OH formation in the photoexcitation of NO2 beyond the dissociation threshold in the presence of water vapor

Article abstract of DOI:10.1021/jp970319e

The pulsed-laser-excitation/resonance fluorescence technique was used to assess the efficiency of OH formation following photoexcitation of NO₂ at discrete wavelengths beyond the photodissociation threshold in the presence of water vapor: NO₂* + H₂O → HONO + OH. Excitation at wavelengths between 432 and 449 nm was found to lead to OH production via a facile sequential two-photon absorption by NO₂, leading to O(¹D) and thus to OH in the presence of H₂O, i.e., NO₂ + hv → NO₂*, NO₂* + hv → NO₂**, NO₂** → NO + O(¹D), O(¹D) + H₂O → 2OH. The cross section for the transition NO₂** ← NO₂* was found to be similar to that for the NO₂* ← NO₂ transition at 435 nm. At 532 nm, the two-photon process is not sufficiently energetic to form I(¹D) and OH is not observed. An upper limit of approximately 7 × 10^-5 is found for the reactive quenching of NO₂* by water vapor relative to collisional quenching. The atmospheric relevance of OH formation via NO₂ excitation is discussed.

Full text of DOI:10.1021/jp970319e

4178
J. Phys. Chem. A 1997, 101, 4178-4184
OH Formation in the Photoexcitation of NO2 beyond the Dissociation Threshold in the
Presence of Water Vapor
John N. Crowley* and Shaun A. Carl
Max-Planck Institut fu¨r Chemie, DiVision of Atmospheric Chemistry, Postfach 3060, 55020 Mainz, Germany
ReceiVed: January 24, 1997; In Final Form: April 3, 1997X
The pulsed-laser-excitation/resonance fluorescence technique was used to assess the efficiency of OH formation
following photoexcitation of NO2 at discrete wavelengths beyond the photodissociation threshold in the presence
of water vapor: NO2* + H2O f HONO + OH. Excitation at wavelengths between 432 and 449 nm was
found to lead to OH production via a facile sequential two-photon absorption by NO2, leading to O(1D) and
thus to OH in the presence of H2O, i.e., NO2 + hν f NO2*, NO2* + hν f NO2**, NO2** f NO + O(1D),
O(1D) + H2O f 2OH. The cross section for the transition NO2** r NO2* was found to be similar to that
for the NO2* r NO2 transition at 435 nm. At 532 nm, the two-photon process is not sufficiently energetic
to form O(1D), and OH is not observed. An upper limit of approximately 7 × 10-5 is found for the reactive
quenching of NO2* by water vapor relative to collisional quenching. The atmospheric relevance of OH
formation via NO2 excitation is discussed.
Introduction
The OH radical is the primary oxidant in the daytime
troposphere, and its concentration determines the lifetimes and
thus the abundance of most natural and anthropogenically
emitted chemical species,1 including hydrocarbons, CO, NO2,
and SO2.
The photolysis of O3 in the presence of water vapor is the
major source of the OH radical.
O3 + hV f O2 + O(1D)
O(1D) + H2O f 2OH
(1)
(2)
(3)
O(1D) + M f O(3P) + M
The production rate of OH via O3 photolysis is given by
ROH ) 2J(O3)[O3]/(1 + k3[M]/k2[H2O])
Figure 1. The UV/vis absorption spectrum (Sigma) of NO2 multiplied
by the quantum yield (QY) for dissociation to O(3P) + NO(2Π) (dotted
line) and by the quantum yield for photoexcitation to NO2* (solid line).
(i)
Additional reactions that oxidize NO to NO2, e.g., with peroxy
radicals,
where J(O3) is the photolysis rate constant for O3 and is equal
to about 1 × 10-6 s-1 at high latitudes in winter, k3 is the rate
constant for collisional deactivation of O(1D) by air, and k2 is
the rate constant for reaction of O(1D) with H2O. This
expression yields OH production rates of 2 × 105 OH cm-3
s-1 if [O3] ) 0.05 ppm, and an H2O pressure of 10 Torr is
assumed.
NO + RO2 f NO2 + RO
(7)
result in a net production of O3 from NO2 photolysis. The above
discussion thus shows how OH, O3, and NO2 are closely linked
in fast photochemical cycles.
The conditions under which O3 loss or production terms
dominate have been investigated in model calculations of the
global troposphere,3 and the critical concentration of NO at
which O3 formation is favored is estimated as about 10-20 ppt
([NOx] ) 40 ppt). Such low concentrations of NOx are usually
only found in very clean air. Recent estimates4 suggest that
about 65% of global tropospheric ozone is formed by the
oxidation of NO by HO2 and the subsequent photolysis of NO2.
The absorption spectrum of NO2 displays two broad continua
centered at ca. 400 and 210 nm with underlying fine structure.
In the troposphere the filtering effect of O3 means that only
light of wavelengths longer than about 300 nm is intense enough
to initiate photochemistry. This paper therefore concentrates
on excitation in the long wavelength continuum of NO2 between
about 400 and 600 nm.
