuming stoichiometric conversion of about
95% of the initial O3 into NO3. The IR absorption coefficients
of O3 were calibrated by measuring the IR absorbance of
an O3 sample simultaneously to its UV absorbance at 254 nm,
using the long-path UV White mirror system described in
ref. 17.
mately 73 kJ mol ,
ꢁ1 15 potentially leading to fission of the weak
RO–ONO2 bond in place of the stronger ROO–NO2 bond:
C6H5O2 þ NO2 ½! C6H5OONOꢄ2ꢅ ! C6H5O þ NO3 ð50Þ
Actually, reaction (5) has been proposed previously by Hendry
and Kenley16 for thermochemical reasons. Using the most
recent thermochemistry for C6H5O11 and C6H5O2,11 reaction
(5) is found to be exothermic by 72 kJ molꢁ1. It is unknown,
however, if there is a considerable energy barrier for this
reaction.
In the present work, kinetic and thermochemical evidence is
presented which supports the idea that the reaction of C6H5O2
with NO2 indeed leads to NO3 and phenoxy radicals. NO3
formation is inferred from the identification of the product
N2O5 that is formed in the reaction
o-Nitrophenol samples (vapour pressure at 298 K: B0.25
mbar18) were directly evaporated into the evacuated reaction
chamber. The evaporation rate was slow and of the same order
as the (reproducible) leak rate of the chamber. Thus the IR
absorption coefficient of o-nitrophenol was estimated from
pressure measurements in the reaction chamber, taking the
leak rate into account.
NO2 þ NO3ðþMÞ ! N2O5ðþMÞ:
ð6Þ
Experimental
Quantum chemical calculations
The experiments were performed at 298.1 ꢆ 0.5 K and a total
pressure of 1 bar in a temperature controlled photoreactor
constructed of quartz (v = 209 L) that has previously been
described in more detail.17 Briefly, the photoreactor consists of
two concentric quartz tubes, being closed by disks made of
stainless steel that are internally covered with a Teflon coating.
The space between the quartz tubes as well as cavities within
the end flanges are filled with silicone oil as a heating/cooling
agent, being circulated between the reaction chamber and a
cryostat. 24 fluorescent lamps of different types can be posi-
tioned around the outer quartz tube to photolyse the gas
mixtures within the reaction chamber. Gas mixtures are
analysed by long-path IR absorption, using a White mirror
system installed within the chamber (optical pathlength =
29.0 m). IR absorption spectra are monitored at a spectral
resolution of 1 cmꢁ1, using an FT-IR spectrometer (Nicolet
MAGNA 560). The pumping system consists of a turbomole-
cular pump (Pfeiffer TMH 260SG) and a 35 m3 hꢁ1 forepump.
Gaseous chemicals can be added to the chamber from flasks of
DFT and ab initio quantum chemical calculations on IR
spectra of potential reaction products and on reaction enthal-
pies were performed in order to assist the conclusions drawn
from the experiments. Geometries were optimised and harmo-
nic frequencies were calculated at the B3LYP/6-31G(d) level
with the Gaussian 98 ab initio quantum chemistry package.19
On the basis of these geometries the electronic energies were
calculated with the MOLPRO software package20 at the RHF-
UCCSD(T) level21 for radical species and the RHF-CCSD(T)
level22 for closed shell species with a cc-pVTZ basis set. For
the NO3 radical, care was taken that the RHF reference wave
function showed D3h symmetry.23
Kinetic simulation of the reaction mechanism
A set of reactions that is able to rationalize the measured
product distributions was compiled (Table 1). Computer-aided
simulation of the photolysis reactions was then employed to
ꢃc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 5036–5051 | 5037

View Article Online
Table 1 Reaction mechanism for the photolysis of C6H5I–O2–NO2–N2 and C6H5NO–O2–NO2–N2 mixtures at 298 K and 1 bar
No. Reaction
k/sꢁ1, cm3 molecꢁ1 sꢁ1 References for k, notes
22a C6H5I + hn - C6H5 + I
6.2 ꢀ 10ꢁ3
1.2 ꢀ 10ꢁ3
2.8 ꢀ 10ꢁ2
3.5 ꢀ 10ꢁ2
1.7 ꢀ 10ꢁ11
3.0 ꢀ 10ꢁ11
1.0 ꢀ 105
This work, from the first order decay of C6H5I
This work, fit to the measured CO yield, see text
22b C6H5I + hn - C6H5* + I
39a C6H5NO + hn - NO + C6H5
39b C6H5NO + hn - NO + C6H5**
This work, from the first order decay of C6H5NO
This work, fit to the measured CO yield, see text
Baulch et al.11 (2005), Yu and Lin14 (1993, 1994)
Arbitrary value, no competing reaction, products assumed
Arbitrary value, no competing reaction, products assumed
Arbitrary value, no competing reaction, products assumed
Arbitrary value, no competing reaction, products assumed
This work, variable, see text
23
40
41
42
43
5
C6H5 + O2 (+M) - C6H5O2 (+M)
C6H5* + O2 - C6H5O* + O
C6H5O* - CO (+C5H5?)
C6H5** + O2 - C6H5O** + O
C6H5O** - CO ( + C5H5?)
C6H5O2 + NO2 - C6H5O + NO3
C6H5O2 + NO - C6H5O + NO2
C6H5O + NO2 - o-nitrophenol
C6H5O + NO2 - p-nitrophenol
o-nitrophenol - CO + product(s)
o-nitrophenol - HONO + product(s)
o-nitrophenol - NO + product(s)
o-nitrophenol - OH + product(s)
o-nitrophenol - other product(s)
NO2 + hn - NO + O
3.0 ꢀ 10ꢁ11
1.6 ꢀ 106
(0–7) ꢀ 10ꢁ12
9.0 ꢀ 10ꢁ12
8.0 ꢀ 10ꢁ14
8.0 ꢀ 10ꢁ14
1.65 ꢀ 1ꢁ0ꢁ4 3
5.2 ꢀ 10ꢁ3
2.1 ꢀ 10ꢁ3
7.3 ꢀ 10ꢁ3
1.9 ꢀ 10ꢁ3
1.2 ꢀ 10
Estimated from Table 5 in ref. 25
This work, fit to the measured build-up of o-nitrophenol
Assumed to be equal to k of previous reaction, see text
This work, fit to the measured CO yield
This work, fit to the measured HONO yield
This work, fit to the measured NO yield
This work, fit to the measured HNO3 yield, see text
This work, fit to the measured total loss of o-nitrophenol
This work, fit to the NO2 loss rate in NO2 photolysis experiment
Atkinson et al.26 (2004)
13
14
15
16
18
7
8
12
9
O + NO2 ( + M) - NO3 ( + M)
O + NO2 - NO + O2
O + O2 (+M) - O3 (+M)
O + I2 - IO + I
2.1 ꢀ 10ꢁ12
1.0 ꢀ 10ꢁ11
1.4 ꢀ 10ꢁ14
1.4 ꢀ 10ꢁ10
3.5 ꢀ 10ꢁ17
2.6 ꢀ 10ꢁ11
1.8 ꢀ 10ꢁ14
1.47 ꢀ 1ꢁ0ꢁ2 12
3.5 ꢀ 10ꢁ2
3.6 ꢀ 10ꢁ3
6.0 ꢀ 10ꢁ5
4.8 ꢀ 10
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
30
10
Sander et al.27 (2006)
NO2 + O3 - NO3 + O2
Atkinson et al.26 (2004)
NO + NO3 - NO2 + NO2
NO + O3 - NO2 + O2
NO2 + NO3 (+M) - N2O5 (+M)
Atkinson et al.26 (2004)
11
6
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
ꢁ6 N2O5 (+M) - NO3 + NO2 (+M)
Atkinson et al.26 (2004)
N2O5 - product(s)
This work, fit to the measured N2O5 concentration
This work, fit to the formation of HNO3
This work, fit to the HONO formation in NO2 photolysis
Jenkin et al.28 (1990)
N2O5 - HNO3 + HNO3
NO2 - HONO
I + I (+M) - I2 (+M)
I + NO (+M) - INO (+M)
I + NO2 (+M) - INO2 (+M)
I + NO3 - IO + NO2
I + O3 - IO + O2
2.9 ꢀ 10ꢁ13
3.7 ꢀ 10ꢁ13
5.0 ꢀ 10ꢁ12
4.5 ꢀ 10ꢁ10
1.2 ꢀ 10ꢁ12
2.6 ꢀ 10ꢁ10
8.3 ꢀ 10ꢁ11
(0–6) ꢀ 10ꢁ11
2.0 ꢀ 10ꢁ11
3.3 ꢀ 10ꢁ12
3.9 ꢀ 100
Sander et al.27 (2006)
Sander et al.27 (2006)
Chambers et al.29 (1992)
Sander et al.27 (2006)
I + INO - I2 + NO
van den Bergh and Troe30 (1976)
25
33
I + INO2 - I2 + NO2
I + IONO2 - I2 + NO3
van den Bergh and Troe30 (1976)
Assumed, variable, see text
IO + NO - I + NO2
Sander et al.27 (2006)
31
IO + NO2 (+M) - IONO2 (+M)
INO2 - I + NO2
Sander et al.27 (2006)
van den Bergh and Troe30 (1976), Sander et al.27 (2006)
Sander et al.27 (2006), van den Bergh and Troe30 (1976)
Photolysis products and F yet unknown, see text
Chambers et al.29 (1992)
26
32
INO2 + INO2 - I2 + NO2 + NO2
IONO2 + hn - I + NO3
NO3 + I2 - IONO2 + I
4.7 ꢀ 10ꢁ15
(0–5.1) ꢀ 10ꢁ2
1.5 ꢀ 10ꢁ12
(0–5) ꢀ 10ꢁ10
4.8 ꢀ 10ꢁ12
1.1 ꢀ 10ꢁ11
37
C6H5O2 + I2 - C6H5O + IO + I
OH + NO (+M) - HONO (+M)
OH + NO2 (+M) - HNO3 (+M)
Variable, see text
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
Fit to C6H5NO2 yield, see text
C6H5NO + C6H5O2 - C6H5NO2 + C6H5O 7.5 ꢀ 10ꢁ12
C6H5O + C6H5O - C12H10O2
C12H10O2 + hn - 2 C6H5O
2 C6H5O2 - 2 C6H5O + O2
C5H5 + O2 - products
1.2 ꢀ 10ꢁ11
1.3 ꢀ 10ꢁ2
Berho and Lesclaux31
Based on UV absorption cross section from ref. 31
Estimate used by Caralp et al.32
Arbitrary value
5.0 ꢀ 10ꢁ12
5 ꢀ 10ꢁ12
29
38
a
C6H5O + O3 - C6H5O2 + O2
( Z ) 2.9 ꢀ 10ꢁ13a
Tao and Li33 (1999), value may be a lower limit
Stated products were not specified in the reference but were assumed for this table.
study details of the reaction mechanisms and to support the
qualitative conclusions drawn from the experimental results.
The program package LARKIN24 was used to numerically
integrate the kinetic differential equations. The mechanism of
Table 1 is quite complex and contains numerous uncertain rate
constants. Nevertheless, useful conclusions can be drawn from
these numerical calculations. Identical rate constants (includ-
ing the wall loss rate constants) were used to model the
following photolysis systems (all at 1 bar):
ꢈ NO2–O2–N2;
ꢈ o-nitrophenol–O2–NO2–N2;
iodoethane–O2–NO2–N2 (reactions related to C2H5
ꢈ
chemistry are omitted from Table 1);
ꢈ iodobenzene–O2–NO2–N2;
ꢈ nitrosobenzene–O2–NO2–N2.
