A temperature-dependent kinetics study of the important stratospheric reaction O(3P) + NO2 → O2 + NO

Article abstract of DOI:10.1021/jp011940o

The kinetics (k₁(T)) of the stratospheric reaction O(³P) + NO₂ → O₂ + NO at 221-425 K was determined using a laser flash photolysis-resonance fluorescence technique. The observed temperature dependence was closely described by the Arrhenius expression k₁T = (4.21 ± 0.25) x 10^-12 exp{(273 ± 18)/T} cc/mol-sec, where the uncertainties represent precision at the 2σ level. The reported values for k₁(T) had an accuracy of ±6% over the temperature range. The use of these data in models of stratospheric chemistry would generate somewhat lower mid-stratospheric ozone levels than results obtained using data from previous studies.

Full text of DOI:10.1021/jp011940o

J. Phys. Chem. A 2001, 105, 9697-9703
9697
A Temperature-Dependent Kinetics Study of the Important Stratospheric Reaction O(3P) +
NO2 f O2 + NO
E. G. Estupin˜a´n,† J. M. Nicovich,‡ and P. H. Wine*,†,‡
School of Earth and Atmospheric Sciences and School of Chemistry and Biochemistry,
Georgia Institute of Technology, Atlanta, Georgia 30332
ReceiVed: May 21, 2001; In Final Form: August 14, 2001
A laser flash photolysis-resonance fluorescence technique has been employed to investigate the kinetics of
the important stratospheric reaction O(3P) + NO2 f O2 + NO (k1) as a function of temperature (221-425
K) with particular attention paid to obtaining the most accurate values possible for k1(T). The following
Arrhenius expression adequately describes the observed temperature dependence: k1(T) ) (4.21 ( 0.25) ×
10-12 exp{(273 ( 18)/T} cm3 molecule-1 s-1, where the uncertainties represent precision at the 2σ level. The
accuracy of the reported values for k1(T) is estimated to be (6% over the entire temperature range investigated.
At room temperature the rate coefficient measured in this study is in excellent agreement with that reported
in a recent paper by Gierczak et al. [J. Phys. Chem. A 1999, 103, 877] while the activation energy of the rate
coefficient is somewhat more negative than that reported by Gierczak et al. Incorporation of the results of the
present study into models of stratospheric chemistry would lead to somewhat lower mid-stratospheric ozone
levels than would be obtained using results of previous studies.
Introduction
in k1(T) have been identified as a major source of uncertainty
in stratospheric models.8 The kinetics of reaction 1 were recently
reinvestigated by Gierczak et al.9 These investigators obtained
a value for k1(298 K) that is 10% faster than the previously
recommended value, and report that k1 increases with decreasing
temperature more rapidly than suggested by earlier studies. The
findings of Gierczak et al.9 stimulated the present study of the
kinetics of reaction 1 with the goal of measuring this critical
rate coefficient to a high degree of accuracy over the temperature
range 220-425 K.
In this study, kinetic information was obtained by monitoring
the temporal profile of oxygen atoms under pseudo-first-order
conditions with NO2 in large excess. Special emphasis was
placed on accurate determination of the NO2 concentration and
on measurements of k1 at stratospheric temperatures.
Active nitrogen constituents, primarily NO and NO2 (together
known as NOx) play important roles both in the troposphere
and stratosphere. In the troposphere, NOx is necessary for the
photochemical production of O3. In addition, NOx directly and
indirectly affects the concentration of the hydroxyl radical,
which is the primary tropospheric oxidant responsible for
converting many primary emissions to secondary products. In
the stratosphere, NOx-catalyzed destruction cycles in combina-
tion with those involving ClOx, BrOx, and HOx radicals strongly
influence the concentration and vertical profile of O3. The
catalytic ozone destruction cycle
O(3P) + NO2 f O2 + NO
O3 + NO f O2 + NO2
(1)
(2)
Experimental Technique
Net: O(3P) + O3 f 2O2
(3)
The laser flash photolysis (LFP)-resonance fluorescence (RF)
apparatus used in this study was similar to systems previously
employed in this laboratory to study atom-molecule reactions
involving O atoms.10 A schematic diagram of the current version
of the apparatus is published elsewhere.11 Important features
of the apparatus and experimental approach that are specific to
this study are described below.
