Formation of HO2 radicals from the 248 nm two-photon excitation of different aromatic hydrocarbons in the presence of O2

Article abstract of DOI:10.1021/jp211520g

The excitation energy dependence of HO₂ radical formation from the 248 nm irradiation of four different aromatic hydrocarbons (benzene, toluene, o-xylene, and mesitylene) in the presence of O₂ has been studied. HO₂ has been monitored at 6638.20 cm^-1 by cw-CRDS, and the formation of a short-lived, unidentified species, showing broad-band absorption around the HO₂ absorption line, has been observed. For all four hydrocarbons, the same HO₂ formation pattern has been observed: HO₂ is formed immediately on our time scale after the excitation pulse, followed by a formation of more HO₂ on a much longer time scale. Taking into account the absorption of the short-lived species, the yields of both types of HO₂ radicals are in agreement with a formation following 2-photon absorption by the aromatic hydrocarbons. The yields do not much depend on the nature of the aromatic hydrocarbon. For practical use in past and future experiments on aromatic hydrocarbons, an empirical value is given, allowing the estimation of the total concentration of HO₂ radicals formed at 40 Torr He in the presence of around [O ₂] = 1 × 10¹⁷cm^-3 as a function of the 248 nm excitation energy: [HO₂]/[aromatic hydrocarbon] ≈ 2 × 10^-6 × E² (with E in mJ cm^-2).

Full text of DOI:10.1021/jp211520g

Article
pubs.acs.org/JPCA
Formation of HO2 Radicals from the 248 nm Two-Photon Excitation
of Different Aromatic Hydrocarbons in the Presence of O2
Chaithanya Jain, Pranay Morajkar, Coralie Schoemaecker, and Christa Fittschen*
́ ̀ ́
Universite Lille Nord de France PhysicoChimie des Processus de Combustion et de l’Atmosphere, PC2A CNRS Universite Lille 1,
UMR 8522, F-59650 Villeneuve d’Ascq, France
ABSTRACT: The excitation energy dependence of HO2 radical
formation from the 248 nm irradiation of four different aromatic hydro-
carbons (benzene, toluene, o-xylene, and mesitylene) in the presence of
O2 has been studied. HO2 has been monitored at 6638.20 cm−1 by cw-
CRDS, and the formation of a short-lived, unidentified species, showing
broad-band absorption around the HO2 absorption line, has been
observed. For all four hydrocarbons, the same HO2 formation pattern has
been observed: HO2 is formed immediately on our time scale after the
excitation pulse, followed by a formation of more HO2 on a much longer
time scale. Taking into account the absorption of the short-lived species,
the yields of both types of HO2 radicals are in agreement with a
formation following 2-photon absorption by the aromatic hydrocarbons.
The yields do not much depend on the nature of the aromatic
hydrocarbon. For practical use in past and future experiments on
aromatic hydrocarbons, an empirical value is given, allowing the estimation of the total concentration of HO2 radicals formed at
40 Torr He in the presence of around [O2] = 1 × 1017cm−3 as a function of the 248 nm excitation energy: [HO2]/[aromatic
hydrocarbon] ≈ 2 × 10−6 × E2 (with E in mJ cm−2).
and leads to wrong conclusions. Following this discovery of the
HO2 formation, Kovacs and co-workers16 investigated the
formation of H-atoms following the 248 nm irradiation of
benzene and toluene and concluded that the observed H-atoms
stem from a 2-photon process. They suggested that the HO2
radicals observed in our experiments15 might as well originate
from a fast reaction of O2 with byproduct of the H-atoms, i.e
from a 2-photon process. Following the publication of Kovacs
and co-workers, we have reinvestigated the energy dependence
of reaction 1 using higher excitation energies17 than in the
earlier study, where experiments had been performed only at
one low excitation energy (13 mJ cm−2).15 Doing so, we
observed that with increasing excitation energy the formation of
a secondary HO2 becomes visible on a much longer time scale
compared to the first, prompt HO2. The rate of formation of
this secondary HO2 was strongly pressure dependent and in
reasonable agreement with the rate constant for the reaction of
H-atoms with O2. The energy dependence of the yield of the
first, prompt HO2 radical showed a linear behavior, whereas the
energy dependence of the yield of the secondary HO2 radical
was in agreement with a 2-photon process.