Tropospheric O3 concentrations are influenced by fast gas-
phase photochemical cycles2 involving NO2. NO2 absorbs
sunlight strongly and is in rapid photochemical equilibrium
(photostationary state) with NOx and O3 (NOx ) NO + NO2).
NO2 + hV f NO + O
O + O2 + M f O3 + M
NO + O3 f NO2 + O2
(4)
(5)
(6)
The steady-state O3 concentration is given by
[O3] ) J(NO2)[NO2]/k6[NO]
(ii)
where J(NO2) is the rate constant for photolysis to O(3P) +
NO (4).
The thermodynamic threshold for dissociation of NO2 into
X Abstract published in AdVance ACS Abstracts, May 1, 1997.
O(3P) atoms and NO is 25 130 cm-1, corresponding to a
S1089-5639(97)00319-8 CCC: $14.00 © 1997 American Chemical Society

OH Formation in the Photoexcitation of NO2
J. Phys. Chem. A, Vol. 101, No. 23, 1997 4179
can be carried into the reaction with H2O (i.e., no internal
relaxation before collision), we find that the generation of OH
radicals and HONO becomes a thermodynamically feasible
reactive quenching mechanism (∆Hr(10) ) -25 kcal/mol)
NO2* + H2O f OH + HONO
(10)
Note that ∆Hr(10) is calculated for products formed in the
electronic ground state, OH(2Π) and HONO(1A), whereby spin
is conserved. Even at wavelengths greater than 600 nm, the
photon energy is sufficient to make reaction 10 exothermic. By
contrast, the reaction of ground-state NO2 with H2O to yield
HONO and OH is 40 kcal/mol endothermic.
Some evidence already exists for the enhanced reactivity of
electronically excited NO2, which reacts with ground-state NO2
Figure 2. J values for photodissociation (left of solid vertical line)
and for photoexcitation (right of solid vertical line). J values are given
for zenith angles (Z) of between 20° and 70° and at altitudes of 0-10
km.
12-14
15,16
to give either 2NO + O2
or NO3 + NO.
NO2* + NO2 f 2NO + O2
f NO3 + NO
(11a)
(11b)
wavelength of 398 nm.5 The experimental observations indicate
that O atom production takes place out to 424 nm, where excess
energy is provided by a combination of rovibrational excitation
and energy transfer from bath-gas molecules.6 The complexity
of the NO2 spectrum is due to strong coupling between the 2A1
ground state and the 2B1 and 2B2 excited states, yielding optically
excited states that have considerable ground electronic state
character. The high density of states results in an excited-state
radiative lifetime (τrad) of between 30 and 200 µs,7-9 depending
upon excitation wavelength. In contrast, the lifetime of the
dissociative state10 is close to 1 × 10-13 s. Figure 1 shows the
absorption spectrum of NO2 between 250 and 650 nm,
multiplied by the wavelength-dependent quantum yields.6 The
dotted line thus represents dissociation to NO and O(3P) and
the solid line excitation to longer lived excited states which we
designate as NO2*.
Both of these reactions are endothermic for interaction between
two ground-state NO2 species. In addition, small yields of N2O
have been observed17 in the reactive collisional quenching of
NO2* by N2. Thus, should NO2* also react with H2O, a
mechanism for the NO2-initiated generation of OH radicals in
the troposphere is available that bypasses the formation of O3
and its photolysis to O(1D), which in the winter months at low
zenith angles is slow due to the reduced solar flux at the shorter
wavelengths.
Under these circumstances, the rate of OH production is given
by
ROH ) Jex(NO2)[NO2]/(1 + k9[M]/k10[H2O])
(iii)
where Jex(NO2) is the rate constant for nondissociative excitation
of NO2, k9 is the quenching rate constant of NO2* by air, and
k10 is the rate constant for reaction of NO2* with H2O.
NO2 + hV f NO2*
(8)
Assuming a NO2 concentration of 1 ppb and an O3 concen-
tration of 0.05 ppm, an excitation rate that is 5 times the
photolysis rate, Jex(NO2) ) 2 × 10-2 s-1, and that HONO is
also rapidly photolyzed to OH and NO, then the rate constant
for NO2* + H2O (k10 in the above text) which is necessary to
generate OH at the same rate as O3 photolysis is found to be 4
× 10-13 cm3 molecule-1 s-1. This rate constant is about 0.2%
of the quenching rate constant for NO2* with H2O. Thus, if
only 2 per mil of excited NO2 is reactively quenched by H2O
(10) rather than collisionally quenched (12), the reaction between
NO2* and H2O would represent an equally important source of
OH in the (weakly illuminated) winter troposphere as the
photolysis of O3 and the subsequent reaction of O(1D) with H2O.