Details of the mechanism in Table 1 and results from these
computer simulations are discussed in the respective Results
and discussion sections below.
ꢃc
This journal is the Owner Societies 2007
5038 | Phys. Chem. Chem. Phys., 2007, 9, 5036–5051

View Article Online
the iodobenzene–O2–NO2 and nitrosobenzene–O2–NO2
photolysis systems. In order to take photolysis products of
o-nitrophenol into account, photolysis of o-nitrophenol under
various reaction conditions (with/without O2 present, with/
without NO2 present) was also investigated. Finally, an
iodoethane–NO2–O2–N2 mixture was photolysed in order to
examine if a special chemistry is effected when I2–NO2 mix-
tures are irradiated with 254 nm light.
IR spectra of iodobenzene, nitrosobenzene, o-nitrophenol,
nitrobenzene, and N2O5
Reference spectra of iodobenzene, nitrosobenzene, o-nitrophe-
nol, nitrobenzene, and N2O5 are shown in Fig. 1. Maxima of
major absorption bands (of Q branches if present) are at the
following wavenumbers (in units of cmꢁ1): iodobenzene 685.3,
730.9, 1021.5, 1066.2, 1476.3, 1582.4, 3080.1; nitrosobenzene
681.6, 750.7, 1112.7, 1454.6, 1521.6, 3080.4; o-nitrophenol
690.1, 748.0, 871.9, 1341.7, 1478.2, 1626.7; nitrobenzene
701.5, 792.5, 851.7, 1355.5, 1548.6 (minimum), 3084; N2O5
743.4, 1245.6, 1716.8 cmꢁ1
.
IR absorption cross sections of the products N2O5 and
o-nitrophenol are presented in Table 2. The absorption cross
section of N2O5 at 1245.6 cmꢁ1 from this work agrees well
with literature data.
Photolysis of NO2 at 253.7 nm
Although the absorption coefficient of NO2 at 253.7 nm is very
small (2.0 ꢀ 10ꢁ20 cm2 moleculeꢁ1 27), there was considerable
photolysis of NO2 in experiments with high initial concentra-
tions of NO2. This increased photolysis rate of NO2 probably
originates from additional weaker emission lines of the TUV
lamps between 300 and 600 nm36 (strongest line at 253.7 nm).
Since the main interest of this work is NO3 formation in
reaction (5), care was taken to identify other sources of
NO3. Two possible sources are reactions (8) and (10) that
can take place after O atom formation in the photolysis of
NO2:
Fig. 1 Reference spectra of iodobenzene (6.15 ꢀ 1013 molecule
cmꢁ3), nitrosobenzene (2.9 ꢀ 1014 molecule cmꢁ3), o-nitrophenol
(5.0 ꢀ 1013 molecule cmꢁ3), nitrobenzene (2.5 ꢀ 1014 molecule
cmꢁ3) and N2O5 (1.42 ꢀ 1014 molecule cmꢁ3) in the 650–4000 cmꢁ1
range; spectral resolution 1 cmꢁ1; optical pathlength 29.0 m; T =
298 K; ptot = 1 bar (N2).
Results and discussion
NO2 þ hn ! NO þ O
O þ NO2 þ M ! NO3 þ M
O þ O2 þ M ! O3 þ M
NO2 þ O3 ! NO3 þ O2
NO þ O3 ! NO2 þ O2
ð7Þ
ð8Þ
The main idea of the present work is to identify the product
NO3 in the reaction of phenylperoxy radicals with NO2 by
scavenging NO3 with NO2. To make the scavenging process
effective, rather high initial NO2 concentrations were applied.
Photolysis of NO2 turned out to influence the reaction me-
chanism. For this reason, photolysis of an NO2–O2–N2 mix-
ture was studied first, using the measured product yields of
NO, O3, N2O5, HONO, and HNO3 to test the mechanism and
the rate constants of the H/N/O subsystem used in the
simulation of the phenylperoxy–NO2 reaction systems. In
addition, o-nitrophenol proved to be a major product in both
ð9Þ
ð10Þ
ð11Þ
Although NO and O3 formed in reactions (7) and (9) largely
neutralize each other by reaction (11), significant amounts of
Table 2 IR peak absorption coefficients of N2O5 and o-nitrophenol
n/cmꢁ1
smax (n)/10ꢁ18 cm2 moleculeꢁ1
Ref.
Temperature/K
Spectral resolution/cmꢁ1
N2O5
1245.6
1246
1245
1479
748.0
1.78 ꢆ 0.09
1.81
1.75 ꢆ 0.01
0.82 ꢆ 0.16a
1.3 ꢆ 0.2a
This work
34
35
This work
This work
298
298
298
298
298
1
1
2
1
1
o-nitrophenol
a
Error margins are quite large (about ꢆ20%) since the leak rate of the reaction chamber had to be taken into account, see experimental section.
ꢃc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 5036–5051 | 5039

View Article Online
Table 3 Product distributions of the photolysis of NO2 at 253.7 nm, T = 298.1 K, pO = 200 mbar, ptot = 1000 mbar, M = O2 + N2; t = total
2
photolysis time; concentrations are in units of molecule cmꢁ3. Percental product yields at time t are relative to the amount of NO2 lost
(= –D[NO2])
[NO2]0
(1013
ꢁD[NO2]
[N2O5]
(1013
[N2O5]
(%)
[HNO3]b
(%)
[NO]
(%)
[HONO]c
(%)
[Nprod]/(–D[NO2])d
(%)
[O3]
(%)
No.
t/s
)
(1013
)
)
1
a
b
c
d
e
a
0
32
98
164
230
296
80.57
—
1.94
4.25
6.25
8.39
10.57
—
—
—
—
—
—
104
103
101
101
101
—
0.231
0.271
0.160
0.115
0.099
11.9
6.4
2.6
1.4
0.9
2.5
1.8
1.5
1.3
1.4
71.8
82.9
88.1
91.0
90.9
6.1
5.8
6.3
6.4
6.6
13.0
3.4
1.7
0.9
0.4
Continuous photolysis was applied for 328 s; 64 scans were averaged to produce one spectrum. Another 15 spectra (not shown in this table) were
b
taken after switching off the lights at t = 330 s. IR absorption cross sections are based on the integrated intensity of the 3552 cmꢁ1 band from
Chackerian et al.37,38 c IR absorption cross sections were taken from Barney et al.39 d [Nprod] ꢇ 2 ꢀ [N2O5] + [HNO3] + [NO] + [HONO].
NO3 may result from reactions (8) and (10). For this reason,
an additional experiment was performed photolysing an
NO2–O2–N2 mixture in order to: (i) determine the true photo-
lysis rate constant of NO2 for the conditions of the present
experiments; (ii) test, by comparing measured and calculated
concentrations of NO, NO2, N2O5, HONO, HNO3, and O3, if
the NO3 chemistry and wall reactions of NO2, N2O5 and
HNO3 are properly taken into account in the model calcula-
tions.
geneous sources are probably responsible for the formation of
both HONO and HNO3. HONO is well-known to form on the
surface of reaction chambers made of quartz glass in the
presence of gaseous NO2.40 In the present reaction chamber,
the HONO yield from NO2 is quite constant (6.2 ꢆ 0.4% of
the consumed NO2). HNO3 is known to be readily formed
from N2O5 on chamber walls as well as on aerosol surfaces
containing adsorbed H2O. In the present work, the HNO3 loss
rate strongly depends on the preceding chemical processes in
the chamber and may be very slow. Due to the transformation
of N2O5 to HNO3 on the reactor walls, the surface of the
chamber may even be a source of HNO3.
The product distribution of the NO2 photolysis experiment
in 1 bar of synthetic air is shown in Table 3. NO is the main
photolysis product, and HONO, HNO3, N2O5, and O3 are
minor products. O3 and N2O5 accumulate only in the early
stage of the photolysis (the first tens of seconds) when the
absolute NO concentration is still moderately low and the O3
concentration is still moderately high. Most of the N2O5 is
formed in this early stage via reaction (10). This mechanism
also generates the small yield of N2O5 that was observed in the
photolysis of C2H5I–NO2–O2–N2 mixtures (see below). NO3
formation from this source also has to be considered in the
photolysis of iodobenzene and nitrosobenzene, in particular,
when the initial NO2 concentration is high. However, the
model calculations show that the much higher yields of
N2O5 in the C6H5I and C6H5NO photolysis systems cannot
be explained by the reactions of NO2 with O3 and O atoms.
There are two peculiar features of this reaction system (NO2
photolysis in synthetic air at a total pressure of 1 bar):
ꢈ The effective first order loss rate constant of NO2 is only
about 25% of the NO2 photolysis rate constant, due to
efficient reformation of NO2 in consecutive reactions.
ꢈ A minor part of the O atoms are not scavenged by O2 to
form ozone but they react with NO2:
Since the intensity of the photolysis lamps slightly increases
during the first 10 s, the NO2 photolysis rate constant and
several wall loss rate constants were estimated by fitting the
calculated to the measured concentrations of NO2, NO, N2O5,
HONO, HNO3, and O3 at t = 32 s (30 s is the standard
photolysis time for o-nitrophenol, iodoethane, and iodoben-
zene in the present work). Experimental concentrations are
well reproduced by the mechanism and rate constants of Table
1. In particular, [N2O5] passes a maximum between 32 and 98 s
both in the kinetic simulation as well as in the experiment.
Photolysis of o-nitrophenol at 253.7 nm
It turned out that o-nitrophenol is a major product of the
photolysis of both iodobenzene and nitrosobenzene in the
presence of O2 and NO2 and partially photolysed during the
course of these reactions. For this reason, the photolysis rate
constant and the photolysis products of o-nitrophenol were
determined under the conditions of the iodobenzene and
nitrosobenzene photolysis experiments in order to take these
products properly into account. Product yields are summar-
ized in Table 4. The following results may be deduced from
Table 4:
O þ NO2 ! NO þ O2
ð12Þ
Reaction (12) increases the NO and reduces the O3 yield
because it forms additional NO and, at the same time, prevents
the O atoms from generating ozone.
(i) N2O5 could not be detected as a product in any of these
experiments.
In photoreactors simulating air chemistry, HONO and
HNO3 are commonly formed by addition of OH radicals to
NO and NO2. In the present NO2 photolysis system, there
must be other sources for HONO and/or HNO3 since the
ratios of the yields of HONO and HNO3 are not compatible
with the NO and NO2 concentrations and the rate constants
for the reactions of these species with OH. Actually, hetero-
(ii) The carbon mass balance of the identified products is
very bad.
(iii) Significant amounts of CO were detected, both with and
without NO2 present.
(iv) HNO3 was a major product in the photolysis of
o-nitrophenol in synthetic air in the presence of NO2, the yield
of HNO3 increasing with the initial amount of NO2.