INTRODUCTION
■
Aromatic hydrocarbons, in particular the so-called BTX
(benzene, toluene, xylene), are an important class of volatile
organic compounds emitted to the atmosphere from vehicle
emissions, from various industrial processes, and from the use
of organic solvents.1 Their presence in the atmosphere was first
detected by Lonneman et al.,2 and a large number of laboratory,
chamber, and field studies has been devoted to unraveling the
oxidation mechanism of this class of hydrocarbons.3−10
Aromatic hydrocarbons also thought to contribute significantly
to the photochemical formation of tropospheric ozone even
though the detailed mechanism of the troposheric degradation
of aromatic hydrocarbons is still controversially discussed.11−14
Using the technique of laser photolysis coupled to a
detection of HO2 radicals by cw-CRDS we have shown in an
earlier work that HO2 radicals are formed after 248 nm
irradiation of benzene in the presence of O2.15
C6H6 + hν248nm + O2 → HO2 + products
From the rate of formation it could be concluded that the
HO2 radicals do not originate from a reaction of H-atoms with
O2 but must rather be the product of a very fast reaction of
some kind of excited species with O2. Even though this finding
had no direct impact on the tropospheric degradation of
aromatic hydrocarbons, it was nevertheless interesting because
248 nm is a commonly used wavelength to initiate chemical
reactions in laboratory studies. Hence, an unknown and
unexpected side reaction such as the observed HO2 formation
can possibly influence the interpretation of experimental data
Following the work on benzene,17 we have continued to
investigate the HO2 formation yields of other aromatic
Special Issue: A. R. Ravishankara Festschrift
Received: November 30, 2011
Revised: March 30, 2012
Published: April 13, 2012
© 2012 American Chemical Society
6231
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239

The Journal of Physical Chemistry A
Article
Figure 1. Schematic diagram of the experimental setup. APD = avalanche photodiode, DL = diode laser, OI = optical isolator, BS = beam splitter,
AOM = acousto-optical modulator, M = mirror, and L = lens.
hydrocarbons (toluene, o-xylene, and mesitylene). During these
experiments, a transient broad-band absorption following the
photolysis pulse has been observed in the wavelength range
used for the HO2 dectection (around 1506 nm). This short-
lived absorption had been overlooked during the earlier C6H6
experiments, because it is much less intense than for the
alkylated aromatic hydrocarbons. Considering this transient
absorption as a time-dependent baseline for the HO2 measure-
ments, it appears that both primary and secondary HO2 radicals
stem from 2-photon absorption. In this work we present the
yield of HO2 radical formation, obtained by taking into account
the transient absorption, following 248 nm excitation for 4
different aromatic hydrocarbons: benzene, toluene, o-xylene,
and mesitylene. The yield of HO2 radicals has been
parametrized to provide a simple tool for an estimation of
the impact of this process in past or future laboratory studies.
of 37 cm between photolyzed volume and cw-CRDS absorp-
tion path. A schematic view of the experimental setup is given
in Figure 1.
cw-CRDS was used to monitor the time-resolved kinetics of
HO2 radicals. All HO2 concentration−time profiles have been
measured at peak of the most intense absorption line in the 2ν1
band21 at 6638.20 cm−1. Ring-down times were converted to
HO2 concentrations using the following equation:
⎛
⎜
⎞
⎟
⎠
RL
c × σ τ
1
1
τ0
[HO2]t =
−
⎝
(1)
t
where σ is the absorption cross section, RL is the ratio between
the cavity length L, i.e., the distance between the two cavity
mirrors, and the length LA over which the absorber is present
(in our case the overlap of photolysis beam and absorption
path), c is the speed of light, and τt and τ0 are the ring-down
times in the presence and absence of HO2, respectively.
Throughout the work we have used an absorption cross section
EXPERIMENTAL SECTION
■
of σ6638.20cm = 2.96 × 10−19 cm−2 for HO2, additional
−1
Laser photolysis coupled to the detection of HO2 radicals by
continuous wave cavity ring-down spectroscopy (cw-CRDS)
has been used in the present work. The principle of this setup
has been published earlier,18,19 and some very recent
improvements on the data acquisition system can be found in
Votava et al.20 The setup consists of the photolysis cell, the
photolysis laser, the cw-CRDS system, and the LIF system
(which has not been used for this work). The photolysis reactor
is a three axis stainless steel cell, internally coated with Teflon,
with one long axis (78 cm) and two short axes. Photolysis is
achieved along the longest axis using an excimer laser (Lambda
Physik LPX 202i) operating at 248 nm and a repetition rate of
0.2 Hz. The cw-CRDS absorption path is installed at a small
angle with respect to the photolysis path, leading to an overlap
broadening through added O2 has not been taken into account,
because it leads to a decrease of the absorption cross section of
less than 2% for the highest O2 concentration.22,23
All experiments were carried out at 295 K. Pressure within
the cell was kept constant by a pressure controller (Leybold
MR16) and was monitored with a 0−1000 Torr Baratron
(MKS). All experiments have been performed at total pressures
of around 40 Torr Helium. Different gases were introduced
to the reactor using calibrated mass flow controllers (Tylan
FC260). Typical total gas flows were 300 cm3 min−1, leading to
a flow velocity of 3 cm/s within the reactor. Gases (He 6.0, O2
4.5: both Praxair) were used without further purification,
aromatic hydrocarbons were prepared as diluted gas mixtures in
6232
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239

The Journal of Physical Chemistry A
Article
glass balloons. Before introducing the hydrocarbons (all Aldrich
with following purities: benzene >99%, toluene >99.5%,
o-xylene and mesitylene >98%), the liquid reservoir was
degassed by pump and freeze cycle.