Although the integrated areas of the dissociative and non-
dissociative parts of the spectrum are similar, the solar spectrum
shows a strong wavelength dependence in this region with a
large increase in ground level irradiance going from 320 to 450
nm when the solar zenith angles are large (e.g., in the winter
troposphere). The relative rates of photodissociation (and thus
O(3P) production) to photoexcitation were calculated by con-
voluting the dotted and solid parts of the NO2 spectrum of Figure
1 with the solar spectrum.11 For a zenith angle of >70°, the
photoexcitation of NO2 is calculated to proceed at a rate about
5 times that of photodissociation. The results are summarized
in Figure 2. The fate of the excited NO2* will be determined
by its fluorescence lifetime (τrad) or quenching by the major
bath gases N2, O2, or H2O. The quenching rate constants have
been measured as 2.7 × 10-11, 3.0 × 10-11, and 1.7 × 10-10
cm3 molecule-1 s-1 for N2, O2, and H2O, respectively.5 At
ground level the total quenching rate is 7.1 × 108 s-1, resulting
in a NO2* lifetime of about 1.4 ns which is thus controlled
almost exclusively by collisional quenching.
NO2* + H2O f NO2 + H2O
(12)
In addition, the rate of NO2* formation via NO + O3 will
increase the rate of NO2* formation by about 10% under the
conditions of NO2 and O3 given above and assuming about 50
ppt NO.
Photochemical experiments in environmental smog chambers
have been carried out for many years with the goal of deriving
reaction mechanisms for tropospheric oxidation processes and
to validate computer models of air pollution. As in the
atmosphere, OH radicals are often the primary oxidant. Several
publications have drawn attention to unknown radical sources
in smog chamber experiments, which have been attributed to
the formation (and subsequent photolysis) of HONO via a
surface reaction of NO2 and H2O18 or between HNO3 and NO.19
NO2* + M f NO2
(9)
As the quenching rate constants show, H2O is a particularly
efficient quencher of NO2* and, despite its lower concentration
when compared to N2 and O2 in the troposphere, contributes
about 8% to the total quenching rate if we assume 150 Torr of
O2, 600 Torr of N2, and 10 Torr of H2O. A NO2* molecule
that has absorbed a photon of wavelength 440 nm has 65 kcal/
mol energy over that of ground-state NO2. If all of this energy

4180 J. Phys. Chem. A, Vol. 101, No. 23, 1997
Crowley and Carl
Experiments have also shown that not only the rate of OH
formation but that of HONO formation is dependent on the
intensity of irradiation by UV-vis light.20 The explanation
given requires a photoenhancement in the rate of the hetero-
geneous NO2 + H2O reaction taking place on the walls of the
reactor. We note that reaction 10 would provide an alternative
mechanism for HONO and OH formation as observed in those
experiments.
To investigate the possibility that reactive quenching of NO2*
by H2O to form OH and HONO can play a role in the
troposphere (and in smog chamber experiments), we have
carried out experiments to investigate the formation of OH in
the pulsed-laser photoexcitation of NO2 in the presence of H2O
vapor.
laser pulse was calculated from the concentration of NO2, the
measured photon fluence, and the known absorption cross
section at the excitation wavelength.
OH generated in the laser pulse were excited using an OH
resonance lamp to promote the OH(2Σ+) r OH(2 Π) transition
at ≈309 nm. The mildly focused light from the resonance lamp
crossed the laser beam at the focal point of a telescopic lens
system which directed resonantly scattered photons to a photon-
counting photomultiplier. The calibration of OH was achieved
by replacing the NO2/H2O/Ar flow with one of N2O/H2O/Ar at
the same flow rate and total pressure of H2O and Ar to preserve
the resonance fluorescence sensitivity, which is dependent on
the quenching of OH(2Σ+). The N2O/Ar/H2O mixture, contain-
ing about 3 × 1015 molecules cm-3 of N2O, was photolyzed at
193 nm using an identical beam geometry as in the 430-450
and 532 nm experiments. Using the well-known N2O cross
section at 193 nm, and a quantum yield for OH formation of 2,
the concentration of OH could be determined via the calculated
extent of N2O dissociation. This required knowledge of the
laser fluence at 193 nm, which was measured using the joule
meter. A small correction (<5%) was applied to take into
account the loss of OH via reaction of O(1D) with N2O (k2/k13
) 1.9 22) and also the photolysis of H2O.
Experimental Section
Formation of OH in the reaction between NO2* and H2O was
investigated by the method of pulsed-laser-excitation/resonance
fluorescence. Pulsed, tunable radiation of wavelengths between
430 and 450 nm (line width ) 0.15 cm-1) was obtained by
pumping a Lambda Physik Scanmate-2 dye laser with either a
Nd:YAG laser (Quantel Brilliant B, pulse width ) 5 ns) or with
an excimer laser (Lambda Physik Lextra 50, pulse width ) 20
ns). The dye used, coumarin 120, was excited at 355 nm (Nd:
YAG pump) or 351 nm (XeF excimer pump) at repetition rates
of between 5 and 10 Hz. The dye laser beam was expanded to
5 mm diameter and reduced to ca. 4 mm diameter by a set of
irises before entering the Teflon-coated stainless steel reaction
vessel21 through quartz windows at the Brewster angle. The
intensity was monitored with a calibrated joule meter upon
exiting through a second Brewster window.
The pulse energy of the pump laser was varied to provide
photon fluences of typically between 3 and 14 mJ/cm2 for
expanded laser beams of 4 mm diameter. Some experiments
were also carried out at 532 nm by coupling the second
harmonic of the Nd:YAG laser directly into the reaction vessel
through the same set of irises. For those experiments in which
the Nd:YAG laser was used to pump the dye laser, the excimer
laser was run with ArF to generate 193 nm pulses. The 193
nm radiation was coupled into the reaction vessel via a beam
splitter, through the same irises, and was used to dissociate N2O
in the presence of H2O vapor to provide a calibration signal for
OH (see below).