ꢃc
This journal is the Owner Societies 2007
5040 | Phys. Chem. Chem. Phys., 2007, 9, 5036–5051

View Article Online
Table 4 Summary of experimental data on the photolysis of o-nitrophenol at 253.7 nm, t = total photolysis time; concentrations are given in
units of molecule cmꢁ3. Product yields (in %) are relative to the amount of converted o-nitrophenol
pO
/
[o-Nitrophenol] –D[o-Nitrophenol] [NO2] –D[NO2]a
[N2O5] [HNO3]
(1013
[NO] [HONO] [Nprod]/
(%)
[CO]
[Nreact]b (%) (%)
2
No. t/s mbar (1013
)
(1013
)
(1013
)
(1013
)
)
(%)
(%)
1
2
3
a
0
30
0
30
0
200
200
200
14.54
0
5.57
0
6.49
0
6.06
0
0
0
0
3.8
0
78.1
0
57.2
0
0
0
0
3.6
0
0
7.9
0
48.1
0
47.0
0
13.4
0
19.2
0
12.6
8.97
18.05
11.56
20.11
14.05
0.36
65.31
55.13
14.32
10.17
ꢁ0.36
0
10.18
0
4.15
o0.1
0
o0.1
40.5 5.0
0
16.6 5.2
0
0
26
o0.1
b
–D[NO2] ꢇ [NO2]0 ꢁ [NO2]t. [Nprod] ꢇ 2 ꢀ [N2O5] + [HNO3] + [NO] + [HONO], [Nreact] ꢇ –(D[o-nitrophenol] + D[NO2]).
(v) There are considerable yields of NO when NO2 is
into account by adding reaction (18) to the mechanism
in Table 1:
initially present. In run 3 of Table 4, only about 15% of this
yield can be explained by NO2 photolysis, according to the
model calculations based on the kinetic data in Table 1,
whereas the surplus must be related to the photolysis of
o-nitrophenol.
o-nitrophenol ! other productðsÞ
ð18Þ
Product spectra and product yields in the photolysis of
C2H5I–O2–NO2–N2 mixtures
(vi) Several % of HONO were detected under all conditions,
also in the absence of initial NO2, thus indicating the existence
of another source of HONO in addition to the well-known
heterogeneous generation from NO2. This observation is in
line with the (much smaller) gas phase source of HONO
reported by Bejan et al.41 for the photolysis of o-nitrophenol
at wavelengths between 300 and 500 nm.
Photolysis of organic iodides is not a very common source of
radicals in stationary experiments. Since abstraction of H
atoms, even of aldehydic H atoms, from organic molecules
by I atoms is very slow, I atoms are essentially inert in those
systems (in the absence of moderate or high O3 concentra-
tions), and predominantly accumulate as I2. Barnes et al.,42
however, observed high yields of N2O5 in the 300–600 nm
photolysis of mixtures of large concentrations of I2 and NO2 in
N2. These authors suggested that NO3 is formed via the
intermediate IONO2. In order to explore the possible NO3
formation from I2 and NO2 in our iodobenzene–NO2–O2–N2
photolysis system, C2H5I was photolyzed as a simple model
compound at initial conditions very similar to the iodobenzene
photolysis experiments (see Table 5 below) except that iodo-
benzene was replaced by iodoethane. Reaction conditions in
(vii) Aerosol bands resembling those in the photolysis of
iodobenzene and nitrosobenzene (see insets of Fig. 2e and 3e)
appeared, indicating the formation of a product with a very
low vapour pressure, possibly of p-nitrophenol.
In the reaction mechanism of Table 1, formation of the
observed photolysis products CO, HONO, and NO is repre-
sented by the first order reactions (13), (14) and (15):
o-nitrophenol ! CO þ productðsÞ
o-nitrophenol ! HONO þ productðsÞ
o-nitrophenol ! NO þ productðsÞ
ð13Þ
ð14Þ
ð15Þ
this experiment were: [C2H5I]0 = 2.46 ꢀ 1014 molecule cmꢁ3
,
[NO2]0 = 1.54 ꢀ 1014 molecule cmꢁ3, p(O2) = 200 mbar,
p(N2) = 800 mbar, –D[C2H5I] = 3.63 ꢀ 1013 molecule cmꢁ3
,
24 low pressure Hg lamps, photolysis time = 30 s. After
switching off the lights, the product spectrum was monitored,
averaging 128 scans. Ethyl peroxynitrate was the only obser-
vable major product which is formed according to:
A possible way to explain the high yields of HNO3 in the
presence of an excess of NO2 is the formation of OH radicals
by C–O bond fission in the photolysis of o-nitrophenol,
followed by the association of OH with NO2. Thus, a first
order rate constant for the formation of OH from o-nitro-
phenol was also included in Table 1:
C2H5I þ hn ! C2H5 þ I
C2H5 þ O2ðþMÞ ! C2H5O2ðþMÞ
ð19Þ
ð20Þ
ð21Þ
C2H5O2 þ NO2ðþMÞ ! C2H5O2NO2ðþMÞ
o-nitrophenol ! OH þ productðsÞ
ð16Þ
C2H5O2NO2 was identified by the positions of its IR absorp-
tion bands centred at 794.0, 1297.4, and 1718.7 cmꢁ1 which
agree within 0.4 cmꢁ1 with the band positions of C2H5O2NO2
reported by Niki et al.43 The observed very low N2O5 yield of
1 ꢀ 1012 molecule cm3 corresponds to the amount of N2O5
that is expected to result from NO2 photolysis (see earlier
section on NO2 photolysis). There was no evidence for the
formation of aerosols in this system or for any peculiar
chemistry related to the simultaneous presence of iodine
and NOx.
However, the energy of 254 nm photons is only a little larger
than the C–O bond energy in phenol. More likely, HO2
radicals are formed during the degradation of o-nitrophenol,
generating OH radicals via reaction (17):
HO2 þ NO ! OH þ NO2
ð17Þ
The total first order loss rate of o-nitrophenol was larger
than the sum of reactions (13)–(16). This fact was taken
ꢃc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 5036–5051 | 5041

View Article Online
Product spectra and product yields in the photolysis of
C6H5I–O2–NO2–N2 mixtures
A product spectrum of the 254 nm photolysis of iodobenzene–
O2–NO2–N2 mixtures at 298 K is shown in Fig. 2a–e. Different
from the photolysis of iodoethane under similar reaction
conditions (see previous section), typical strong absorption
bands close to 800, 1300, and 1740 ꢆ 20 cmꢁ1 which are
specific to all known peroxynitrates are missing. Thus, there is
no indication for the formation of a peroxynitrate according
to
C6H5I þ hn ! I þ C6H5 ðexcitation unspecifiedÞ
ð22Þ
C6H5I þ hn ! I þ C6H5 ðthermally distributedÞ ð22aÞ
C6H5 þ O2 ðþMÞ ! C6H5O2 ðþMÞ
ð23Þ
ð24Þ
C6H5O2 þ NO2 ðþMÞ ! C6H5O2NO2 ðþMÞ
Reaction (22a) is introduced in order to distinguish essentially
thermally distributed C6H5 radicals from those carrying im-
portant amounts of excitation energy, thus potentially leading
to other products (see below). N2O5 was identified as a major
product, in addition to o-nitrophenol (Fig. 2b–d). Fig. 2e
depicts the unidentified residual absorption after spectral
subtraction of N2O5. In the inset of Fig. 2e, a weak broad
band absorption between 3000 and 3500 cmꢁ1 indicates the
presence of an aerosol which may be related to p-nitrophenol
(see later).
The measured product yields are summarized in Table 5. In
run 1, spectra were monitored in dark periods suspending the
photolysis in order to take the product spectra at better signal-
to-noise ratios (128 rather than 10 scans per spectrum as in the
time resolved experiments). In runs 2–5, product spectra based
on 128 accumulated scans were taken after the noted photo-
lysis periods. In all of these experiments, heterogeneous reac-
tions on the chamber walls were strongly reduced during the
dark periods by keeping the fans arrested. In runs 6 and 7,
spectra were continuously monitored during photolysis, aver-
aging 10 scans per spectrum. The given times indicate the
middle of the periods accumulating 10 scans to produce one
spectrum (dead time per spectrum: 2 s). For clarity, the
average product yields obtained for total photolysis times of
E30 s from the data in Table 5 (runs 1c, 2–5, 6c) are collected
in Table 6.
Fig. 2 Product spectra in the C6H5I–O2–NO2–N2 photolysis system
at 298 K after 30 s photolysis (run 3 from Table 5); [C6H5I]0 = 2.43 ꢀ
1014 molecule cmꢁ3, [NO2]0 = 1.48 ꢀ 1014 molecule cmꢁ3, [O2]0
=
4.86 ꢀ 1018 molecule cmꢁ3 (200 mbar), ptot = 1000 mbar, 128 scans.
(a) Reaction mixture before irradiation; (b) product spectrum minus
residual absorption of the reactant iodobenzene; (c) spectrum (b)
minus absorptions from NO2, HNO3, HONO, CO and CO2 (please
note that the y axis is enlarged by a factor of 2 as compared to (b)); (d)
spectrum (c) minus absorption from o-nitrophenol; (e) spectrum (d)
minus absorption from N2O5, inset: absorption from aerosol in the
3000–3500 cmꢁ1 region.
Several conclusions may be drawn from the product spectra
in Fig. 2 and the experimental data listed in Tables 5 and 6:
(i) The carbon mass balance is very poor (about 30%),
o-nitrophenol being the major carbon containing product. It
may be noted that comparable yields of o- and p-nitrophenol
are obtained when nitrophenol is synthesized from benzene
and nitric acid in the liquid phase. The same may happen in
the gaseous phase. Here, however, p-nitrophenol could remain
undetected due to its extremely low vapour pressure (ca. 4 ꢀ
absorption bands at 748.0, 1200, and 1478.2 cmꢁ1. A very
similar observation was made by Niki et al.9 in their paper on
the FT-IR product spectra of the photolysis of mixtures
containing ppm quantities of Cl2, benzaldehyde, and NO in
purified air. Niki et al. showed that the degradation of
benzaldehyde in the presence of NO leads to phenyl radicals
as was later confirmed by Caralp et al.32 They identified
o-nitrophenol as a major photolysis product that rapidly
decayed during aging the reaction mixture for about 10 min
in the dark. They also showed that the residual of their
photolysis product spectrum, including the most prominent
absorption band at 1343 cmꢁ1, was very similar to that of a
thin film of deposited p-nitrophenol, and they tentatively
18
10ꢁ5 mbar at 25 1C44 as compared to about 0.25 mbar for
o-nitrophenol). In fact, only about one half of the second
largest IR absorption band of o-nitrophenol centred at 1342
cmꢁ1 (see Fig. 2) disappears when the absorption of o-nitro-
phenol is subtracted from the product spectra, based on its
ꢃc
This journal is the Owner Societies 2007
5042 | Phys. Chem. Chem. Phys., 2007, 9, 5036–5051

View Article Online
Table 5 Summary of experimental data on the photolysis of iodobenzene at 298.1 K;abc t = total photolysis time, concentrations are given in
units of molecule cmꢁ3. Product yields (in %) refer to the total amount of converted iodobenzene. (Note: NO has been added before irradiation in
run 7.)