concentration decreased instead at low photolysis energies, as
can be seen in Figure 2b for o-xylene, where this effect is
particularly strong. It turned out that a short-lived species,
absorbing around the HO2 absorption line at 6638.20 cm−1, is
formed after 248 nm excitation of aromatic hydrocarbons,
inducing thus an error in the calculation of the HO2 con-
centration, as explained further down. Figure 3 illustrates for
RESULTS AND DISCUSSIONS
■
Correction of HO2 Concentration for Absorption by a
Short-Lived Species. Following our earlier work on the
formation of HO2 radicals after 248 nm excitation of benzene,
we have performed additional experiments for a series of
alkylated aromatic hydrocarbons: toluene, o-xylene, and
mesitylene. Indeed, aromatic hydrocarbons present absorption
features in the UV region, attributed to the S1 ← S0 electronic
transition, reaching down to wavelengths below 248 nm for
many of them. Absorption cross sections at 248 nm have been
published for some aromatic hydrocarbons by Trost et al.24 as
well as more recently by Fally et al.,25 and it can be noticed that
the cross sections for the aromatic hydrocarbons investigated in
this work are on the same order of magnitude: benzene, 2.3 ×
10−19 cm2; toluene, 1.8 × 10−19 cm2; o-xylene, 3.3 × 10−19 cm2;
mesitylene, 2.5 × 10−19 cm2 (private communication U. Platt).
In the experiments with toluene, o-xylene, and mesitylene,
curious and very different signals compared to the benzene
experiments (Figures 2a) were observed at low photolysis
energy: no secondary HO2 was observed anymore, but the HO2
Figure 3. Ring-down times obtained from the photolysis of [o-xylene]
= 3.1 × 1014 cm−3 and [O2] = 1.0 × 1017 cm−3, E248nm = 10 (circles)
and 35 (square) mJ cm−2 at two different wavelengths: online (filled
symbols, 6638.20 cm−1) and offline (open symbols, 6638.30 cm−1).
o-xylene the time-resolved evolution of the ring-down time at
two different wavelengths: 6638.20 cm−1 (filled symbols, center
of HO2 absorption line) and 6638.30 cm−1 (open symbols, off
any HO2 absorption) for two different photolysis energies.
Please note that the example in Figure 3 has been chosen for
illustration and is one of the most extreme cases, the amplitude
of the transient absorption was much less in most of the
experiments. In our earlier work15,17 the ring-down times have
been converted to HO2 concentrations using eq 1, with τ0
being the average of all ring-down times occurring before the
photolysis pulse. However, the presence of a “time-resolved
baseline” such as shown by the open symbols in Figure 3 does
not allow for such a simple treatment of the raw data but needs
to be taken into account for the conversion of individual ring-
down times into HO2 concentrations.
Therefore, we have first investigated if the transient
absorption is due to a broad-band absorbing species or if the
absorption is structured. In the first case, correction would be
easy by measuring the baseline off any HO2 absorption line; in
the second case, i.e., a structured absorption spectrum of the
transient species, the task would be much more difficult and
unreliable. For this purpose, we have measured the absorption
spectrum around the HO2 absorption line in a similar way as
recently explained in detail for the measurement of the HONO
absorption spectrum:26 the result is shown in Figure 4. Each data
point in this spectrum is obtained from one time-resolved
measurement such as shown in the inset: the data points of the
upper line (a) represent the average of all ring-down times
having occurred before the laser pulse, data points on the middle
line (b) represent the average of all ring-down times having
occurred at delays between 1 and 50 ms after the photolysis
pulse, and data points on the lower line (c) are obtained by
averaging all ring-down times having occurred at delays up to
1 ms after the photolysis pulse. The decay time of the unknown,
absorbing species is short (less than 1 ms) compared to the
Figure 2. (a) Plot of [HO2] against time at different energies. The
open symbols and dashed lines represent the HO2 concentrations
profiles without baseline correction; full symbols and solid lines
represent the HO2 concentration profiles after the baseline correction.
The benzene and O2 concentrations used were 3.35 × 1015 and 1.0 ×
1017 cm−3, respectively. The total pressure was 40 Torr He. (b) Laser
energy dependence of HO2 concentration obtained from the
photolysis of a mixture of [o-xylene] = 3.0 × 1014 cm−3 and [O2] =
9 × 1016 cm−3, showing an unusual behavior at the lower energies.
6233
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239

The Journal of Physical Chemistry A
Article
time-dependent parameters are then used to calculate an
individual ring-down time τ0 for each individual ring-down
event, to be used subsequently in eq 1. The impact on such a
baseline correction on the HO2 concentration−time profiles is
shown in Figure 5 for o-xylene, using two different photolysis
Figure 4. Spectrum measured around the HO2 absorption line:
[o-xylene] = 3.2 × 1014 cm−3, [O2] = 1.0 × 1017 cm−3, photolysis energy
30 mJ cm−2. The inset shows an example (6638.43 cm−1) of how the
spectrum has been obtained from time-resolved measurements. (a) is
the average of all ring-down times before the photolysis pulse, (b) is
the average of ring-down times at delays 0−1 ms after photolysis pulse,
(c) is the average of ring-down times at delays 1−50 ms, (d) = (c) − (b)
(right-hand scale). The dashed, vertical line indicates the wavenumber,
where background measurements have been performed for all offline
measurements.