NO2 was diluted as a ca. 5% mixture in Ar and its flow
regulated by a mass flow controller. A second, calibrated, Ar
flow was directed through a H2O-filled bubbler at room
temperature and through a glass spiral maintained at tempera-
tures between 1 and 5 K below that of the bubbler to ensure
saturation. The resultant H2O/Ar mixture was mixed with the
NO2/Ar flow and a third calibrated flow of pure Ar before being
directed into a long-path optical absorption cell connected in
series to the RF cell. Absorption of light from a halogen lamp
over a path length of 975 cm was detected by a diode-array
detector and was used to determine the NO2 concentration. In
some experiments, the attenuation of the laser light upon
addition of the NO2 flow to the reaction volume was measured
using the joule meter. In combination with the literature value
of the absorption cross section of NO2 at the excitation
wavelength and the optical path between the two Brewster
windows (≈18 cm), this enabled the NO2 concentration to be
determined. The values obtained were consistent with the [NO2]
measurement described above, but was useful only for high
[NO2] when the averaged joule meter reading was sufficiently
accurate to detect the small absorption over 18 cm. The amount
of NO2 that was promoted into the excited state following the
N2O + hν (193 nm) f N2 + O(1D)
O(1D) + H2O f 2OH
(13)
(2)
O(1D) + N2O f products
(14)
The conditions most conducive to optimize the amount of
NO2* that is quenched by H2O rather than bath gas were those
of maximum H2O vapor pressure. However, the sensitivity of
the resonance fluorescence detection of OH decreased rapidly
with increasing H2O partial pressure due to the very high
quenching rate constant23 for OH(2Σ+) with H2O of (2-5) ×
10-10 cm3 molecule-1 s-1
. Similarly, N2 is a much more
efficient quencher of OH(2Σ+) than Ar and at a pressure of 20
Torr was found to reduce the OH signal to just 20% of that
obtained in the same pressure of Ar. These considerations
therefore constrained the experimental conditions to total
pressures of Ar bath gas of about 20 Torr and H2O concentra-
tions of 2 Torr or below. The quenching rate constants for NO2*
by H2O and Ar are 1.7 × 10-10 and 1.85 × 10-11 cm3
molecule-1 s-1, respectively.5 The first-order loss rates for
NO2* is thus 2 × 104 s-1 for fluorescence (based on τrad ) 50
µs), 1.1 × 107 s-1 for quenching by 19 Torr of Ar, and 5.8 ×
106 s-1 for quenching by 1 Torr of H2O. This means that 33%
of the NO2* is quenched by H2O under the experimental
conditions. (Self-quenching by NO2 could be neglected as its
concentration is too low.) The quenching rate constants of NO2*
by Ar and He are almost identical, and quenching of OH(2Σ+)
by both gases is very small.23 Ar was therefore used in
preference of He to reduce diffusion. The typical total flow
rate was about 250 sccm (cm3 (STP) min-1), resulting in a total
pressure in the UV cell of 24 Torr and in the fluorescence cell
of 20 Torr. The residence time in the fluorescence cell was
about 1 s (linear velocity ) 8 cm s-1) and 20 s in the UV cell.
NO2, purchased as N2O4 (Merck, 99.5%), N2O (Hoechst,
99.5%), H2O (distilled, Millipore), CH4 (Linde 99.9%), and Ar
(Linde 99.999%) were used without further purification.
Results
432-449 nm Excitation. Figure 3 displays fluorescence
intensity as a function of time following the pulsed photolysis

OH Formation in the Photoexcitation of NO2
J. Phys. Chem. A, Vol. 101, No. 23, 1997 4181
Figure 3. Production and decay of OH following 439.4 nm photolysis
of a flowing NO2/H2O/Ar mixture. The solid line is a pseudo-first-
order fit to the decay as described in the text. The integration time per
channel was 25 µs; 3000 laser pulses were averaged at 10 Hz. cps )
counts per second.
Figure 5. The 420-460 nm visible absorption spectrum of NO2 (line,
left y-axis) and the OH signal height as a function of excitation
wavelength (solid circles, right y-axis). The OH signal has been
normalized for the number of photons per pulse, signal integration time,
and laser fluence squared.
at the experimental temperature and pressure in air (1.3 × 10-12
cm3 molecule-1 s-1) and confirms that the detected species is
OH.
Having established that OH is formed in the 439.4 nm
photolysis of NO2/Ar/H2O mixtures, a series of diagnostic tests
were carried out to ascertain whether reaction 10 is its source.
Figure 5 shows a plot of initial OH signal intensity as a function
of wavelength for a small section of the NO2 spectrum. The
OH signal is gained by back-extrapolation of the exponential
decay of OH to t ) 0 and has been normalized for number of
laser pulses and integration time and the laser fluence squared
(see below). The NO2 concentration was held constant at about
2.5 × 1015 molecules cm-3 for the whole experiment; that of
H2O was held at ≈0.5 Torr. The measured laser fluence at
each wavelength was converted to a number of photons/pulse
to make comparison between each wavelength meaningful.