A Pyrex, jacketed reaction cell with an internal volume of
∼150 cm3 was used in all experiments. The cell was maintained
at a constant temperature by circulating ethylene glycol (T >
298 K) or methanol (T < 298 K) from a thermostated bath
through the outer jacket. A copper-constantan thermocouple
with a stainless steel jacket was periodically injected into the
reaction zone through a vacuum seal to measure the gas
temperature under the precise pressure and flow rate conditions
of the experiment. The thermocouple was retracted during
kinetic experiments and the temperature measurement in the
reaction zone is estimated to be accurate to (1 K.
is the most important NOx cycle in the stratosphere. Reaction 1
is the rate-determining step in the cycle and it accounts for the
majority of odd-oxygen destruction in the 25 to 40 km altitude
regime.1 Hence, it is desirable to know k1(T) with a very high
degree of accuracy, especially at temperatures that are typical
of the mid-stratosphere, i.e., 220-260 K.
While several studies of the kinetics of reaction 1 have been
reported in the literature,2-6 the 1997 NASA panel for chemical
kinetics and photochemistry data evaluation7 suggested that the
uncertainties in k1(T) values remained much higher than de-
sired, particularly at temperatures below 240 K. Uncertainties
* Author to whom correspondence should be addressed at School of
Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, GA
30332-0400. E-mail: pw7@prism.gatech.edu.
† School of Earth and Atmospheric Sciences.
‡ School of Chemistry and Biochemistry.
10.1021/jp011940o CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/27/2001

9698 J. Phys. Chem. A, Vol. 105, No. 42, 2001
Estupin˜a´n et al.
Oxygen atoms were produced by 355 nm laser flash pho-
tolysis of NO2 using third harmonic radiation from a Nd:YAG
laser (Quanta Ray model DCR-2A, pulse width ∼6 ns) as the
photolytic light source:
stream of the reaction cell. Using a partially reflecting optic,
the output beam of the argon ion laser operating at 457.9 nm
was split into two approximately equal components which were
then multipassed using White cell optics12 through two separate
Pyrex absorption cells. The windows on the absorption cells
had antireflection coatings (400-500 nm) and the White cell
mirrors were coated for high reflectivity (425-485 nm). For
these experiments, the upstream cell was 37.3 cm in length with
52 passes of the laser beam and the downstream cell was 35.5
cm in length with 56 passes. The intensity of the laser light
exiting the White cells was continuously monitored by separate
silicon photodiodes. To reference the signals from these two
photodiodes to the initial intensity of the laser, a thin, glass
optic was used to pick off a portion of the laser beam before
the beam splitter and direct it onto a third photodiode. Each
photodiode was equipped with a stack of thin Teflon diffusers
to prevent saturation. The detector signals were electronically
amplified and processed through an analog-to-digital card in a
personal computer. An electromechanical shutter allowed the
laser beam to be periodically blocked in order to measure the
small background for each detector. On a separate gas handling
system, the NO2 absorption cross section at 457.9 nm and room
temperature was measured using the same Ar+ laser, a 37.4 cm
long absorption cell, and manometrically measured samples of
NO2. Again, the light intensity of the laser beam exiting the
absorption cell was referenced to the initial intensity using a
pick-off mirror and a second photodiode. Absorption measure-
ments were made over a wide range of [NO2] (0.11 to 0.87
Torr), path length (1 to 5 passes of the laser beam), and total
pressure (N2 was added for some measurements). The pressure
gauges used for these determinations were calibrated by the
expansion of known masses of various gases into the known
volumes of the gas manifold and absorption cell. Results of
both methods of determining the NO2 concentrations in the
kinetics experiments are discussed below.
NO2 + hV f O(3P) + NO
(4)
At λ ) 355 nm, the yield of O(3P) in reaction 4 is known to be
unity.7 An atomic resonance lamp, situated perpendicular to the
photolysis laser, excited resonance fluorescence in the pho-
tolytically produced atoms. The resonance lamp consisted of
an electrodeless microwave discharge through ∼1 Torr of a
flowing mixture containing a trace of O2 in helium. The flow
of a 0.1% O2 in He mixture and pure He into the lamp were
controlled by separate needle valves, thus allowing the total
pressure and O2 concentration to be adjusted for optimum signal-
to-noise. Radiation was coupled out of the lamp through a MgF2
window and into the reaction cell through a MgF2 lens. Dry N2
was used as a purge gas in the volume between the lamp window
and the reaction cell lens to exclude room air and thus allow
transmission of vacuum-UV radiation.