Figure 5. HO2 concentration time profiles from [o-xylene] = 2.54 ×
1014 cm−3, [O2] = 1.0 × 1017 cm−3 for two different photolysis
energies. Open symbols are obtained by using constant τ0 in eq 1;
filled symbols are obtained by taking into account the absorption of
the unknown species for the calculation of τ0 using eq 2.
lifetime of HO2 under our conditions (several 10 ms; Figure 2).
Therefore, line (b) does not contain any absorption from the
unknown species; only the HO2 spectrum is obtained. The lower
line (c) on the other hand contains the sum of the absorption
both from the unknown species and from HO2. The difference
between lines (b) and (c) corresponds to the spectrum of the
unidentified species and is shown as line (d) of the graph (right
y-axis applies): it can be seen that the difference between both
spectra is constant (≈1.5 μs) over the entire wavelength range,
except for a slight maximum around the HO2 absorption lines at
6638.11 and 6638.20 cm−1. This is probably due to a slightly
higher, average HO2 concentration in the time window 0−1 ms
compared to the HO2 concentration in the time window
1−50 ms. From this result we have decided that it is reasonable
to use the time-resolved evolution of the ring-down time,
measured at 6638.30 cm−1 (dashed, vertical line in Figure 4),
as baseline for converting the ring-down times obtained at
6638.20 cm−1 into HO2 concentrations.
energies. The open symbols represent HO2 concentration−
time profiles calculated the usual way, i.e., using the same τ0 in
eq 1 for all ring-down events (obtained from the average of all
ring-down events before the photolysis pulse), whereas the
HO2 concentration−time profiles represented by the filled
symbols have been obtained by calculating an individual τ0 for
each ring-down event. It can be seen that now the curious
signal at low photolysis energy is not curious anymore but,
rather, has the typical shape. Again, these two signals are
extreme cases and have been chosen for illustration. In Figure
2a it can be seen that the effect is much smaller for benzene:
open symbols and dashed lines show uncorrected HO2
concentration−time profiles, filled symbols and full lines
represent the baseline corrected profiles.
Energy Dependence of HO2 Formation for Different
Aromatic Hydrocarbons. Having realized that the absorp-
tion of the short-lived species has to be taken into account as
“time-dependent baseline” and has thus an impact on the yield
of the HO2 formation, we have repeated all experiments,
including benzene, now carefully measuring an offline signal
before or after each online measurement. Though the impact
was most important for o-xylene, it had also some
consequences for the reinterpretation of the benzene experi-
ments, already published recently.17 Systematic analysis of all
baseline signals (see further down) showed that the intensity of
the unknown absorption α, i.e., (τt=∞ − τt=0) in eq 2, rapidly
approached saturation with increasing photolysis energy (see
the small increase from 10 to 35 mJ cm−1 in Figure 3 and
the dependence of α on photolysis energy, Figure 10). As a
consequence, the baseline correction changes the HO2
concentration (decreasing the initial and increasing the
secondary) roughly by the same absolute number, independent
of the photolysis energies. In Figure 6 are shown the results
for the benzene experiment from Figure 2a with and without
baseline correction: the HO2 concentration−time profiles in
Figure 2a have been fitted (dashed lines for uncorrected, full
This baseline correction has been carried out by a Labview
program as follows: offline signals such as shown in Figure 3
have been fitted to an exponential function:
τ = τ + (τ
t=∞
− τt=0) × (1 − e−k×t
)
(2)
t
t=0
where k is the decay constant of the short-lived, absorbing
species, τt=0 is the ring-down time just after the photolysis pulse,
τt=∞ is the ring-down time at very long delays (identical to the
conventional τ0, i.e., the ring-down time just before the laser
pulse). In practice, the value of τt=∞ has first been calculated
from the average of all data points before the photolysis pulse
and is then fixed as a constant in the baseline fit. In Figure 3 it
can be seen that the ring-down times before the laser pulse are
slightly lower for the online measurements compared to the
offline measurements (0.05 μs in the example of Figure 3).
This shift is due to an absorption of the aromatic hydrocarbon,
slightly different for the online and offline wavelengths: to
correct for this absorption, the Labview program calculated
this difference and deducted it from the τt=∞ value. These
6234
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239

The Journal of Physical Chemistry A
Article
From these fits, the initial, prompt HO2 concentration
[HO2]ini can be obtained as the intercept whereas [HO2]sec is
the concentration of the HO2 radicals, formed in a slow
reaction on a longer time scale. Although the impact on the
secondary HO2 is very minor, the yield of the initial HO2
radical experiences a fundamental change: without baseline
correction, the increase of the yield with photolysis energy
seems linear and is thus in agreement with a 1-photon process,
whereas after baseline correction this yield is not anymore
linear with energy. This energy dependence study has been
extended to three other aromatic compounds: toluene,
o-xylene, and mesitylene. The results for all hydrocarbons are
in good agreement with a two photon processes being at the
origin of both HO2 radicals, initial and secondary: the results
are summarized in Figure 7 with full lines showing a forced
square-fit for the yields of both initial and secondary HO2 to
the equation:
Figure 6. Plot of [HO2] against photolysis energy with and without
correction obtained for benzene. Experiments were carried out with
[C6H6] = 3.35 × 1015 cm−3 in the presence of [O2] = 1.0 × 1017 cm−3.