There is clearly good correlation between the OH signal intensity
and the features of the NO2 spectrum, with maxima and minima
in the OH signal correlating with peaks and troughs of the NO2
spectrum, and confirms that the initial absorption process
responsible for OH formation is by NO2. Note however that
this good agreement was only obtained when the OH signal
was normalized to the squared laser energy; otherwise, the
features of the NO2 spectrum were reproduced, but the relative
magnitude of the signal did not match.
The energy dependence of the OH signal was therefore
measured as a function of laser fluence at a fixed wavelength
(439.4 nm) and fixed NO2 and H2O concentrations similar to
above. The results are displayed in Figure 6 in which the ln-
(relative pulse energy) is plotted against the extrapolated ln-
(signal) at t ) 0 for experiments with the Nd:YAG laser used
to pump the dye laser. The slope of the plot is 2.06 ( 0.12.
For comparison, the slope for the same plot for 193 nm
photolysis of a N2O/H2O/Ar mixture yields a slope of 1.14 (
0.15. A slope of unity is expected for a single-photon process,
whereas a slope of two is expected for a two-photon process.
Using the Nd:YAG or the excimer laser to pump the dye laser
yielded identical slopes. An approximate dependence on the
square of the initial NO2* concentration could indicate not only
two-photon absorption but also collisional interaction of two
singly excited NO2*. This could be ruled out by conducting
experiments at fixed energy and varying the NO2 concentration.
In this case the dependence on [NO2] was found to be linear as
shown in Figure 7.
Figure 4. Dependence of the pseudo-first-order decay rate (k1st) of
OH on the initial concentration of NO2. The fit to the data gives a rate
constant for OH + NO2 of (1.1 ( 0.1) × 10-12 cm3 molecule-1 s-1 at
20 Torr of Ar and at 295 K. The error bars are statistical (2σ).
(5 ns pulse) of a flowing mixture of NO2/Ar/H2O at 439.4 nm.
The total pressure was 20 Torr with 19 Torr of Ar, 1 Torr of
H2O, and [NO2] ) 3.33 × 1015 molecules cm-3. The measured
fluence was 14 mJ cm-2. The laser was triggered at time )
500 µs in this trace. Twenty channels of data (500 µs) were
acquired prior to the pulse to establish the intensity of back-
ground scattered light from the resonance lamp. Once corrected
for this background signal, the decays were found to be
exponential and the decay rate linearly dependent on the initial
NO2 concentration. Stray light following the laser pulse
precluded the analysis of the data up to 75 µs after the 20 ns
pulse was complete, and these data were not included in the
nonlinear least-squares fit to the data represented by the solid
line in Figure 3.
The background-corrected data were fit to an algorithm of
the form
[OH]t ) [OH]0 exp(-k1stt)
(iv)
where k1st is the pseudo-first-order decay constant (in s-1) of
OH and is defined as k1st ) kbi[NO2], where kbi is the
bimolecular rate constant for the reaction of OH with NO2 at
the given temperature and pressure.
A plot of pseudo-first-order decay rate versus the optically
determined [NO2]0 (Figure 4) yields a rate constant of kbi
)
(1.1 ( 0.1) × 10-12 cm3 molecule-1 s-1. This is close to the
The question remains how two-photon absorption by NO2
can lead to OH formation in the Ar/NO2/H2O mixture. Note
literature rate constant for the reaction between OH and NO2

4182 J. Phys. Chem. A, Vol. 101, No. 23, 1997
Crowley and Carl
Figure 8. OH generated in the 435 nm photolysis of NO2/H2O/Ar
(solid circles) or NO2/CH4/Ar (open squares). Concentrations are given
in the text. cps ) counts per second.
Figure 6. Dependence of [OH]0 on the laser pulse energy. Open
squares: 193 nm photolysis of N2O /H2O, slope ) 1.14 ( 0.15. Full
circles: 439.4 nm excitation of NO2/H2O/Ar, slope ) 2.06 ( 0.12.
Open circles: 439.4 nm excitation of NO2/H2/Ar, slope ) 2.00 ( 0.04.
The error bars are statistical (2σ) errors on the back-extrapolated OH
signal at t ) 0.
an experiment in which ca. 4 × 1015 molecules cm-3 of NO2
was photolyzed at 435 nm in the presence of either 0.5 Torr of
H2O and 20 Torr of Ar or 0.65 Torr of CH4 and 20 Torr of Ar
are displayed in Figure 8. The OH profiles in both experiments
are clearly very similar, and as there is no obvious route to OH
formation via a reaction of NO2* in the CH4 experiments, we
conclude that O(1D) is responsible in both cases. The almost
identical magnitude of the initial OH signal in both experiments
is thought to be fortuitous and is probably the result of the
greater quantum yield of OH production in the reaction of O(1D)
with H2O being compensated by the lower quenching rates of
NO2* and possibly also reduced quenching of OH(2Σ+) by CH4
as compared to H2O. Experiments were also carried out in
which H2O was replaced by H2. Again OH was observed, and
the dependence upon energy was determined. The plot of
relative energy versus OH signal yielded a slope of 2.00 ( 0.04,
and we conclude that the two-photon formation of O(1D) and
reaction with H2 are responsible. These data are also shown in
Figure 6. In this case a stronger OH signal is seen due to the
formation of a second OH radical by
Figure 7. Dependence of [OH]0 generated by photolysis at 435 nm
on the initial NO2 concentration. The OH signal is normalized for total
integration time (cps ) counts per second). Error bars are statistical
error (2σ) on the back-extrapolated OH signal.