Fluorescence from excited O(3P) atoms within the reaction
zone was collected by a MgF2 lens on an axis orthogonal to
both the photolysis laser beam and the resonance lamp beam
and imaged onto the photocathode of a solar blind photomul-
tiplier. The region between the reaction cell and the photomul-
tiplier was purged with dry N2 and contained a CaF2 window
to prevent detection of Lyman-R emission from the resonance
lamp. The fluorescence signals were processed using photon-
counting techniques in conjunction with multichannel scaling.
For each O(3P) decay rate measured, 500-5000 temporal
profiles were co-added to obtain a well-defined temporal profile.
The multichannel analyzer sweep was triggered approximately
3.2 ms prior to the photolysis laser in order to determine the
background light level immediately before the laser flash.
To avoid accumulation of photolysis or reaction products,
all experiments were carried out under “slow-flow” conditions.
The N2 used in this study was UHP grade with a minimum
purity of 99.999%; it was used as supplied. NO2 was prepared
by reacting NO with excess O2 and leaving the mixture
overnight. NO2 was then collected in a liquid nitrogen cooled
trap while O2 and other impurities were pumped away. NO and
O2 had minimum stated purities of 99.0% and 99.996%,
respectively.
The linear flow rate through the reactor was typically 3 cm s-1
,
and the laser repetition rate was typically 5 Hz. Since photolysis
occurred on an axis perpendicular to the direction of flow, no
volume element of the reaction mixture was subjected to more
than a few laser shots. The reactant and photolytic precursor,
NO2, was flowed into the reaction cell from a 12-L bulb
containing dilute mixtures in nitrogen buffer gas, while N2 was
flowed directly from a high-pressure tank. The NO2/N2 bulb
was blackened to prevent photolysis by room lights.
Results and Discussion
1. NO2 Concentration Results. The NO2 absorption cross
section at the argon ion laser wavelength (457.9 nm) was
determined thirty-two times in order to maximize the precision
of the measurement. These results are summarized in Figure 1,
where the calculated NO2 cross section is plotted as a function
of absorbance. Two parameters were varied to ensure that the
measured NO2 cross section was independent of initial condi-
tions; these were the initial NO2 pressure and the total laser
absorption path length. No systematic change in cross section
results was observed. The NO2 concentration was corrected for
the formation of N2O4 using the equilibrium constant recom-
mended by the NASA panel.7 The corrections were very small
(<1%). Calibration of the pressure gauge used in the cross
section measurements showed that the pressure reading was 1.9
( 2.5% higher than expected from masses of the expanded ideal
gases (uncertainty is 2σ). Therefore, the NO2 cross section was
corrected upward by 1.9%. Two experiments were performed
at a total pressure of 100 Torr using N2 as the buffer gas (open
symbols in Figure 1). From our experiments, the NO2 absorption
cross section at 457.9 nm is measured to be (5.06 ( 0.08) ×
To minimize systematic error in the NO2 concentration
measurement, two independent methods were employed. In one
method, the flow of the NO2/N2 mixture and the flow of pure
N2 were both measured with calibrated mass flow meters. The
concentration of NO2 was calculated using measurements of
the mass flow rates of the two components of the flow, the
total pressure in the reaction cell (measured with a 1000 Torr
full scale capacitance manometer), and the temperature in the
reaction cell. The mole fraction of NO2 in the NO2/N2 mixture
was checked frequently by UV photometry at 366 nm. The
photometric measurements were carried out on a separate high
vacuum gas handling system employing a mercury penray lamp
and a photomultiplier tube equipped with an interference filter
to isolate the three Hg transitions at λ ∼ 366 nm from other
lamp emissions. The effective NO2 absorption cross section was
measured using the same lamp and filter to be (6.05 ( 0.14) ×
10-19 cm2 molecule-1
.