Open symbols/dashed lines: without baseline correction. Filled
symbols, full lines: with baseline correction. Error bars represent the
statistical errors only (95% confidence interval).
[HO2] = a × E2
(4)
with E being the photolysis energy in mJ cm−2.
Table 1 lists the parameters a from these fits for both the
initial and secondary HO2 yields for six different series of
experiments. To compare the results by accounting for the
different initial aromatic hydrocarbon concentration, the a
values have been divided by this concentration. Except for the
only series with a low O2 concentration ([O2] = 1.12 × 1016
cm−3), the HO2 yields do not seem to be strongly dependent
on the nature of the aromatic hydrocarbon. We propose for
lines for corrected profiles) to the following equation (for
details see Jain et al.17):
slowt
[HO2] = [HO2]ini × e−k + [HO2]sec
kfast
kslow − kfast
fastt
slowt
×
(e−k − e−k
)
(3)
Figure 7. Results of energy dependence studies of all four different aromatic compounds in a log−log plot. Full lines represent square dependence of
HO2 yields. Error bars in all figures represent the statistical errors only (95% confidence interval).
6235
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239

The Journal of Physical Chemistry A
Article
a
Table 1. Factor a from Eq 4 per Molecule of Aromatic Hydrocarbon for the Initial and Secondary HO2
initial HO2
secondary HO2
compound
(a/[aromatic])/10−6
(a/[aromatic])/10−6
no. of measurements
[aromatic]/1015 cm−3
[O2]/1017 cm−3
benzene
benzene
benzene
toluene
0.48 0.05
0.58 0.06
0.28 0.03
0.48 0.05
0.45 0.07
0.36 0.05
0.47
1.3 0.06
1.8 0.08
0.73 0.03
2.0 0.3
2.0 0.3
1.3 0.3
1.68
4
6
6
5
7
5
3.3
1.0
0.65
3.4
1.3
0.112
1.0
1.0
o-xylene
mesitylene
average
0.315
1.0
1.0
1.15
a
Values have been obtained by fitting the experimental data to eq 4. Error represents 95% confidence interval.
practical purposes to estimate the concentration of HO2
radicals, formed per molecule of aromatic hydrocarbon excited
at 248 nm, for any aromatic hydrocarbon [A] using the average
value of all series in Table 1 (except low O2):
[HO2]/[A] = ([HO2]ini + [HO2]sec)/[A]
= (0.47 + 1.68) × 10−6 × E2
([5])
with E being the energy of the 248 nm laser pulse in mJ cm−2.
This empiric equation can be used to estimate the
concentration of HO2 radicals that has been generated in
former experiments using the 248 nm photolysis of H2O2 as
OH precursor for studying the OH-initiated degradation of
aromatic hydrocarbons: Bohn and Zetzsch5 used a laser fluence
of E = 1.5 mJ cm−2 and initial concentrations of [C6H6] ≈ 5 ×
1015 molecules cm−3, leading to the formation of unwanted
[HO2] = 2.9 × 1010 cm−3. To evaluate the possible impact on
the reaction mechanism, this concentration needs be compared
with the expected radical concentration from H2O2 photolysis:
their typical [OH] concentration has been 1.7 × 1011 cm−3,
obtained from the simultaneous photolysis of 5 × 1014 cm−3
H2O2 under the same conditions. Therefore, the unwanted
HO2 concentration is close to 20% of the total radical con-
centration. The situation is much worse for the experimental
conditions used in the work of Johnson et al.,8 where much
higher photolysis energies (30−40 mJ cm−2) were employed.
The calculated unexpected [HO2] is 2 × 1013 and 3.5 × 1013
cm−3 for 30 and 40 mJ cm−2, respectively, for their highest
C6H6 concentration, to be compared to their typical initial
OH concentration of 3.5 × 1013 and 4.6 × 1013 cm−3: under
these conditions, the concentration of the parasite HO2 radicals
comes close to the concentration of the expected OH radicals.
A similar situation can be found in the work of Johnson et al.6
on the measurement of the UV-absorption spectrum of methyl-
substituted hydroxylcyclohexadienyl radicals. Taking into
account the fact that each HO2 radical is accompanied by at
least one other radical, the concentration of unwanted (and
unaccounted) radicals is higher than the known radical
concentration.
Figure 8. Decay rate kslow of HO2 radicals as a function of sum of
[HO2]ini + [HO2]sec for different aromatic compounds. Error bars
represent the statistical errors only.
o-xylene is much faster: (1.1 0.1) × 10−10 cm3 s−1, too fast for
classical HO2 + RO2 reactions.