O(1D) + H2 f OH + H
H + NO2 f OH + NO
(15)
(16)
that the action spectrum for OH formation matches the structure
in the NO2 absorption spectrum, implying that sequential two-
photon absorption is taking place. Simultaneous two-photon
absorption is in any case highly unlikely considering the low
photon densities in the unfocused laser beam (e.g., at 10 mJ/
cm2 and 20 ns pulse, the irradiance is 5 × 109 J s-1m-2 and the
photon density is calculated as 4 × 1013 photons cm-3). We
suggest that a transition from the initially populated, mixed
NO2*(2B1/2B2/2A1) state to the dissociative NO2**(2 2B2) state
is possible at wavelengths close to 440 nm. This is an allowed
transition due to the partial ground-state nature of the NO2*
excited state and corresponds to absorption of a single photon
Previous experiments have also identified two-photon effects
in the beyond-threshold dissociation of NO2,25,26 and have shown
2
that pulsed (10 ns) laser excitation of the B2 state (693.4 nm)
with laser energies between 1 and 2 J (giving rise to photon
densities approximately 2 orders of magnitude greater than in
the present study) results in O(3P) generation. As in the present
study, sequential two-photon absorption was thought to be
responsible.
Taking the data from a typical NO2/Ar/H2O experiment in
which the OH concentration was calibrated, we can derive the
approximate cross section for the absorption of the second
photon by NO2*. With a starting concentration of [NO2] )
2.8 × 1015 molecules cm-3 and a photon fluence of 13.6 mJ
cm-2 at 439.4 nm and by using the known absorption cross
section of NO2 (σ ) 7.08 × 10-19 cm2 molecule-1), we calculate
that 6.0 × 1013 molecules cm-3 NO2* is generated in the pulse.
As the identical slopes were obtained for Nd:YAG or excimer
pumps, quenching of NO2* must be negligible within the
duration of the laser pulse (5 or 20 ns, respectively). In addition,
quenching of O(1D) by Ar is negligible compared to reaction
with H2O, and we assume that the concentration of OH (8.4 ×
1011 molecules cm-3) is the same as the number of photons
absorbed by NO2* per unit volume. (A small correction of
<10% is necessary to take into account reaction of O(1D) with
2
of wavelength close to 220 nm to the 2 B2 state, as has been
2
observed in absorption spectroscopic studies. The 2 B2 state
dissociates to give O(1D) and NO(2Π) with a single-photon
absorption threshold of 244 nm,5 although yields of O(3P)
remain substantial at wavelengths between 214 and 239 nm with
the O(1D) yield close to 40%.24
O(1D) is thus identified as the source of OH via reaction 2.
As a test of this supposition, experiments were carried out in
which H2O was replaced by CH4. The rate constants for
reaction of O(1D) with H2O and CH4 are 2.2 × 10-10 and 1.7
× 10-10 cm3 molecule-1 s-1, respectively, and H2O and CH4
pressures of 0.5 and 0.65 Torr were chosen to ensure that the
relative rates of quenching of O(1D) with Ar (the quenching
rate constant with Ar is 7.0 × 10-13 cm3 molecule-1 s-1) and
reaction with H2O or CH4 were kept constant. The results of

OH Formation in the Photoexcitation of NO2
J. Phys. Chem. A, Vol. 101, No. 23, 1997 4183
NO2 that does not give OH.) This of course also assumes that
all NO2** dissociates to O(1D). If, however, we equate NO2**
with the 2 2B2 state, and assume that the O(1D) yields obtained
by direct excitation24 at 214-239 nm are applicable here, then
the OH generated represents only about 40% of the total NO2*
excited. We thus derive a very approximate cross section of
ca. 5 × 10-19 cm2 molecule-1 for absorption of a photon of
wavelength 439.4 nm by NO2*. Consistent results ((20%) were
obtained when the laser fluence was halved to 6.8 mJ/cm2 with
[NO2] held constant and when the NO2 concentration was
increased to 4.9 × 1015 molecules cm-3 with the laser fluence
(16.6 mJ/cm2) held constant. Large potential errors on this cross
section arise from the indirect estimation of [NO2*] and [OH]
via joule meter measurements of laser fluence and assumptions
concerning the yield of O(1D) from NO2**. The magnitude of
the cross section appears, however, to be similar that for
absorption of a similar frequency photon by ground-state NO2.