[o-Nitro-
phenol]
(%)
[Nprod]/
d
pO
/
[C6H5I]0 ꢁD[PhI]
[NO2]0 ꢁD[NO2] [NO]
[N2O5] [N2O5] [HNO3] [HONO] [Nreact
(1013
]
[CO]
(%)
2
Run
t/s
mbar (1013
)
(1013
)
(1013
)
(%)
(1013
)
)
(%)
(%)
(%)
(%)
1
a
b
c
d
2
0
10
20
30
40
0
30
0
30
0
200
25.28
0
15.49
0
0
0
0
0
0
0.7
14.0
25.3
32.8
0
35.5
0
21.5
0
20.8
0
15.7
0
7.2
4.6
10.3
0
r8.5
3.7
3.8
0
0
0
1.52
3.14
4.75
6.34
0
5.63
0
4.89
0
4.67
0
4.78
0
1.81
3.93
5.80
0
0.82
2.71
4.47
182.2
181.2
171.9
161.9
0
150.5
0
187.7
0
203.6
0
177.4
0
210.7
204.5
196.2
0
56.1
32.8
32.9
r0.12
r0.12
r0.12
r0.12
0
r0.12
0
r0.12
0
r0.12
0
0.98
0
0.13
r0.12
r0.12
5.42
4.71
3.45
2.22
34.3
30.7
28.3
28.5
0
25.2
0
27.4
0
24.9
0
32.3
0
30.6
25.9
25.6
0
17.1
29.2
30.9
0.75
1.20
1.20
1.02
0
1.08
0
1.90
0
2.28
0
1.49
0
1.04
2.01
2.69
0
r0.25
r0.25
0.28
49.3
38.2
25.2
16.1
0
19.2
0
38.8
0
48.9
0
31.3
0
57.4
52.3
46.4
0
6.6
5.4
5.9
5.5
0
8.0
0
7.2
0
7.5
0
6.8
0
76.2
65.8
58.7
55.1
0
62.8
0
71.1
0
69.1
0
61.2
0
85.1
88.4
67.1
0
—
—
14.1
15.6
16.5
17.4
0
20.4
0
16.8
0
17.7
0
17.6
0
14.5
15.4
16.1
0
18.3
17.3
17.7
1000
200
200
200
200
25.01
24.33
24.60
25.30
24.15
11.52
14.81
44.15
128.7
15.14
3
4
5
30
0
30
0
6
a
b
c
7
a
b
c
5
r11
17
29
0
5
17
29
r5.1
r3.4
0
200
24.87
16.17
r30
19.5
13.3
14.5
r9.2
6.3
—
a
Runs are arranged in chronological order. The reaction chamber was completely set apart, cleaned and set together between runs 2 and 3. This may explain the
b
significantly larger yield of N2O5 at the expense of HNO3 in the later experiments. 1–2.5 % CO2 is formed in all experiments but not included in Table 5. The CO2
c
yield is 2–2.5 % in runs 1 and 2 and 1–2% in runs 3–7. This fact points to the heterogeneous origin of the small CO2 yield (see previous note). In run 6, 20 consecutive
d
spectra were taken after switching off the lights. [Nprod] ꢇ [NO] + [o-nitrophenol] + 2 ꢀ [N2O5] + [HNO3] + [HONO], [Nreact] ꢇ ꢁD[NO2].
assigned their residual gas phase spectrum to airborne
crystalline p-nitrophenol. Since the residual spectra in our
iodobenzene photolysis systems look very much like the
residual spectra of Niki et al., the prominent band in our
residual spectrum at 1343 cmꢁ1 (Fig. 2e) as well as the broad
band absorption between 3000 and 3500 cmꢁ1 and additional
absorption bands at 1113, 1169, 1293, and 1517 cmꢁ1 may also
originate from an aerosol consisting of solid p-nitrophenol. In
the dark, o-nitrophenol decayed very slowly whereas the
abovementioned bands tentatively assigned to p-nitrophenol
decayed much faster, exhibiting about the same time beha-
viour. This aerosol, together with o- and p-nitrophenol depos-
ited on the reactor walls, may be responsible for most of the
bad carbon mass balance and, at the same time, for the
deficiency in the nitrogen balance (see Table 6). It covers most
of the absorption in the residual IR spectra. For our purpose,
it is important to note that there are no major unidentified
bands characteristic of NOx groups in the residual spectra.
(ii) The nitrogen balance (second last column in Table 5) is
better than the carbon balance, the major part of the missing
nitrogen possibly being buried in the aerosol and in material
condensed or adsorbed on the chamber walls (e.g. p-nitrophe-
nol, see (i) above, and HNO3). The loss of NO2 approaches
200% of the converted iodobenzene at very short reaction
times. This strong NO2 loss probably reflects the large im-
portance of reaction (5), each NO3 molecule formed consum-
ing another NO2 molecule to generate N2O5.
(iii) There is a major (CO) and a minor (CO2) C1 product
showing up in the product spectrum. Making allowance for
the corresponding C5 or C components which have not yet
o5
been identified would help to improve the carbon balance.
(iv) The yield of o-nitrophenol is nearly independent of the
initial concentration of NO2 (see runs 3–5 in Table 5).
Reaction mechanism in the iodobenzene photolysis system based
on reaction (5)
Table 6 Photolysis of C6H5I–O2–NO2–N2 mixtures — average pro-
duct yields from 6 experiments at 298.1 K in 1 bar of N2 + O2 after
about 30 s photolysis and conversions of about 20% (runs 1c, 2–5, and
6c from Table 5); initial concentrations: see Table 5
Alkyl iodides are known to photolyse at 254 nm with a
quantum yield of 1,45 generating alkyl radicals and iodine
atoms. In the present work, rate constants of the 254 nm
photolysis of iodoethane (5.0 ꢀ 10ꢁ3
s
ꢁ1) and iodobenzene
Product(s)
Molar yield (%)
(6.2 ꢀ 10ꢁ3
s
ꢁ1) were measured under identical reaction
o-Nitrophenol
CO
CO2
27 ꢆ 3
18 ꢆ 2
1.8 ꢆ 0.7
19–49
conditions. The photolysis rate constant ratio (0.81) proved
to be close to the ratio of their absorption cross sections at
253.7 nm (= 0.66, based on s = 1.22 ꢀ 10ꢁ18 cm2 moleculeꢁ1
for C2H5I27 and s = 1.86 ꢀ 10ꢁ18 cm2 moleculeꢁ1 for
C6H5I46,47), suggesting that the total quantum yield for I atom
formation from iodobenzene is similar to that of iodoethane,
N2O5
HNO3
N2O5 + HNO3
10–36
46 ꢆ 10
1
2
[Nprod]/(–DNO2)
65 ꢆ 6
ꢃc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 5036–5051 | 5043

View Article Online
i.e. E1. Iodine atoms will probably be converted to molecular
iodine via INO2:
ꢈ A major fraction of the phenoxy radicals might add NO2
to form phenyl nitrate:
C6H5O þ NO2ðþMÞ ! C6H5ONO2ðþMÞ
ð27Þ
I þ INO2 ! I2 þ NO2
ð25Þ
Recently, Platz et al.50 generated phenoxy radicals in the
presence of NO2 and monitored the decay of phenoxy by
following its UV absorption at 235 nm on a time scale of tens
of ms. They observed a residual absorption that was tentatively
assigned to phenyl nitrate. However, we are not aware of any
other information about this compound in the literature. Our
DFT calculations on the vibrational frequencies of phenyl
nitrate do not support its formation in our reaction system: the
by far strongest IR absorption bands of phenyl nitrate which
we obtained from B3LYP/6-31G(d) calculations using the
INO2 þ INO2 ! I2 þ NO2 þ NO2
ð26Þ
Photolysis of molecular iodine is not an effective source of
iodine atoms since the absorption coefficient is below 1 ꢀ
10ꢁ18 cm2 moleculeꢁ1 throughout the wavelength region
250–470 nm.48 At 298 K, phenyl radicals rapidly add O2 to
14
form the thermally stable phenylperoxy radicals (phenyl–OO
bond energy E177 kJ molꢁ1 11). In the presence of NO2 and
the absence of NO, phenylperoxy radicals are expected to add
NO2 to form a fairly stable peroxynitrate. However, in the
present reaction system there is no evidence for the typical
peroxynitrate absorption bands close to 800, 1300, and 1740 ꢆ
20 cmꢁ1, as was already stated previously by Kirchner.5 On
the other hand, early thermochemical estimates16 as well as
ab initio calculations from this work (see Table 10 and Fig. 4)
suggest that the O–O bond is considerably less stable than the
O–N bond in phenyl peroxynitrate. The detection of N2O5
(and of high HNO3 yields, probably resulting from the reac-
tion of additional N2O5 on the chamber walls) is in line with
these thermochemical considerations.
Gaussian 98 program package19 (812, 1369, and 1794 cmꢁ1
,
all unscaled) are missing in the residual product spectrum
(Fig. 2e), thus indicating that the main part of the residual
absorption is not from phenyl nitrate. In addition, B3LYP/
6-31G(d) calculations result in a reaction enthalpy of ꢁ59 kJ
molꢁ1 for the addition of NO2 to phenoxy. Assuming a
negligible barrier for the association reaction and a preexpo-
nential factor of 1015 sꢁ1 for the dissociation reaction, a
thermal lifetime of phenyl nitrate at 298 K of tens of micro-
seconds results that is consistent with the characteristic time of
approach to equilibrium measured by Platz et al.50 for the
decay of phenoxy radicals in the presence of NO2. Thus,
phenyl nitrate may be present in equilibrium with phenoxy
and NO2 on the time scale of tens of microseconds in the
experiments of Platz et al. but unstable on our time scale of
tens of seconds.
In the equilibrium
!
NO þ NO þ M N O þ M
ð6;
3
2
2
5
-6Þ only about 0.1% of the NO3 either present as free NO3 or
masked as N2O5 at our typical reaction conditions (298 K,
[NO2] E 2 ꢀ 1014 molecule cmꢁ3, [N2O5] E (1–2) ꢀ 1013
molecule cmꢁ3) is free NO3 and can thus react with other
ꢈ In addition to the identified product o-nitrophenol,
p-nitrophenol can be formed and may produce solid particles
due to its extremely low vapour pressure. The formation of
p-nitrophenol is supported by the residual IR spectrum (see
previous section).
reactants, e.g. with NO (kNO +NO,
= 2.6 ꢀ 10ꢁ11
298K
3
molecule cmꢁ3 ꢁ1 26), o-nitrophenol (kNO +o-nitrophenol, 298K
s
3
o 2 ꢀ 10ꢁ14 molecule cmꢁ3 sꢁ1 49), I atoms (kNO + I = 4.5 ꢀ
3
10ꢁ10 molecule cmꢁ3
s
ꢁ1 29), and I2 (kNO +I , 298K = 1.5 ꢀ
ꢈ There is no simple explanation for the formation of the
observed by-products CO (major) and CO2 (minor) in this
reaction system (see Table 6). If they are the result of thermal
or photolytic decomposition of C6 compounds, part of the
residual IR absorption (Fig. 2e) and of the missing carbon
may originate from the complementary C5 compound(s). It is
well known that thermal decomposition of phenoxy radicals
leads to cyclopentadienyl radicals and CO at high tempera-
tures:
3
2
10ꢁ12 molecule cmꢁ3 sꢁ1 29). In most of these experiments, the
yield of NO is below its detection limits of 3 ꢀ 1012 molecule
cmꢁ3 for spectra based on 10 scans and 1 ꢀ 1012 molecule
cmꢁ3 for spectra based on 128 scans, and its reaction with NO3
cannot be a significant loss process for NO3. The same is true
for o-nitrophenol due to its slow reaction with NO3.49 The
model calculations based on the kinetic data in Table 1 suggest
that the most important loss processes of NO3 in the iodo-
benzene system are its reaction with I2 and the wall loss of
N2O5.