The concentration of NO2 was also determined by in situ
long-path absorption measurements both upstream and down-

Stratospheric Reaction O(3P) + NO2 f O2 + NO
J. Phys. Chem. A, Vol. 105, No. 42, 2001 9699
Figure 2. Typical resonance fluorescence temporal profiles observed
in the study of reaction 1. Experimental conditions: T ) 244 K; P )
15 Torr; [NO2] in units of 1013 molecules cm-3 ) (A) 3.09, (B) )
6.58, (C) ) 12.20; number of laser shots averaged ) (A) 800, (B) )
1000, (C) ) 2000. Solid lines are obtained from least squares analyses
and give the following first-order decay rates in units of s-1: (A) 490,
(B) 940, (C) 1620.
Figure 1. Plot of the calculated NO2 absorption cross section at the
argon ion laser wavelength (457.9 nm) versus absorbance. Solid circles
indicate absorption cross-section measurements made with only NO2,
while open circles indicate absorption cross-section measurements made
with NO2/N2 mixtures at a total pressure of 100 Torr.
TABLE 1: Quantitative Comparison of k1 Obtained from
This Study Using the NO2 Concentration Determined by
Absorption with the NO2 Concentration Determined by Flow
reaction cell. These pressure drop corrections were small (<2%)
and could be made very accurately.
Ta
kabsorptionb/kflow
Ta
kabsorptionb/kflow
c
c
2. Kinetic Results. All the experiments were carried out under
pseudo-first-order conditions with NO2 in large excess over
O(3P). Thus, in the absence of secondary reactions that enhance
or deplete the O(3P) atom concentration, the O(3P) atom
temporal profile is dominated by the reactions
425
406
374
346
324
294
295
295
1.01
1.00
1.01
0.99
1.02
0.99
1.00
1.01
295
295
295
295
271
244
221
0.98
1.06
1.01
1.05
1.01
1.01
1.00
O(3P) + NO2 f O2 + NO
(1)
a Units are T (K). b Rate coefficient for reaction 1 determined using
an average between the upstream and downstream NO2 concentrations
measured by absorption. c Rate coefficient for reaction 1 determined
using the NO2 concentration measured by flow.
O(3P) f loss by diffusion from the detector field of
view and/or reaction with background impurities (5)
Integration of the rate equations for the above scheme yields
the following simple relationship:
10-19 cm2 molecule-1, where the stated uncertainty represents
precision at the 95% confidence level. It is difficult to make a
direct comparison between the NO2 cross section obtained at
457.9 nm in this study and NO2 cross sections obtained in
previous studies because of the higher resolution of our
measurement and the structured nature of the NO2 absorption
spectrum. Nonetheless, there is reasonably good agreement
between our measured NO2 cross section and the 0.5 nm
averages reported by Harder et al.,13 ∼5.1 × 10-19 cm2
molecule-1, and Schneider et al.,14 ∼4.5 × 10-19 cm2 molecule-1.
In the kinetics experiments, excellent agreement was found
between the NO2 concentration measurements made by in situ
absorption and the corresponding concentrations calculated from
flow and pressure measurements. Furthermore, excellent agree-
ment was found between the upstream and downstream NO2
concentrations determined by absorption. Table 1 shows the ratio
of the rate coefficient for reaction 1 determined using an average
between the upstream and downstream NO2 concentrations
measured by absorption to the same rate coefficient determined
using the NO2 concentration measured by flow. The difference
between the measurements was typically less than 2% and was
never greater than 6%. Hence, the three NO2 measurements were
averaged, ([NO2]upstream + [NO2]downstream + [NO2]flow)/3, to yield
a single NO2 concentration determination for each pseudo-first-
order decay rate. Since the experiments were performed under
flow conditions, the NO2 concentration measured in each of
the two absorption cells was corrected to the pressure in the
ln{S0/St} ) (k1[NO2] + k5)t ) k’t
(6)
In eq 6, S0 is the O(3P) fluorescence signal at a time shortly
after the laser fires and St is the O(3P) fluorescence signal at
time t. The bimolecular rate coefficient, k1, is determined from
the slope of a k′ vs [NO2] plot.
O(3P) atom decays were found to be exponential and the
pseudo-first-order O(3P) decay rates were found to increase
linearly with increasing NO2 concentration except at T < 240
K (see discussion below). These kinetic observations are
consistent with eq 6. Furthermore, observed decay rates were
independent of laser photon fluence. This set of observations
strongly supports the contention that reactions 1 and 5 are the
only processes which affected the post-laser-flash O(3P) time
history in the experiments performed at T > 240 K and the
dominant processes at T < 240 K.