The origin of both types of HO2 radicals is not clear. The
prompt HO2 is necessarily formed from a very fast reaction
with O2, because we are not able to resolve the increase of this
prompt HO2 on our time scale, even at the lowest O2
concentrations. The origin of the secondary HO2 has been
investigated in some detail in our earlier work on benzene:17 we
concluded that the reaction of H-atoms with O2 is at least partly
responsible, because we observed a strong pressure dependence
of the rate of formation ((2.1 0.2) and (8.7 1.2) × 10−14
cm3 s−1 at 22 and 102 Torr, respectively). H-atoms had also
been observed as product of 2-photon absorption by Kovacs
et al.,16 however, at a very low yield not in line with our
observations. In this study, we have measured the rate of
formation as a function of the O2 concentration for different
aromatic hydrocarbons at 40 Torr He. The obtained depend-
ence of kfast as a function of O2 is shown in Figure 9: it can be
seen that the formation rate for the secondary HO2 radical is
not dependent on the structure of the aromatic hydrocarbon.
Also, the initial concentration or the photolysis energy does not
have a noticeable impact on the formation rate. The rate
constant, obtained from a linear regression to all data points, is
(4.7 0.4) × 10−14 cm3 s−1. Again, this rate is slightly higher
than the expected value for the reaction of H-atoms with O2
from literature data27 but is in excellent agreement with the
pressure-dependent rate from our earlier work (interpolation to
40 Torr of Jain et al.17 leads to 4.9 × 10−14 cm3 s−1 prior to
correction for the role of O2 as a third body in the H + O2
reaction; see ref 17 for details).
It has already been stated in our earlier work17 that the decay
rate of the HO2 radicals depended on the photolysis energy, i.e.,
the absolute radical concentration. We have confirmed this
behavior for all aromatic hydrocarbons, as can be seen in
Figure 8: the exponential decay rates such as obtained from the
fits to eq 3 over the first 5 ms are plotted against the sum of
primary and secondary HO2 radicals, serving as proxy for the
accompanying RO2 radicals. The obtained rate constants for
benzene, toluene, and mesitylene are similar ((1.5
0.1),
(1.2 0.2), and (2.3 0.2) × 10−11 cm3 s−1), only the rate for
6236
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239

The Journal of Physical Chemistry A
Article
hydrocarbon concentration) is plotted as a function of the
excitation energy for all four hydrocarbons. It can be seen what
has already been mentioned earlier: the absorption is very
strong for o-xylene, whereas it is minor for benzene. What can
also be seen from Figure 10 is that the intensity rapidly
saturates with increasing excitation energy, which has also been
mentioned already and can be seen in Figure 3. From this
behavior we suggest that the absorbing species originates from
the absorption of one photon: the absorption cross section for
the second photon at 248 nm is much higher than for the first
photon;16 therefore, the reservoir of 1-photon excited aromatic
hydrocarbons will be more and more depleted with increasing
excitation energy. A simple model considering a sequential
absorption of both photons and a σ2‑photon = 2 × 10−17 cm2
shows that at 10 mJ cm−2 78% of the species having initially
absorbed 1-photon are still available, whereas at 40 mJ cm−2
this theoretical rate falls to 37%. This simple model thus
reproduces very well the observations made in this study
(Figure 10). In Figure 10 (upper part) it can be seen that the
decay rate slightly increases with the photolysis energy,
indicating a very fast reaction with other photolysis products:
the rate constant of this reaction is close to collision frequency,
when considering formation of one active species per absorbed
photon. Figure 11 shows the dependence of α with the initial
Figure 9. Dependence of kfast as obtained from eq 3 for dif-
ferent aromatic hydrocarbons. Total pressure was 40 Torr for all
experiments.
The results from this work, now confirming two earlier
aromatic hydrocarbon studies on the unexpected formation of
HO2 radicals after the 248 nm excitation of benzene,15,17 might
have some impact on the decay kinetics measured in the above-
mentioned studies.4−8,28,29 This could especially have an impact
on the determination of rate constants for radical−radical
reactions such as the self-reaction of aromatic−OH adducts or
the reaction of aromatic−OH adduct with HO2 radicals. In his
work on the formation of peroxy radicals from the toluene−
OH adduct with O2, Bohn4 had to consider a very fast rate
constant of 1.8 × 10−10 cm3 s−1 for the reaction of the toluene−
OH adduct with HO2 to correctly reproduce his experimental
results. Therefore, the presence of even low, unaccounted
concentrations of HO2 radicals might have an impact on the
observed decays and a re-evaluation of former experiments
might be interesting.