C6H5O þ M ! C5H5 þ CO þ M
ð28Þ
When the reaction starts, the concentration of N2O5 in-
creases parallel to the decay of iodobenzene, whereas [HNO3]
increases with a delay (see run 1 in Table 5). Later in the
reaction, however, the N2O5 concentration passes a maximum
whereas the HNO3 concentration increases all the time. After
switching off the photolysis lights and with the fans on, N2O5
rapidly disappears within about 30 s whereas the HNO3
concentration slowly decays and approaches a constant value.
This behaviour suggests that N2O5 is transformed to HNO3 on
the time scale of the experiment, probably by reaction with
H2O absorbed on the chamber walls.
At room temperature, this reaction is too slow in a thermal
system due to the high activation energy of 183.5 kJ molꢁ1 11
.
DFT and ab initio calculations on the fate of phenylperoxy
radicals (e.g. by Fadden et al.51 and Tokmakov et al.52)
demonstrate a number of ways in which CO and CO2 could
be formed from phenylperoxy radicals, accompanied by the
formation of cyclopentadienyl radicals. However, the reaction
barriers involved are not reliably known. Nothing is known
about the kinetics of the cyclopentadienyl (C5H5) radical. In
the model calculations, its loss reactions have been taken into
account very fragmentary by the reaction
There are several reaction paths possibly accounting for the
lacking carbon:
C5H5 þ O2 ! products
ð29Þ
ꢃc
This journal is the Owner Societies 2007
5044 | Phys. Chem. Chem. Phys., 2007, 9, 5036–5051

View Article Online
Excess energy in the phenoxy radicals formed from phenylper-
oxy by reaction (5) could help to overcome the energy barrier.
Since the yield of CO in the photolysis of iodobenzene–O2–N2
mixtures (i.e. in the absence of NOx) increases with the partial
pressure of O2,53 the reactions of C6H5, C6H5O, and C6H5O2
radicals with O2 have to be considered as well as initiation
processes for the CO formation.54 This will be discussed in the
section on C6H5NO photolysis below. Finally, photolysis of
the product o-nitrophenol is also a source of CO (see earlier
section on the photolysis of o-nitrophenol). However, this
source can only account for a minor fraction of the CO yields
measured in the photolysis of iodobenzene and nitroso-
benzene.
in the present work. These authors photolysed mixtures of I2,
NO2, and N2 in the 300–600 nm range with the aim of
generating high concentrations of IONO2. In fact, they iden-
tified IONO2 for the first time in the gaseous phase by its IR
spectrum. In their reaction system, IONO2 is formed the
following way:
NO2 þ hn ! NO þ O
O þ I2 ! IO þ I
ð7Þ
ð30Þ
ð31Þ
IO þ NO2 ðþMÞ ! IONO2 ðþMÞ:
Surprisingly, however, Barnes et al.42 detected large yields of
N2O5 which they assigned to either of two mechanisms—
photolysis of IONO2,
ꢈ At 254 nm, photolysis of ozone is very efficient. Since the
electronically excited photolysis products O(1D) and O2(1D)
are relaxed too rapidly to undergo other reactions,26 the
photodissociation of O3 in synthetic air leads back to O3.
For this reason, the photolysis of ozone has not been incor-
porated into the mechanism in Table 1.
IONO2 þ hn ! I þ NO3;
or reaction of IONO2 with I atoms:
I þ IONO2 ! I2 þ NO3:
ð32Þ
ð33Þ
Formation of the identified major product o-nitrophenol
(isomeric to phenyl nitrate) in mixtures of phenylperoxy with
Reactions (30)–(33) are also included in the mechanism of
Table 1. In fact, photolysis of IONO2 may be an efficient
source of NO3 at 253.7 nm since the absorption cross section
of IONO2 is very large (s253.7 nm = 1.24 ꢀ 10ꢁ17 cm2 59). On
the other hand, the photolysis products of IONO2 are not
known at the present time. The channel leading to I + NO3 is
energetically less favourable by about 40 kJ molꢁ1 as com-
pared to the formation of IO and NO2 (see thermochemical
data in Table 11). An upper limit of 5.1 ꢀ 10ꢁ2 sꢁ1 (for a
maximum quantum yield F of 1) was estimated for the
photolysis rate constant of IONO2 under our reaction condi-
tions (24 low pressure mercury lamps). This figure is based on
the assumption that the rate constant ratio for the I atom
formation in the photolysis of iodoethane and IONO2 is equal
to the ratio of their absorption cross sections at 253.7 nm,
5,9
NO2
NO3
as well as in mixtures of phenols with NO2 and
55–57
has been reported previously. Although the precise
mechanism of its formation is not yet known, it is now
generally assumed that o-nitrophenol is generated in the
reaction of phenoxy radicals with NO2.25 Presuming that this
is correct, formation of the observed products N2O5 and
o-nitrophenol can easily be explained by reactions of NO2
with NO3 and phenoxy radicals. Both these latter intermedi-
ates are products of reaction (5), and thus reaction (5) is able
to explain the observed products. The identified N-containing
products (excluding p-nitrophenol) account for about 65% of
the NO2 lost.
In the OH-initiated degradation of aromatic compounds
like benzene, toluene and xylene, glyoxal and/or methylglyoxal
are prominent products. If glyoxal were formed in our system,
NO3 could result from the reaction of glyoxylperoxy radicals
with NO2 as was proposed by Orlando and Tyndall.6 How-
ever, an upper limit of 1 ꢀ 1012 molecule cmꢁ3 (corresponding
to a yield of r2% based on our IR absorption cross section
s(1731 cmꢁ1, base e) E 4.6 ꢀ 10ꢁ19 cm2 moleculeꢁ1) could be
placed on the maximum concentration of glyoxal as measured
by FT-IR. Since, in our reaction system, OH radical concen-
trations are too small to abstract sufficient H atoms from
glyoxal to produce significant yields of glyoxyl radicals, and
since photolysis of glyoxal as a possible source of glyoxyl is
very slow due to the small absorption cross section of glyoxal
combining the literature values sC2H5I,
= 1.22 ꢀ
253.7 nm
10ꢁ18 cm2 moleculeꢁ1 27 and FC2H5I, 253.7 nm = 1 with the
45
photolysis rate constant of iodoethane measured in this work
(khn = 4.7 ꢀ 10ꢁ3
s
ꢁ1). There are several reasons, however,
why the formation of IO via reaction (30) is not effective in the
present work:
ꢈ Formation of O atoms is much less efficient than in the
work of Barnes et al. who photolysed NO2 at 300–600 nm due
to the small absorption coefficient of NO2 at 254 nm. In
addition, most of the O atoms formed are scavenged by O2
at our reaction conditions (200 mbar O2) whereas Barnes et al.
used an oxygen-free system (M = N2).
at the photolysis wavelength (s254
= 1.6 ꢀ 10ꢁ20 cm2
ꢈ Much higher concentrations of the precursors of IO, i.e. of
I2 and NO2, were used in the work of Barnes et al.
(I2: 200-fold, NO2: 25-fold).
nm
moleculeꢁ1 58), formation of glyoxyl from glyoxal and NO3
formation via this route can be excluded.
However, there is another possibility to produce NO3 in our
reaction system involving the intermediate IONO2 (iodine
nitrate). This will be discussed in the next section.
ꢈ Whereas Barnes et al. detected IR absorbances of the
order of 0.02 for both IONO2 (strong bands centred at 813,
1276, and 1673 cmꢁ1) and N2O5, we observed N2O5 absor-
bances of the same magnitude but no indication for any
absorbance from IONO2 (for example, the 813 cmꢁ1 band of
IONO2 is completely missing in Fig. 2e).
Reaction mechanism in the iodobenzene photolysis system
including IONO2 reactions
According to the model calculations, the two mechanisms
suggested by Barnes et al.42 for the formation of N2O5 in the
photolysis of I2–NO2–N2 mixtures via IONO2 contribute very
IONO2 has been suggested by Barnes et al.42 as a possible
effective source of NO3 in a photoreactor similar to that used
ꢃc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 5036–5051 | 5045

View Article Online
little to the N2O5 yields observed in the photolysis of iodo-
benzene in the presence of NO2 and O2. However, in our
iodobenzene system, another mechanism via a phenylperoxy
iodide, though speculative at the present time, could also lead
to effective NO3 formation if phenylperoxy iodide decomposes
extremely fast by either photolysis or thermal decomposition
to form IO radicals:
also consider NO3 formation in the reaction I + IONO2.
Simulation of the present reaction system shows that photo-
lysis of C6H5OOI with a rate constant as high as 5 ꢀ 10ꢁ2 sꢁ1
would not be fast enough to explain the measured N2O5 yields.
For this reason, fast thermal decomposition of C6H5OOI via
reaction (36) (or decomposition of chemically activated
C6H5OOI formed in reaction (34)) would be necessary to
explain the detected N2O5 yield if reaction (5) were not
effective (see Table 7).