Typical observed oxygen atom temporal profiles are shown
in Figure 2 and typical plots of k′ versus [NO2] are shown in
Figure 3. The kinetic data for reaction 1 are summarized in Table
2. k1 was measured at 10 different temperatures, ranging from
221 to 425 K. Some experiments were performed at a very low
NO2 flow through the system (i.e., not measurable by either
absorption or flow methods) in order to measure the O(3P) atom
background decay rate in essentially the absence of NO2. In
these experiments a few percent of NO2 was lost due to

9700 J. Phys. Chem. A, Vol. 105, No. 42, 2001
Estupin˜a´n et al.
Figure 4. Arrhenius plot for reaction 1 and comparison with the results
of Gierczak et al.9 and the 1997 NASA panel7 recommendation. Solid
data points are from this study and error bars represent our best esti-
mate of accuracy. The solid line is obtained from a weighted least-
squares analysis (of data from this study only) and represents the
Arrhenius expression given in the text. The dashed line represents the
Arrhenius expression reported by Gierczak et al.9 and the dotted line
represents the Arrhenius expression recommended by the NASA panel7
in 1997.
Figure 3. Typical plots of k′ versus [NO2] for data obtained in the
study of reaction 1. Solid lines are obtained from least squares analyses;
their slopes give the bimolecular rate coefficients in units of 10-12 cm3
molecule-1 s-1: 12.7 at 244 K, 10.5 at 295 K, and 7.89 at 425 K.
TABLE 2: Summary of Kinetic Data for the Reaction O(3P)
+ NO2 f O2 + NO
no. of
a,c
a,e
Ta Pa
[NO2]a,b
[O]0
expts.d range of k′a
k1
425 15
406 15
374 15
346 15
324 15
5-1750
44-3010
6-1780
5-3500
0-1780
3.0
3.6
3.0
4.0
3.0
3.3
3.2
1.7
8.9
3.1
3.1
3.1
3.2
2.9
3.1
6
7
6
8
6
5
6
6
7
6
6
6
7
7
6
180-1550
190-2600
140-1700
170-3450
120-1850
530-3980 10.4 ( 0.2
120-2230 10.5 ( 0.5
94-1970 10.6 ( 0.2
94-2970 10.6 ( 0.2
95-1890 10.9 ( 0.3
130-2000 10.9 ( 0.3
120-2050 10.9 ( 0.4
95-3370 11.6 ( 0.3
90-1620 12.8 ( 0.3
90-1110 14.3 ( 0.4
7.89 ( 0.29
8.17 ( 0.40
8.82 ( 0.14
9.50 ( 0.32
9.68 ( 0.21
the rate coefficient for reaction 1 is characterized by a small
negative activation energy. A linear least-squares analysis of
the ln k1 vs 1/T data gives the following Arrhenius expression:
k1(T) ) (4.21 ( 0.25) ×
294 15 380-3700
295 15
295 15
295 15
295 15
295 15
295 15
271 16
244 15
221 15
0-2010
0-1760
0-2700
0-1620
0-1700
0-1700
0-2860
0-1200
0-700
10-12 exp{(273 ( 18)/T} cm3 molecule-1 s-1 (8)
Uncertainties in the above expression are 2σ and represent
precision only. These uncertainties refer to the Arrhenius
parameters only. Error estimates for individual rate coefficients
are derived below.
The temperature dependence of k1 can also be parametrized
as k1(T) ) A(T/300)-n where T is in units of degrees Kelvin:
a Units are T (K), P (Torr); concentrations (1011 per cm3), k′ (s-1),
k1 (10-12 cm3 molecule-1 s-1). b Average concentration based on flow
and absorption measurements. A value of zero indicates that the NO2
concentration was below our detection limit. c Average concentration
for each reported value of k calculated based on the measured laser
fluence and NO2 concentration. d expt t determination of a single
pseudo-first-order decay rate. e Uncertainties are 2σ and represent
precision only.
k1(T) ) (10.57 ( 0.12) ×
10-12 (T/300)-(0.85 ( 0.06) cm3 molecule-1 s-1 (9)
Again, uncertainties in the above expression are 2σ and represent
precision only.