Sequential two-photon absorption and O(1D) formation also
most probably occurred in the experiments of Nizkorodov et
al.,16 who used photon fluences at 436.45 nm that were up to a
factor of 5 greater than in the present work. The linear
dependence of NO3 formation on pulse energy observed by these
authors probably reflects the minor role of this process compared
to the NO2* + NO2 reaction to give NO3 in the photolysis of
pure NO2 samples,
fluence at 532 nm, similar errors in the calculation of [OH] in
the N2O photolysis at 193 nm, and any errors associated with
the relative rate constants for quenching of NO2* by H2O and
Ar. Up until now we have used the data of ref 5 for the
quenching rate constants (kq(Ar) ) 1.85 × 10-11 cm3 molecule-1
s-1). We note, however, that a value of 3.2 × 10-11 cm3
molecule-1 s-1 for Ar quenching of NO2* has been measured
for 639 nm excitation.27 Use of this quenching constant would
result in 40% of NO2* that is quenched by H2O instead of 56%.
The upper limit to the relative contribution of OH formation in
the quenching of NO2* by water vapor then becomes ≈7 ×
10-5
.
Discussion
In the laser excitation experiments described above it was
shown that the two-photon absorption of NO2 at wavelengths
between 430 and 450 nm is facile and leads to O(1D) production.
We now examine the possibility that this can occur with
sufficient efficiency in the troposphere to make it a significant
source of O(1D). At a concentration of 1 ppb, the production
rate of NO2* is approximately JEX(NO2)*[NO2] ) 5 × 108
molecules cm-3 s-1, and the loss rate, mainly due to collisional
quenching (see above), is about 7 × 108 s-1. The production
and loss terms combine to result in a steady-state concentration
of about 1.4 molecules cm-3
.
By assuming similar oscillator strengths for the NO2* r NO2-
(2A1) and NO2** r NO2* transitions as discussed above, and
that the 2 2B2 state (NO2**) dissociates to O(1D), we calculate
a production rate of 2.8 × 10-2 molecules cm-3 s-1 for O(1D)
by the process of sequential two-photon absorption by NO2. A
comparison with the O(1D) production rate via O3 photolysis
of approximately 1 × 106 O(1D) cm-3 s-1 renders this process
uninteresting for the atmosphere. Indeed, any reactions of NO2*
with species other than H2O, O2, and N2 are unlikely to be of
importance under atmospheric conditions due to their low
abundance.
By assuming that the reactivity of the excited state of NO2
formed by 532 nm excitation is applicable to the entire
nondissociative part of the NO2 spectrum, the value derived
above of k10/k12 ≈ 7 × 10-5 (or k10 ≈ 1.2 × 10-14 cm3
molecule-1 s-1) can be used to assess the potential significance
of reaction 10 as a source of OH in the troposphere. Using eq
iii and the conditions outlined in the Introduction, we derive an
upper limit to the rate of OH formation of about 3 × 103 OH
cm-3 s-1, less than 2% of the production rate via O3 photolysis.
532 nm Excitation. The interfering feature of the two-photon
process is that it generates OH via O(1D). The thermodynamic
threshold for two-photon generation of O(1D) from NO2
excitation is 488 nm. This is calculated from the known 244
nm threshold for O(1D) production via one photon and assuming
no internal relaxation of NO2* before the second photon is
absorbed. Thus, although excitation at λ > 488 nm may involve
two-photon absorption, this cannot lead to O(1D), but possibly
to O(3P) formation as previously observed.25,26 Experiments
were therefore carried out with excitation at 532 nm, by coupling
the second harmonic of the Nd:YAG laser directly into the
reaction cell. The cross section of NO2 at 532 nm is smaller
than that around 400-450 nm (σ532 ) 1.44 × 10-19 cm2
molecule-1), but two-photon absorption no longer presents a
limitation to the photon fluence, and the low cross section can
be compensated by high pulse energies to ensure sufficient
population of NO2*.
A flowing mixture of [NO2] ) 4.4 × 1015 molecules cm-3
(measured by visible absorption), 2 Torr of H2O, and 16 Torr
of Ar was irradiated with 157 mJ/cm2 per pulse at 532 nm.
Under these conditions, about 2.7 × 1014 molecules cm-3 NO2*
was generated in the pulse. OH was not detected in these
experiments. The results of an OH calibration experiment
carried out immediately afterward enabled the sensitivity to be
established under identical conditions. This was found to be 1
× 10-7 counts/s (cps) per [OH]. An upper limit to OH
formation in the 532 nm excitation was obtained by fitting the
scattered fluorescence signal to an exponential decay with a
forced slope of 5300 s-1, which corresponds to the expected
decay rate of OH in the presence of 4.4 × 1015 cm-3 NO2. The
back-extrapolated signal at t ) 0 was found to be 800 ( 600
cps. This is equivalent to about 8 × 109 OH cm-3. Taking
the relative efficiencies of quenching of NO2 by Ar and H2O
and the relative concentrations of H2O and Ar, it can be
calculated that for this experiment 53% (1.4 × 1014 molecules
cm-3) of the NO2* was quenched by H2O. An upper limit of
about k10/k12 ) 8 × 109/1.4 × 1014 ≈ 6 × 10-5 can be placed
on the relative contribution of OH formation in the quenching
of NO2* by water vapor. The value of 6 × 10-5 contains errors
due to the estimation of NO2* from joule meter reading of laser
Conclusions
The potential role of the reaction between NO2* and H2O as
a source of OH in the troposphere has been assessed by exciting
NO2 at discrete wavelengths between 430 and 450 nm and at
532 nm. The experiments at 430-450 nm revealed a facile
sequential two-photon process that leads, via O(1D), to OH. This
process will be inefficient in the troposphere due to the low
photolysis intensities and rapid collisional quenching of NO2*.