C6H5OO þ I2 ! C6H5OOI þ I
C6H5OO þ I ðþMÞ ! C6H5OOI ðþMÞ
C6H5OOI þ hn; ðþMÞ ! C6H5O þ IO ðþMÞ
ð34Þ
ð35Þ
ð36Þ
Kinetic simulation of the iodobenzene photolysis system
In order to investigate the possible importance of reactions
(34)–(36) for NO3 formation in our reaction system, the
photolysis of iodobenzene in the presence of NO2 and O2
was simulated, based on the reaction mechanism and the rate
constants summarized in Table 1. In order to estimate an
upper bound for the yield of NO3 by the IONO2 mechanism,
reactions (34) and (36) were combined to
Addition of iodine atoms to methylperoxy and hydroxyethyl-
peroxy radicals to form the corresponding peroxyiodides has
been suggested by Jenkin and Cox.60 Reaction (35), however,
is slow under our reaction conditions due to the small sta-
tionary concentrations of both radical species. Reactions (34)
and (36) are speculative at the present time. If IO radicals are
formed from phenylperoxy iodide in reaction (36) by either
photolysis or by spontaneous or collision induced decomposi-
tion, NO3 could be formed via reaction (31), followed by
reactions (32) and/or (33). Although Barnes et al. favour the
photolysis of IONO2 as the source of NO3 in their system, they
C6H5O2 þ I2 ! C6H5O þ IO þ I;
ð37Þ
assigning a safe upper limit of 5 ꢀ 10ꢁ10 cm3 moleculeꢁ1 sꢁ1 to
the rate constant of this reaction. Rate constants for reactions
(5) and (37) were varied between 0 and 7 ꢀ 10ꢁ12 cm3
moleculeꢁ1 sꢁ1 and between 0 and 5 ꢀ 10ꢁ10 cm3 moleculeꢁ1
Table 7 Measured and calculated product concentrations in the photolysis of iodobenzene at 253.7 nm in synthetic air (run 3 of Table 5); ptot =
1 bar, T = 298.15 K, [C6H5I]0 = 2.43 ꢀ 1014 molecule cmꢁ3, [NO2]0 = 1.48 ꢀ 1014 molecule cmꢁ3, photolysis time = 30 s; concentrations in units
of 1013 molecule cmꢁ3, rate constants in units of cm3 moleculeꢁ1 ꢁ1; numerical calculations based on data of Table 1
s
Calculated based on data in
Table 1, k5 = 7 ꢀ 10ꢁ12
k37 = 0
Calculated based on data in
Table 1, k5 = 0,
Calculated based on data in
Table 1, k5 = 7 ꢀ 10ꢁ12
,
,
Measured
k37 = 5 ꢀ 10ꢁ10
k37 = 5 ꢀ 10ꢁ10
–D[C6H5I]
–D[NO2]
[NO]
4.9
9.18
o0.1
1.9
4.8
12.0
0.005
2.3
4.8
11.7
0.005
2.3
4.8
11.9
0.004
2.3
[N2O5]
[o-Nitrophenol]
[HNO3]
[CO]
1.3
1.0
0.8
0.3
1.4
0.7
0.8
0.5
1.4
0.6
0.8
0.5
1.4
0.7
0.8
0.5
[O3]
Table 8 Experimental product yields in the photolysis of nitrosobenzene at 298 K in the presence of NO2, O2 (200 mbar), and N2 (800 mbar); t =
total photolysis time, concentrations are given in units of molecule cmꢁ3. Product yields (in %) refer to the total amount of converted
nitrosobenzene
[o-Nitro-
phenol]
(%)
[Nprod]/
a
[PhNO]
(1013
–D[PhNO] [NO2]
–D[NO2] [NO]
[N2O5] [HNO3] [Nreact
(%)
]
[HONO] [C6H5NO2]
(%)
Run
t/s
)
(1013
)
(1013
)
(%)
(%)
(%)
(%)
(%)
[CO] (%)
1
a
b
c
0
4
14
24
0
20
0
20
0
23.8
21.4
9.2
3.3
25.0
4.2
25.0
3.5
25.0
20.7
8.9
0
83.2
80.9
74.1
72.9
77.5
70.6
9.6
0
93.3
61.8
50.0
0
32.9
0
6.0
0
0
0
r4.2
16.2
20.3
0
24.6
0
20.6
0
0
24.3
12.4
8.0
0
2.8
0
r0.5
0
0
0
0
0
4.7
6.7
0
7.8
0
13.7
0
0
0
2.4
14.6
20.5
0
20.8
0
21.5
0
4.3
16.1
21.7
71.4
41.1
33.6
0
21.2
0
2.9
0
62.4
40.6
33.0
3.8
4.0
8.2
0
9.7
0
2.8
0
1.1
4.6
8.9
68.8
60.9
62.1
0
59.7
0
57.5
0
64.4
64.5
66.2
5.0
7.8
8.3
0
10.5
0
19.9
0
5.5
7.8
10.3
55.1
53.0
53.5
0
56.7
0
52.9
0
56.2
55.2
54.6
2
3
8.3
4
a
b
c
84.6
79.4
74.4
73.0
4
14
24
117.7
62.9
53.4
27.1
20.7
22.3
19.3
12.1
9.2
5.5
7.1
8.7
3.3
a
[Nprod] ꢇ [NO] + [o-nitrophenol] + 2[N2O5] + [HNO3] + [HONO] + [C6H5NO2], [Nreact] ꢇ –D[C6H5NO] ꢁ D[NO2].
ꢃc
This journal is the Owner Societies 2007
5046 | Phys. Chem. Chem. Phys., 2007, 9, 5036–5051

View Article Online
s
ꢁ1, respectively. The following mechanisms for the formation
different final products are taken into account by reactions
(39a) and (39b):
of NO3 were distinguished:
Mechanism I: k5 ¼ 7 ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1
;
C6H5NO þ hn ! NO þ C6H5 ðthermally distributedÞ ð39aÞ
C6H5NO þ hn ! NO þ C6H5ꢄꢄ ðhighly excitedÞ ð39bÞ
k37 ¼ 0 cm3 moleculeꢁ1 sꢁ1
Mechanism II: k5 ¼ 0 cm3 moleculeꢁ1 sꢁ1
;
The reaction conditions and the product distributions of 4
experiments are summarized in Table 8. Due to the very large
absorption cross section of nitrosobenzene (s254 nm = 1.8 ꢀ
10ꢁ17 cm2 molecule1 62), the photolysis is very fast. For this
reason, shorter photolysis times were chosen as compared to
the iodobenzene experiments. Product spectra of run 1b are
shown in Fig. 3.
k37 ¼ 5 ꢀ 10ꢁ10 cm3 moleculeꢁ1 sꢁ1
Mechanism III : k5 ¼ 7 ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1
k37 ¼ 5 ꢀ 10ꢁ10 cm3 moleculeꢁ1 sꢁ1
;
The results are shown in Table 7. All of the mechanisms can
reproduce the experimental product distributions pretty well.
The smaller consumption of NO2 and the very low yield of
N2O5 obtained with initial NO present (run 7 in Table 5) are
also reproduced by all mechanisms. The modelled ozone
concentrations are significantly larger than the measured
values. This deficit, however, is not considered to be a serious
drawback of the mechanisms since, applying the rate constant
k
38 = 2.9 ꢀ 10ꢁ13 cm3 moleculeꢁ1 sꢁ1 of Tao and Li33 for the
reaction of ozone with phenoxy radicals,
C6H5O þ O3 ! C6H5O2 þ O2;
ð38Þ
a significant part of the ozone is already destroyed by this
reaction and the authors of this work state that their value for
k38 may be too low.
The following conditions must be valid in order that the
IONO2 mechanism of NO3 formation be effective in the
C6H5I–NO2–O2–N2 photolysis system:
ꢈ Phenylperoxy radicals must react very fast with molecular
iodine to form IO radicals (either directly or by thermal
decomposition of phenylperoxy iodide).
ꢈ IONO2 must photolyse with a high quantum yield gen-
erating NO3 and I atoms.
Since both these requirements are speculative at the present
time, the N2O5 formation in the C6H5I–NO2–O2–N2 photo-
lysis is in favour of the existence of reaction (5).
Photolysis of nitrosobenzene (C6H5NO) in the presence of O2
and NO2
In order to further support the existence of reaction (5), an
iodine free reaction system leading to phenylperoxy radicals
was investigated, i.e. the photolysis of nitrosobenzene in the
presence of O2, NO2, and N2 at 298 K and a total pressure of 1
bar. Since NO is formed in the primary photolysis step,61
C6H5NO þ hn ! NO þ C6H5ðwith unspecified excitationÞ;
ð39Þ
Fig. 3 Product spectra in the C6H5NO–O2–NO2–N2 photolysis
system at 298 K (run 1b from Table 8); [C6H5NO]0 = 2.38 ꢀ 1014
molecule cmꢁ3, [NO2]0 = 8.32 ꢀ 1014 molecule cmꢁ3, [O2]0 = 4.86 ꢀ
1018 molecule cmꢁ3, ptot = 1000 mbar, photolysis time 14 s, 8 scans.
(a) Reaction mixture before irradiation; (b) product spectrum minus
residual absorption of the reactant nitrosobenzene; (c) spectrum (b)
minus absorptions from NO2, HNO3, HONO, HO2NO2, CO and
CO2; (d) spectrum (c) minus absorption from o-nitrophenol; (e)
spectrum (d) minus absorptions from N2O5 and nitrobenzene, inset:
absorption from aerosol in the 3000–3500 cmꢁ1 region.
the yield of NO3 is heavily reduced by the reaction of NO3
with NO. On the other hand, (i) the uncertainties related to
reactions of several iodine containing species are avoided in
this system, and (ii) simulation of the reaction system suggests
that a major part of the NO3 formed in reaction (50) would be
scavenged as N2O5 provided that enough NO2 (about 1 ꢀ 1015
molecule cmꢁ3) is present. In the mechanism of Table 1,
different internal energies of C6H5 potentially leading to
ꢃc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 5036–5051 | 5047

View Article Online
Main features of the experimental product distributions in
the photolysis of nitrosobenzene in the presence of O2 and
NO2 are as follows:
in the simulations that are low by a factor of 13, thus strongly
supporting the existence of reaction (5).
By fitting the calculated to the measured N2O5 yields with k5
as the only parameter, a rough estimate of k5 can be derived
from the data in Table 8. Since the calculated N2O5 concen-
tration is very sensitive to the value of k5 for k5 r 10ꢁ12 cm3
ꢈ Fairly high concentrations of N2O5 were observed in these
experiments (see Table 8 and Fig. 3). Yet, they are significantly
lower as compared to the iodobenzene photolysis experiments
mainly because of two reasons: the direct reaction of NO3 with
NO is more important and NO competes with NO2 for
phenylperoxy radicals, thus reducing the extent of reaction
(5) and the formation of NO3 and N2O5.
moleculeꢁ1
s
ꢁ1, a lower limit of k5 4 1 ꢀ 10ꢁ12 cm3
moleculeꢁ1 sꢁ1 can be derived. The best fit gives k5 E 7 ꢀ
10ꢁ12 cm3 moleculeꢁ1 sꢁ1, even though with large error limits.
Coincidentally, this value is equal to the recombination rate
constants measured for other RO2 + NO2 reactions at a total
ꢈ CO yields are much higher than in the photolysis experi-
ments of o-nitrophenol (Table 4) and iodobenzene (Table 5).
They are independent of the initial concentration of NO2.
ꢈ It proved to be difficult to rationalize, by the same
mechanism, the different CO yields in the photolysis of
iodobenzene and nitrosobenzene (E17% vs. E55%) as well
as their independence of the NO and NO2 concentrations (see
Tables 5 and 8). The most plausible reason is to trace the
different yields back to the different bond energies in iodo-
benzene and nitrosobenzene (280 vs. 227 kJ molꢁ1), providing
the phenyl radicals resulting from photolysis with 254 nm
photons with different excess energies: r192 kJ molꢁ1 in the
iodobenzene photolysis and r244 kJ molꢁ1 in the nitroso-
benzene photolysis. Since every fifth collision partner of C6H5
radicals is a reactive O2 molecule, energy-rich phenoxy radi-
cals could be formed:
pressure of 1 bar (E7 ꢀ 10ꢁ12 cm3 moleculeꢁ1
s
ꢁ1 27). Thus,
we propose
k5 ꢂ 7 ꢀ 10ꢁ12 cm3 moleculeꢁ1 sꢁ1 ð298 K; 1 barÞ:
It may be noted that a value of k = 1.0 ꢀ 10ꢁ11 cm3
moleculeꢁ1 sꢁ1 at 298 K was assumed by Caralp et al. for
the reaction C6H5O2 + NO2 - products (without specifying
the products) to model the UV absorption traces in their
experiments on the flash photolysis of Cl2–C6H5C(O)-
H–O2–NO–N2 mixtures.10
In run 3 of Table 8 the initial NO2 concentration is too small
to scavenge NO3 effectively—in agreement with the model
calculated product distribution, the N2O5 yield is below the
detection limit.
C6H5I þ hn ! C6Hꢄ5 þ I
C6Hꢄ5 þ O2 ! C6H5Oꢄ þ O
ð22bÞ
ð40Þ
A striking feature in Table 9 is that the measured distinct
loss of NO2 is poorly reproduced in the computer simulations.
This result may be the clue to another finding: in the nitroso-
benzene photolysis (see spectrum (e) in Fig. 3), there is a
prominent, yet unidentified product band near 1720 cmꢁ1
which might originate from a C5 product containing an NO2
group. This product would explain both the large NO2 loss
and the missing carbon complementary to the CO yield. In the
iodobenzene photolysis, on the other hand, the absorption at
1720 cmꢁ1 is very small and, at the same time, the NO2 loss is
better represented by the simulations (see Table 5), and the
yields of CO and the expected, yet unidentified, C5 compound,
are much lower.