At temperatures below 240 K, it becomes necessary to
consider the effect of NO2 dimerization on observed kinetics.
photolysis because very high laser powers were needed to obtain
a measurable fluorescence signal. In all other experiments the
NO2 concentration was always in large excess over the O(3P)
atom concentration (pseudo-first-order conditions). In the room-
temperature experiments (294-295 K), the initial O(3P) con-
centration was varied by about a factor of 5 [(1.7-8.9) × 1011
atoms cm-3] resulting in no observed systematic change in the
rate coefficients. All the experiments were performed at a
relatively low pressure of 15 Torr in order to avoid significant
contribution to observed kinetics from the addition reaction
NO2 + NO2 T N2O4
(10)
Unfortunately, the uncertainty in the equilibrium constant, K10,
is very large at low temperatures,7 thus making quantitative
evaluation of the concentration of NO2 and N2O4 in the reaction
cell difficult. To shed some light on this problem, we carried
out a set of experiments at T ) 210 K where O(3P) kinetics
were studied over a wide range of concentrations of NO2
monomer units. As expected, and as shown by the solid symbols
in Figure 5, the slope of a plot of k′ vs [NO2 monomer units]
decreased with increasing [NO2 monomer units]. The rate
coefficient for the reaction
O(3P) + NO2 + M f NO3 + M
(7)
The maximum contribution of reaction 7 (at the lowest
temperature investigated) is less than 1% based on the 1997
NASA panel7 recommendation and less than 2% using the value
derived from the recent study by Burkholder and Ravishankara.15
An Arrhenius plot for reaction 1 is shown in Figure 4. Like
many radical-radical reactions, the temperature dependence of
O(3P) + N2O4 f products
(11)
is thought to be much smaller than k1, i.e., k11 < 2 × 10-12
cm3 molecule-1 s-1 at 199 K.6 Assuming this to be the case,

Stratospheric Reaction O(3P) + NO2 f O2 + NO
J. Phys. Chem. A, Vol. 105, No. 42, 2001 9701
K10(T) ) 6.62 × 10-30 exp(7258/T) cm3 molecule-1 (14)
At T ) 220 K, the above expression gives an equilibrium
constant that is about a factor of 2 larger than the current NASA
panel7 evaluation (expression 13), although well within the
evaluated uncertainty of almost a factor of 3. Interestingly,
Wollenhaupt and Crowley,16 in a recent study of the CH3O +
NO2 reaction, arrived at the same conclusion, i.e., the low-
temperature value for K10 is somewhat larger than currently
recommended. In fact, their derived value for K10 at 233 K is
within 2% of the 233 K value obtained from expression 14.
Correcting our NO2 concentration data for the formation of N2O4
at 221 K using expression 14 to obtain K10 gives a value for k1
of 1.47 × 10-11 cm3 molecule-1 s-1. Negligible corrections are
needed for the NO2 concentration at the higher temperatures
investigated (i.e., 244 K and above).
Figure 5. Plot of k′ versus [NO2 monomer units] for data obtained in
the study of the NO2 dimerization equilibrium constant, K10, at T )
210 K. The solid curve is obtained from a nonlinear fit analysis. k11
was fixed to 2.0 × 10-12 cm3 molecule-1 s-1. Results of the fit are the
following: k1 ) 1.57 × 10-11 cm3 molecule-1 s-1, and K10 ) 6.77 ×
10-15 cm3 molecule-1. The dotted curve is obtained from a nonlinear
fit analysis by fixing k11 to zero. Results of this fit are the following:
k1 ) 1.40 × 10-11 cm3 molecule-1 s-1, and K10 ) 3.23 × 10-15 cm3
Once again, assuming that the 298 K equilibrium constant
recommended by the NASA panel7 is correct and that the 210
K equilibrium constant derived from fitting our kinetic data,
assuming k11 to be zero, is also correct, we obtain the following
expression for the temperature dependence of the equilibrium
constant:
molecule-1
.