Characterization of the Absorption from the Short-
Lived Species. With the wish to better identify the unknown
absorbing species, we have evaluated the two parameters, k and
α, obtained from the baseline fits to eq 2, as a function of
different parameters. The intensity α has been calculated using
eq 1 with σ = 1, τt = (τt=∞ − τt=0), and τ0 = τt=∞. In Figure 10
the intensity α (for comparison divided by the initial aromatic
Figure 11. Absorption of the unknown reactive intermediate
absorption for toluene and o-xylene concentrations (right y-axis) and
decay rate of the unknown reactive intermediate (left y-axis) as a func-
tion of initial concentration, photolysis energy was 35 and 62 mJ cm−2
for o-xylene and toluene, respectively.
aromatic hydrocarbon concentration: a rapid saturation of the
intensity is visible with increasing concentration, being visible
already at the lowest hydrocarbon concentration (the signal
decreases to zero very shortly after shutting off the hydro-
carbon flow). Figure 11 shows in the upper part that the decay
rate k increases linearly as a function of the aromatic
hydrocarbon concentration: the rate constant for the quenching
process, obtained for toluene and o-xylene is 4.6 and 7.2 ×
10−12 cm3 s−1, respectively. These values are on the order of
magnitude for the rate constant that has been found for the
quenching of triplet benzene with ground state benzene:30
1.1 × 10−13 cm3 s−1. In the same work, the decay of triplet
benzene has been found to be 5 × 103 s−1, the same order of
magnitude as the decay rates found in this work for the
unknown species. Figure 12 finally shows the evolution of α
and k with O2: though the intensity increases linearly with O2,
the decay rates do not depend on O2 concentration. O2 is
Figure 10. Absorption of the unknown reactive intermediate
absorption measured at 6638.30 cm−1 (right y-axis) and decay rate
of the unknown reactive intermediate (left y-axis) as a function of
photolysis energy. Initial hydrocarbon concentrations: 0.3, 1.0, 1.0, and
3.2 × 1015 cm−3 for o-xylene, toluene, mesitylene, and benzene,
respectively.
6237
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239

The Journal of Physical Chemistry A
Article
the French Research Ministry and by the European Fund for
Regional Economic Development (FEDER). C.J. thanks the
EU for financial support through project MEST-CT-2005-
020659.
REFERENCES
■
(1) Calvert, J. G.; Atkinson, R.; Becker, K. H.; Kamens, R. M.;
Seinfeld, J. H.; Wallington, T. H.; Yarwood, G. Oxford University
Press: Oxford, U.K., 2002.
(2) Lonneman, W. A.; Bellar, T. A.; Altshuller, A. P. Environ. Sci.
Technol. 1968, 2, 1017−1020.
(3) Nehr, S.; Bohn, B.; Fuchs, H.; Hofzumahaus, A.; Wahner, A. Phys.
Chem. Chem. Phys. 2011, 13, 10699−10708.
(4) Bohn, B. J. Phys. Chem. A 2001, 105, 6092−6101.
(5) Bohn, B.; Zetzsch, C. Phys. Chem. Chem. Phys. 1999, 1, 5097−
5107.
Figure 12. Absorption of the unknown reactive intermediate (6638.30
cm−1) as a function of O2 concentrations (left y-axis) and dependence
of the decay rate of the unknown reactive intermediate (right y-axis) as
a function of the O2 concentration. Initial aromatic hydrocarbon
concentrations were 0.2, 1, and 2.2 × 1015 cm−3 for o-xylene, toluene,
and benzene, respectively.
(6) Johnson, D.; Raoult, S.; Lesclaux, R.; Krasnoperov, L. N. J.
Photochem. Photobiol., A 2005, 176, 98−106.
́
(7) Estupinan, E.; Villenave, E.; Raoult, S.; Rayez, J. C.; Rayez, M. T.;
̃
Lesclaux, R. Phys. Chem. Chem. Phys. 2003, 5, 4840−4845.
(8) Johnson, D.; Raoult, S.; Rayez, M.-T.; Rayez, J.-C.; Lesclaux, R.
Phys. Chem. Chem. Phys. 2002, 4, 4678−4686.
known to be a good quencher for singlet benzene, leading to an
increased yield in triplet benzene.31 The increase in intensity is
therefore in agreement with triplet benzene being involved in
the formation of the absorbing species. However, the
quenching of triplet benzene with O2 has been found31 to be
2.1 × 10−11 cm3 s−1, which is not at all in agreement with the
independence of the decay rate with O2: even under the lowest
O2 concentrations, the decay should already be 2 × 105 s−1,
around 100 times faster than our observation. Therefore, triplet
benzene itself cannot be the absorbing species; it is, however, in
line with our observations that triplet benzene is a precursor for
the absorbing species: a possible candidate could be vibration-
ally excited benzene in the ground state obtained through
quenching of the triplet state.
(9) Nishino, N.; Arey, J.; Atkinson, R. J. Phys. Chem. A 2010, 114,
10140−10147.
(10) Noda, J.; Volkamer, R.; Molina, M. J. J. Phys. Chem. A 2009, 113,
9658−9666.
(11) Volkamer, R.; Klotz, B.; Barnes, I.; Imamura, T.; Wirtz, K.;
Washida, N.; Becker, K. H.; Platt, U. Phys. Chem. Chem. Phys. 2002, 4,
1598−1610.