C6H5Oꢄ ! CO þ productðsÞ ðC5H5 ?Þ
C6H5NO þ hn ! NO þ C6H5ꢄꢄ
C6H5ꢄꢄ þ O2 ! C6H5Oꢄꢄ þ O
ð41Þ
ð39bÞ
ð43Þ
C6H5Oꢄꢄ ! CO þ productðsÞ ðC5H5 ?Þ
ð43Þ
The larger internal energy of most of the phenyl radicals
generated in the photolysis of nitrosobenzene (reaction
(39b)) might then explain the larger CO yield in this system
as compared to the iodobenzene photolysis. For this reason,
reactions (22b), (40), (41), (39b), (42) and (43), though com-
pletely speculative at the present time, were included in the
reaction mechanism (Table 1). These reactions make allow-
ance for the experimental findings of very different CO yields
in the photolysis of iodobenzene and nitrosobenzene and of
the independence of the CO yields on the NO and NO2
concentrations.
Table 9 Measured and calculated product concentrations in the
photolysis of nitrosobenzene at 253.7 nm in synthetic air, ptot = 1
bar (800 mbar N2 + 200 mbar O2), T = 298.1 K, [C6H5NO]0 ꢁ3
=
,
2.38 ꢀ 1014 molecule cmꢁ3, [NO2]0 = 8.32 ꢀ 1014 molecule cm
photolysis time = 14 s; concentrations in units of 1013 molecule cmꢁ3
(run 1b of Table 8, numerical calculations based on data in Table 1)
In the presence of NO, phenoxy radicals may add NO to
form phenyl nitrite.63 The formation of this species, however,
was not included in the reaction mechanism of Table 1 since its
thermal lifetime at our reaction conditions (0.12 s at 298 K, 1
bar, M = N2 63) is too small to affect the product distribution.
In Table 9, measured and model calculated concentrations
are compared for run 1b from Table 8. Main conclusions from
the data in Table 9 are that inclusion of reaction (5) is able to
reasonably describe the measured product distribution
whereas exclusion of reaction (5) leads to N2O5 concentrations
Calculated with
k5 = 7ꢀ10ꢁ12
Calculated with
k5 = 0
Measured
–C6H5NO
–NO2
NO
14.6
9.0
6.0
1.8
2.4
0.6
7.7
o0.3
14.6
3.7
5.9
1.8
2.6
0.4
7.6
0.4
16.4
0.1
6.4
0.14
2.4
0.2
6.8
0.3
N2O5
o-Nitrophenol
HNO3
CO
O3
ꢃc
This journal is the Owner Societies 2007
5048 | Phys. Chem. Chem. Phys., 2007, 9, 5036–5051

View Article Online
Table 10 Calculated enthalpies of reaction (4) for R = H, C2H5, CH3CO, and C6H5
DrH0298/kJ molꢁ1
UCCSD(T)/cc-pVTZ
(this work)
From tabulated enthalpies of formation
(see Table 11)
Reaction
HOO + NO2 - OH + NO3
C2H5OO + NO2 - C2H5O + NO3
CH3C(O)O2 + NO2 - CH3C(O)O + NO3
C6H5O2 + NO2 - C6H5O + NO3
a
73.3
62.6
22.9
ꢁ31.9
63.2
54.3
(+2.4a)
ꢁ72.4
This value is deduced by Miller et al.64 from a variety of thermochemical data to get best internal consistency.
transformation of the resulting phenylperoxy radicals to phe-
Discussion and conclusions
noxy radicals can take place via their reaction with NO or, via
reaction (5), with NO2. Similar reactions could occur during
the degradation of other methyl substituted benzenes (e.g.
xylenes). An estimate based on typical trace gas concentrations
in polluted air shows that reaction (5) probably will not
significantly increase the production rate of NO3 in the atmo-
sphere (in the order of several % during the day) which is
dominated by the reaction of NO2 with ozone. In laboratory
systems dealing with the degradation of aromatic compounds,
In the presence of NO2 and the absence of NO, organic peroxy
radicals usually add NO2 to form peroxynitrates. The experi-
ment on the photolysis of iodoethane from the present work
behaves exactly like this: ethyl peroxynitrate is the only major
product that appears in the product IR spectrum. In the
photolysis of iodobenzene and nitrosobenzene, on the other
hand, peroxynitrate bands do not appear. Instead, large yields
of N2O5 are formed: 20–50% of the iodobenzene converted
after 30 s photolysis, 12–24% of the nitrosobenzene converted
at 5–15 s photolysis. These observations point to the forma-
tion of a peroxynitrate intermediate that is unstable with
respect to fission of the O–O bond, thus leading to NO3 and,
finally, N2O5 or HNO3.
This conclusion is supported by another finding: in the
iodobenzene photolysis system, the decrease of the initial
NO2 concentration is very strong (approaching 200% of the
converted iodobenzene, see Table 5), in agreement with each
NO3 radical using up another NO2 molecule to form N2O5.
Ab initio calculations at the RHF-UCCSD(T) level were
performed on reaction (4),
RO2 þ NO2 ! RO þ NO3
ð4Þ
for R = H, C2H5, CH3CO and C6H5, using the Gaussian 98
and MOLPRO program packages. The aim of these calculations
was to get an idea about the dependence of the thermochemistry
of reaction (4) on the chemical nature of the group R. The
results are presented in Table 10 and Fig. 4.
The calculated energies are consistent with the experimental
results from the literature that peroxynitric acid, ethyl peroxy-
nitrate and acetyl peroxynitrate are either thermally stable at
room temperature, in particular in the presence of an excess of
NO2, or decompose to RO2 and NO2 on a time scale of
seconds or minutes.3,27,68 In addition, the enthalpy calculated
for reaction (5) supports the conclusion from the present
experiments that phenyl peroxynitrate is thermally unstable
at room temperature with respect to O–O bond fission. Such
an O–O bond rupture was already suggested earlier by Orlan-
do and Tyndall for glyoxyl peroxynitrate.6 For phenylperoxy
radicals, the notable resonance energy of the phenoxy radical15
may be the driving force for reaction (5) to occur. With the
enthalpies of formation in Table 11, reaction enthalpies were
estimated and also included in Table 10.
The side-chain oxidation of toluene leads to phenyl radicals,
via benzaldehyde, benzoylperoxy radicals, their reaction with
NO and rapid elimination of CO2.9,10 After the addition of O2,
Fig. 4 Calculated O–N– and O–O– bond energies ((U)CCSD(T)/
cc-pVTZ) in HOONO2, C2H5OONO2, CH3C(O)OONO2 and
C6H5OONO2.
ꢃc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 5036–5051 | 5049

View Article Online
Table 11 Enthalpies of formation used in this work
DfH0
Walker and J. Warnatz, J. Phys. Chem. Ref. Data, 2005, 34,
757–1397.
12 P. Frank, J. Herzler, Th. Just and C. Wahl, Proceedings of the 25th
Symposium (Int.) on Combustion, The Combustion Institute,
Pittsburgh, 1995, pp. 833–840.
13 S. S. Kumaran and J. V. Michael, Proceedings of the 21th
Symposium on Shock Tubes and Waves, ed. A. F. P. Houwing
et al., Panther Publishing, Australia, 1997, pp. 241–244.
14 (a) T. Yu and M. C. Lin, J. Am. Chem. Soc., 1993, 115, 4371–4372;
(b) T. Yu and M. C. Lin, J. Am. Chem. Soc., 1994, 116,
9571–9576.
15 A. J. Colussi, F. Zabel and S. W. Benson, Int. J. Chem. Kinet.,
1977, 9, 161–178.
16 D. G. Hendry and R. A. Kenley, Atmospheric Chemistry of
Peroxynitrates, in Nitrogenous Air Pollutants, ed. D. Grosjean,
Ann Arbor Press, Ann Arbor, MI, 1979.
17 H. G. Libuda, O. Shestakov, J. Theloke and F. Zabel, Phys. Chem.
Chem. Phys., 2002, 4, 2579–2586.
18 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC,
Boca Raton, FL, 74th edn, 1993.
19 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, R.
E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D.
Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V.
Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo,
S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K.
Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B.
Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Stefanov,
G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R.
L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. G.
Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M.
Head-Gordon, E. S. Replogle and J. A. Pople, GAUSSIAN 98
(Revision A.7), Gaussian, Inc., Pittsburgh, PA, 1998.
/
298
Species
kJ molꢁ1
Ref.
H
O
OH
HO2
NO
NO2
NO3
N2O5
I
I2
IO
IONO2
CO
CO2
CH3
CH3O
CH3OO
C2H5O
C2H5OO
CH3C(O)O
CH3C(O)OO
C6H5 (phenyl)
C6H5O (phenoxy)
C6H5OO (phenylperoxy)
C6H5OH (phenol)
217.998 ꢆ 0.006 Atkinson et al.26 (2004)
249.18 ꢆ 0.10 Atkinson et al.26 (2004)
37.3 ꢆ 0.3
14.6
90.25
Ruscic et al.65 (2005)
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
33.18
73.72 ꢆ 1.4
11.3
106.76 ꢆ 0.04 Atkinson et al.26 (2004)
62.42 ꢆ 0.08
115.9 ꢆ 5.0
70 ꢆ 16
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
Atkinson et al.26 (2004)
ꢁ110.53 ꢆ 0.17 Atkinson et al.26 (2004)
ꢁ393.51 ꢆ 0.13 Atkinson et al.26 (2004)
146.7 ꢆ 0.3
21.0 ꢆ 2.1
9.0 ꢆ 5.1
ꢁ13.6 ꢆ 4.0
ꢁ27.4 ꢆ 9.9
ꢁ192.5 ꢆ 5.8
ꢁ154.4 ꢆ 5.8
338.0
Ruscic et al.65 (2005)
Ruscic et al.65 (2005)
Atkinson et al.26 (2004)
Ruscic et al.65 (2005)
Baulch et al.11 (2005)
Miller et al.64 (2004)
Miller et al.64 (2004)
Baulch et al.11 (2005)
Baulch et al.11 (2005)
Baulch et al.11 (2005)
Baulch et al.11 (2005)
NIST66 (2005)
47.70
160.66
ꢁ96.39
C6H5NO (nitrosobenzene) 201.0 ꢆ 4.2
C6H5NO2 (nitrobenzene)
68.53 ꢆ 0.67
NIST66 (2005)
C6H5ONO (phenyl nitrite) 51.5 ꢆ 8.0
Berho et al.63 (1998)
NIST66 (2005)
C6H5I (iodobenzene)
C5H5 (cyclopentadienyl)
C6H5–C6H5 (biphenyl)
165.0 ꢆ 5.9
261.50
182.0 ꢆ 0.7
Baulch et al.11 (2005)
Chirico et al.67 (1989)
20 H.-J. Werner, P. J. Knowles, R. Lindh, M. Schutz, P. Celani, T.
¨
Korona, F. R. Manby, G. Rauhut, R. D. Amos, A. Bernhardsson,
A. Berning, D. L. Cooper, M. J. O. Deegan, A. J. Dobbyn, F.
Eckert, C. Hampel, G. Hetzer, A. W. Lloyd, S. J. McNicholas, W.
Meyer, M. E. Mura, A. Nicklass, P. Palmieri, R. Pitzer, U.
Schumann, H. Stoll, A. J. Stone, R. Tarroni and T. Thorsteinsson,
MOLPRO, version 2002.6, a package of ab initio programs, see
http://www.molpro.net.
however, reaction (5) may significantly modify the distribution
of different nitrogen oxides.