At 532 nm no OH production was observed, enabling an upper
limit of about 7 × 10-5 to be placed on the relative rate constant
for reactive quenching of NO2* to nonreactive quenching
(collisional) by H2O. If the reactivity of NO2* following
excitation at this wavelength is representative of the entire
nondissociative part of the NO2 absorption spectrum, we place
an upper limit of about 2% to formation of OH via reaction 10
in the troposphere when solar zenith angles are high.
Acknowledgment. We thank Paul Crutzen for proposing that
the reaction of NO2* with H2O may be a source of tropospheric
OH, thereby stimulating this research.

4184 J. Phys. Chem. A, Vol. 101, No. 23, 1997
Crowley and Carl
(15) Jones, I. T. N.; Bayes, J. D. J. Chem. Phys. 1973, 59, 4836.
(16) Nizkorodov, S. A.; Makarov, V. I.; Khmelinskii, I. V.; Kotschubei,
S. A.; Amosov, K. A. Chem. Phys. Lett. 1994, 222, 135.
(17) Zellner, R.; Hartmann, D.; Rossner, I. Ber. Bunsen-Ges. Phys.
Chem. 1992, 96, 385.
(18) Pitts, J. N.; Sanhueza, E.; Atkinson, R.; Carter, W. P. L.; Winer,
A. M.; Harris, W. H.; Plun, C. N. Int. J. Chem. Kinet. 1984, 16, 919 and
references therein.
(19) Besemer, A. C.; Nieboer, H. Atmos. EnViron. 1985, 19, 507.
(20) Akimoto, H.; Takagi, H.; Sakamaki, F. Int. J. Chem. Kinet. 1987,
19, 539.
(21) Crowley, J. N.; Campuzano-Jost, P.; Moortgat, G. K. J. Phys. Chem.
1996, 100, 3601.
(22) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling, JPL Publication 94-26, Pasadena, 1994.
(23) Clyne, M. A. A.; Down, S. J. Chem. Soc., Faraday Trans. 2 1974,
70, 253.
(24) Uselman, W. A.; Lee, E. K. C. J. Chem. Phys. 1976, 65, 1948.
(25) Gerstmayr, J. W.; Harteck, P.; Reeves, R. R. J. Phys. Chem. 1972,
76, 474.
References and Notes
(1) Crutzen, P. J. Atmospheric Interactions-Homogeneous Gas Reac-
tions of C, N and S containing compounds (and references therein). In The
Major Biogeochemical Cycles and Their Interactions; Bolin, B., Cook, R.
B., Eds.; 1983; Chapter 3, SCOPE 21.
(2) Crutzen, P. J. Annu. ReV. Earth Planet. Sci. 1979, 7, 443.
(3) Crutzen, P. J.; Gidel, L. T. J. Geophys. Res. 1983, 88, 6641.
(4) Hough, A. J. Geophys. Res. 1991, 96, 7352.
(5) Okabe, H. Photochemistry of Small Molecules; John Wiley and
Sons: New York, 1978.
(6) Roehl, C. M.; Orlando, J. J.; Tyndall, G. S.; Shetter, R. E.; Vazquez,
G. J.; Cantrell, C. A.; Calvert, J. G. J. Phys. Chem. 1994, 98, 7837 and
references therein.
(7) Paech, F.; Schmiedl, R.; Demtro¨der, J. Chem. Phys. 1975, 63, 4369.
(8) Haas, Y.; Houston, P. L.; Clark, J. H.; Moore, C. B.; Rosen, H.;
Robrish, P. J. Chem. Phys, 1975, 63, 4195.
(9) Donnelly, V. M.; Kaufman, F. J. Chem. Phys. 1977, 66, 4100.
(10) Busch, G. E.; Wilson, K. R. J. Chem. Phys. 1972, 56, 3638.
(11) LUTHER program (adapted from: Luther, F. M.; Gelinas, R. J. J.
Geophys. Res. 1976, 81, 1125. Isaksen, I. S. A.; Midtbo, K.; Sunde, J.
Crutzen, P. J. Institut Report Series, No. 20, Institut for Geofysikk,
University of Oslo, 1976).
(26) Hakala, D.; Harteck, P.; Reeves, R. R. J. Phys. Chem. 1974, 78,
1583.
(27) Hakala, D. F.; Reeves, R. R. Chem. Phys. Lett. 1976, 28, 510.
(12) Norrish, R. G. W. J. Chem Soc. 1929, 1158, 1611.
(13) Dickinson, R. G.; Baxter, W. P. J. Am. Chem. Soc. 1928, 50, 774
(14) Creel, C. L.; Ross, J. J. Chem. Phys. 1976, 64, 3560.