(12) Klotz, B.; Volkamer, R.; Hurley, M. D.; Andersen, M. P. S;
Nielsen, O. J.; Barnes, I.; Imamura, T.; Wirtz, K.; Becker, K.-H.; Platt,
U.; Wallington, T. J.; Washida, N. Phys. Chem. Chem. Phys. 2002, 4,
4399−4411.
(13) Bloss, C.; Wagner, V.; Jenkin, M. E.; Volkamer, R.; Bloss, W. J.;
Lee, J. D.; Heard, D. E.; Wirtz, K.; Martin-Reviejo, M.; Rea, G.;
Wenger, J. C.; Pilling, M. J. Atmos. Chem. Phys. 2005, 5, 641−664.
(14) Bloss, C.; Wagner, V.; Bonzanini, A.; Jenkin, M. E.; Wirtz, K.;
Martin-Reviejo, M.; Pilling, M. J. Atmos. Chem. Phys. 2005, 5, 623−
639.
CONCLUSIONS
■
The energy-dependent study of the formation of HO2 radicals
after 248 nm irradiation of four different aromatic hydrocarbons
in the presence of O2 has revealed that the formation of the
HO2 radicals involves a two photon excitation processes. Our
earlier work on HO2 formation from excited benzene17 had to
be revised in this study: the formation of an unknown, short-
lived species, absorbing at the same wavelength than HO2 has
been taken into account as baseline for the calculation of HO2
concentrations. The intensity as well as the decay time of the
absorption of the unknown species has been analyzed to
identify its nature. However, no explanation could be found
that would explain all characteristics of this absorption. An
empiric equation has been proposed to allow an estimation of
the impact of this HO2 formation for all aromatic hydrocarbons
for past and future laboratory studies.
(15) Aluculesei, A.; Tomas, A.; Schoemaecker, C.; Fittschen, C. Appl.
Phys. B: Lasers Opt. 2008, 92, 379−385.
(16) Kovacs, T.; Blitz, M. A.; Seakins, P. W.; Pilling, M. J. J. Chem.
Phys. 2009, 131, 204304.
(17) Jain, C.; Parker, A. E.; Schoemaecker, C.; Fittschen, C.
ChemPhysChem 2010, 11, 3867−3873.
(18) Thiebaud, J.; Fittschen, C. Appl. Phys. B: Lasers Opt. 2006, 85,
383−389.
(19) Parker, A.; Jain, C.; Schoemaecker, C.; Szriftgiser, P.; Votava, O.;
Fittschen, C. Appl. Phys. B: Lasers Opt. 2011, 103, 725−733.
́
(20) Votava, O.; Masat, M.; Parker, A.; Jain, C.; Fittschen, C. Rev. Sci.
Instrum. 2012, DOI: http://dx.doi.org/10.1063/1.3698061.
(21) Thiebaud, J.; Crunaire, S.; Fittschen, C. J. Phys. Chem. A. 2007,
111, 6959−6966.
(22) Ibrahim, N.; Thiebaud, J.; Orphal, J.; Fittschen, C. J. Mol.
Spectrosc. 2007, 242, 64−69.
(23) Tang, Y.; Tyndall, G. S.; Orlando, J. J. J. Phys. Chem. A 2010,
114, 369−378.
AUTHOR INFORMATION
Corresponding Author
*E-mail: Christa.fittschen@univ-lille1.fr.
■
(24) Trost, B.; Stutz, J.; Platt, U. Atmos. Environ. 1997, 31, 3999−
4008.
(25) Fally, S.; Carleer, M.; Vandaele, A. C. J. Quant. Spectrosc. Radiat.
Transfer 2009, 110, 766−782.
Notes
The authors declare no competing financial interest.
(26) Jain, C.; Morajkar, P.; Schoemaecker, C.; Viskolcz, B.; Fittschen,
C. J. Phys. Chem. A 2011, 115, 10720−10728.
(27) Fernandes, R. X.; Luther, K.; Troe, J.; Ushakov, V. G. Phys.
Chem. Chem. Phys. 2008, 10, 4313−4321.
ACKNOWLEDGMENTS
■
This work is supported by the CNRS through the French-
German program “ATMOCHEM” and by the Nord-Pas de
Calais region in the frame of the IRENI research program, by
(28) Berho, F.; Rayez, M.-T.; Lesclaux, R. J. Phys. Chem. A 1999, 103,
5501−5509.
6238
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239

The Journal of Physical Chemistry A
Article
(29) Raoult, S.; Rayez, M.-T.; Rayez, J.-C.; Lesclaux, R. Phys. Chem.
Chem. Phys. 2004, 6, 2245−2253.
(30) Ueno, T.; Kouchi, N.; Takao, S.; Hatano, Y. J. Phys. Chem. 1978,
82, 2373−2377.
(31) Morikawa, A.; Cvetanovic, R. J. J. Chem. Phys. 1970, 52, 3237−
3239.
6239
dx.doi.org/10.1021/jp211520g | J. Phys. Chem. A 2012, 116, 6231−6239