21 (a) P. J. Knowles, C. Hampel and H.-J. Werner, J. Chem. Phys.,
1993, 99, 5219–5227; (b) P. J. Knowles, C. Hampel and
H. J. Werner, J. Chem. Phys., 2000, 112, 3106–3107; (c) J. D.
Watts, J. Gauss and R. J. Bartlett, J. Chem. Phys., 1993, 98,
8718–8733.
Acknowledgements
Financial support of this work by the Deutsche Forschungs-
gemeinschaft (DFG) is gratefully acknowledged.
22 C. Hampel, K. A. Peterson and H.-J. Werner, Chem. Phys. Lett.,
1992, 190, 1–12.
23 W. Eisfeld and K. Morokuma, J. Chem. Phys., 2000, 113,
5587–5597.
24 P. Deuflhard and U. Nowak, Ber. Bunsen-Ges. Phys. Chem., 1986,
90, 940–946.
References
1 P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau, G. D.
Hayman, M. E. Jenkin, G. K. Moortgat and F. Zabel, Atmos.
Environ., 1992, 26A, 1805–1961.
25 R. Atkinson and J. Arey, Chem. Rev., 2003, 103, 4605–4638.
26 R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F.
Hampson, R. G. Hynes, M. E. Jenkin, M. J. Rossi and J. Troe,
Atmos. Chem. Phys., 2004, 4, 1461–1738.
2 T. J. Wallington, P. Dagaut and M. J. Kurylo, Chem. Rev., 1992,
92, 667–710.
3 F. Kirchner, A. Mayer-Figge, F. Zabel and K. H. Becker, Int. J.
Chem. Kinet., 1999, 31, 127–144.
27 S. P. Sander, R. R. Friedl, A. R. Ravishankara, D. M. Golden, C.
E. Kolb, M. J. Kurylo, R. M. J. Molina, G. K. Moortgat, H.
Keller-Rudek, B. J. Finlayson-Pitts, P. H. Wine, R. E. Huie and V.
L. Orkin, Chemical kinetics and photochemical data for use in
atmospheric studies, Evaluation 15, Jet Propulsion Laboratory,
Pasadena, CA, JPL Publication 06-2, 2006.
28 M. E. Jenkin, R. A. Cox, A. Mellouki, G. Le Bras and G. Poulet, J.
Phys. Chem., 1990, 94, 2927–2934.
29 R. M. Chambers, A. C. Heard and R. P. Wayne, J. Phys. Chem.,
1992, 96, 3321–3331.
4 A. Mayer-Figge, PhD Thesis, University of Wuppertal, 1997.
5 F. Kirchner, PhD Thesis, University of Wuppertal, 1993.
6 J. J. Orlando and G. S. Tyndall, Int. J. Chem. Kinet., 2001, 33,
149–156.
7 S. Jagiella and F. Zabel, Abstract Book of the 17th International
Symposium on Gas Kinetics, Essen, Germany, 2002, contribution
AP 24.
8 S. Jagiella and F. Zabel, Abstract Book of the Annual Meeting of
the Deutsche Bunsengesellschaft fur Physikalische Chemie, Dresden,
2004, contribution no. B13.
¨
30 H. van den Bergh and J. Troe, J. Chem. Phys., 1976, 64, 736–742.
31 F. Berho and R. Lesclaux, Chem. Phys. Lett., 1997, 279,
289–296.
32 F. Caralp, V. Foucher, R. Lesclaux, T. J. Wallington and M. D.
Hurley, Phys. Chem. Chem. Phys., 1999, 1, 3509–3517.
33 Z. Tao and Z. Li, Int. J. Chem. Kinet., 1999, 31, 65–72.
34 C. A. Cantrell, J. A. Davidson, A. H. McDaniel, R. E. Shetter and
J. G. Calvert, Chem. Phys. Lett., 1988, 148, 358–363.
9 H. Niki, P. D. Maker, C. M. Savage and L. P. Breitenbach, Fourier
Transform Infrared (FTIR) Studies of Gaseous and Particulate
Nitrogenous Compounds, in Nitrogenous Air Pollutants, ed. D.
Grosjean, Ann Arbor Press, Ann Arbor, MI, 1979.
10 F. Caralp, V. Foucher, R. Lesclaux, T. J. Wallington and M. D.
Hurley, Phys. Chem. Chem. Phys., 1999, 1, 3509–3517.
11 D. L. Baulch, C. T. Bowman, C. J. Cobos, R. A. Cox, Th. Just, J.
A. Kerr, M. J. Pilling, D. Stocker, J. Troe, W. Tsang, R. W.
ꢃc
This journal is the Owner Societies 2007
5050 | Phys. Chem. Chem. Phys., 2007, 9, 5036–5051

View Article Online
35 R. A. Graham and H. S. Johnston, J. Phys. Chem., 1978, 82,
254–268.
52 I. V. Tokmakov, G.-S. Kim, V. V. Kislov, A. M. Mebel and M. C.
Lin, J. Phys. Chem. A, 2005, 109, 6114–6127.
36 Philips Product Information, Data Sheets no. 1817, 1995.
37 C. Chackerian, S. W. Sharpe and T. A. Blake, J. Quant. Spectrosc.
Radiat. Transfer, 2003, 82, 429–441.
53 D. Schuler, S. Jagiella and F. Zabel, 2006, unpublished results.
54 However, a very low upper limit of ko 5 ꢀ 10ꢁ21 cm3 moleculeꢁ1
sꢁ1 has been determined for the reaction of phenoxy radicals with
O2 at 298 K by Platz et al. (ref. 50).
55 E. Bolzacchini, M. Bruschi, J. Hjorth, S. Meinardi, M. Orlandi, B.
Rindone and E. Rosenbohm, Environ. Sci. Technol., 2001, 35,
1791–1797.
56 J. Hjorth, J. Viidanoja, R. Manca, J.-P. Putaud, D. Ferri, N. R.
Jensen, C. Astorga, B. R. Larsen, R. Van Dingenen, A. Dell’Ac-
qua, R. Winterhalter and E. Bolzacchini, Proc. EUROTRAC-2
Symposium 2002, Garmisch-Partenkirchen, ed. P. M. Midgley and
M. Reuther, Margraf Publishers, Weikersheim, Germany, 2003.
57 R. I. Olariu, B. Klotz, I. Barnes, K. H. Becker and R. Mocanu,
Atmos. Environ., 2002, 36, 3685–3697.
38 The IR absorption cross section s(HNO3, 3551.3 cmꢁ1, base e) =
6.2 ꢀ 10ꢁ19 cm2 moleculeꢁ1 was used, applying the integrated
absorption of the 3552 cmꢁ1 band from ref. 37 (A = 1.238 ꢀ 10ꢁ17
cm moleculeꢁ1) to our reference spectrum. This value for A is
supported by results of Okumura et al. (personal communication,
2006) whose preliminary value for A agrees with that of Chack-
erian et al. within their uncertainties.
39 W. S. Barney, L. M. Wingen, M. J. Lakin, T. Brauers, J. Stutz and
B. J. Finlayson-Pitts, J. Phys. Chem. A, 2000, 104, 1692–1699.
40 P. Wiesen, J. Kleffmann, R. Kurtenbach and K. H. Becker, Fara-
day Discuss., 1995, 100, 121–127.
41 I. Bejan, Y. Abd El Aal, I. Barnes, T. Benter, B. Bohn, P. Wiesen
and J. Kleffmann, Phys. Chem. Chem. Phys., 2006, 8, 2028–2035.
42 I. Barnes, K. H. Becker and J. Starcke, J. Phys. Chem., 1991, 95,
9736–9740.
58 R. Volkamer, P. Spietz, J. Burrows and U. Platt, J. Photochem.
Photobiol., A, 2005, 172, 35–46.
59 J. C. Mossinger, D. M. Rowley and R. A. Cox, Atmos. Chem.
¨
Phys. Discuss., 2002, 2, 765–788.
43 H. Niki, P. D. Maker, C. M. Savage and L. P. Breitenbach, Chem.
Phys. Lett., 1978, 55, 289–292.
60 M. E. Jenkin and R. A. Cox, J. Phys. Chem., 1991, 95,
3229–3237.
44 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC,
Boca Raton FL, 86th edn, 2005.
45 J. G. Calvert and J. N. Pitts, Jr, Photochemistry, John Wiley, New
York, 1966.
61 C.-M. Tseng, Y. M. Choi, C.-L. Huang, C.-K. Ni, Y. T. Lee and
M. C. Lin, J. Phys. Chem. A, 2004, 108, 7928–7935.
62 K. Tabei and S. Nagakura, Bull. Chem. Soc. Jpn., 1965, 38,
965–971.
46 In ref. 47, the photolysis of iodobenzene has been studied using
laser photolysis at several wavelengths including 248 nm. The
authors present a UV spectrum (relative s scale) in their Fig. 7.
From their stated absorption coefficient at 248 nm (1.54 ꢀ 10ꢁ18
cm2 moleculeꢁ1) and the ratio of the relative absorption coefficients
at 248 and 253.7 nm from their Fig. 7, s(253.7 nm) = 1.86 ꢀ 10ꢁ18
cm2 moleculeꢁ1 is estimated.
47 W. H. Pence, S. L. Baughcum and S. R. Leone, J. Phys. Chem.,
1981, 85, 3844–3851.
48 A. Saiz-Lopez, R. W. Saunders, D. M. Joseph, S. A. Ashworth and
J. M. C. Plane, Atmos. Chem. Phys., 2004, 4, 1443–1450.
49 R. Atkinson, S. M. Aschmann and J. Arey, Environ. Sci. Technol.,
1992, 26, 1397–1403.
63 F. Berho, F. Caralp, M.-T. Rayez, R. Lesclaux and E. Ratajczak,
J. Phys. Chem. A, 1998, 102, 1–8.
64 C. E. Miller, J. I. Lynton, D. M. Keevil and J. S. Francisco, J.
Phys. Chem. A, 1999, 103, 11451–11459.
´ ´
65 B. Ruscic, J. E. Boggs, A. Burcat, A. G. Csaszar, J. Demaison, R.
Janoschek, J. M. L. Martin, M. L. Morton, M. J. Rossi, J. F.
Stanton, P. G. Szalay, P. R. Westmoreland, F. Zabel and T.
Berces, J. Phys. Chem. Ref. Data, 2005, 34, 573–656.
´
66 H. Y. Afeefy, J. F. Liebman and S. E. Stein, Neutral Thermo-
chemical Data, in NIST Chemistry WebBook, NIST Standard
Reference Database Number 69, ed. P. J. Linstrom and W. G.
Mallard, National Institute of Standards and Technology,
Gaithersburg MD, June 2005, 20899, http://webbook.nist.gov.
67 R. D. Chirico, S. E. Knipmeyer, A. Nguyen and W. V. Steele, J.
Chem. Thermodyn., 1989, 21, 1307–1331.
68 R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F.
Hampson, R. G. Hynes, M. E. Jenkin, M. J. Rossi and J. Troe,
Atmos. Chem. Phys., 2006, 6, 3625–4055.
50 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, 7964–7974.
51 M. J. Fadden, C. Barckholtz and C. M. Hadad, J. Phys. Chem. A,
2000, 104, 3004–3011.
ꢃc
This journal is the Owner Societies 2007
Phys. Chem. Chem. Phys., 2007, 9, 5036–5051 | 5051