K10(T) ) 3.82 × 10-29 exp(6735/T) cm3 molecule-1 (15)
our data were fit to an equation of the form
At T ) 220 K, expression 15 gives an equilibrium constant
that is only about 12% faster than the current NASA panel7
evaluation (expression 13). Using expression 15 to correct our
NO2 concentration at 221 K gives a lower limit value for k1 of
k′ ) k1[NO2] + k11[N2O4] + k′0
(12)
where k′0 was measured to be 94 s-1 using a very small
concentration of NO2 (<1 × 1012 molecules cm-3) and [NO2]
is expressed in terms of the dimerization equilibrium constant
for reaction 10 and the concentration of NO2 monomer units.
k1 and K10 were allowed to vary and k11 was set to 2.0 × 10-12
cm3 molecule-1 s-1. This nonlinear fit to the data is shown as
the solid curve in Figure 5. The value for the dimerization
equilibrium constant obtained from this fit was 6.77 × 10-15
cm3 molecule-1 and the k1 value was 1.57 × 10-11 cm3
molecule-1 s-1, which is in excellent agreement with the value
for k1 obtained from extrapolating expression 8 to a temperature
of 210 K. The k′ vs [NO2 monomer units] data were also fit
assuming k11 to be zero (dotted curve in Figure 5). The results
of this fit were 3.23 × 10-15 cm3 molecule-1 for K10 and 1.40
× 10-11 cm3 molecule-1 s-1 for k1(210 K). This value of K10
(i.e., assuming k11 to be zero) is in better agreement (only 13%
larger) with the currently recommended value for K10(210 K)7
(see below). In addition, the k′ vs [NO2 monomer units] data at
210 K were also fit allowing all three parameters, k1, K10, and
k11, to vary and we obtained 1.54 × 10-11 cm3 molecule-1 s-1
for k1, 6.04 × 10-15 cm3 molecule-1 for K10, and 1.8 × 10-12
cm3 molecule-1 s-1 for k11.
1.37 × 10-11 cm3 molecule-1 s-1
.
On the basis of the analysis described above, we report our
221 K rate coefficient as (1.43 ( 0.07) × 10-11 cm3 mole-
cule-1 s-1, where the stated error includes precision and the
uncertainty in the correction for NO2 dimerization. This value
weights the rate coefficient obtained using K10 from expression
14 by 60% and the rate coefficient obtained using expression
K10 from expression 15 by 40%. The higher weight for
expression 14 comes from the slightly better fit of k′ vs [NO2
monomer units] at 210 K using a value for k11 of 2.0 × 10-12
cm3 molecule-1 s-1
.
3. Estimated Accuracy of Reported Rate Coefficients. The
two major contributors to the overall accuracy of the rate
coefficients in this temperature-dependent study of the O + NO2
reaction are the precision of the rate coefficients and the
accuracy of the NO2 concentration determinations. Examination
of Table 2 shows that the precision of the rate coefficients is
quite good (typically (3%).
As discussed above, to determine the NO2 concentration by
either “absorption” or “flow,” the NO2 absorption cross section
at the chosen wavelength needs to be determined; this is a
potential source of error. In the case of the NO2 absorption
measurements, the NO2 cross section at the argon ion laser
wavelength (457.9 nm) was determined with a precision of (2%
at the 95% confidence level. We conservatively estimate that
the absolute calibration of the pressure gauge is within (2.5%.
There is also some error associated with correcting the NO2
concentration for the formation of N2O4. However, the uncer-
tainty in the equilibrium constant is a negligible source of error
in the NO2 absorption cross section determinations since the
contribution of N2O4 to the total pressure was less than 1%.
The precision of the determinations of both I and I0 in both the
NO2 cross section measurements at 457.9 and in the determi-
nation of the NO2 concentration during the kinetic experiments
The NASA panel for chemical kinetics and photochemical
data evaluation7 currently recommends the following expression
for the temperature dependence of K10:
K10(T) ) 2.5 × 10-19 exp(6643/T) cm3 molecule-1 (13)
Assuming that the 298 K equilibrium constant recommended
by the NASA panel,7 is correct and that the 210 K equilibrium
constant derived from fitting our kinetic data assuming k11 to
be 2.0 × 10-12 cm3 molecule-1 s-1 is also correct, we obtain
the following expression for the temperature dependence of the
equilibrium constant:

9702 J. Phys. Chem. A, Vol. 105, No. 42, 2001
Estupin˜a´n et al.
was also quite good, i.e., each one about (0.5%. As a result,
the estimated accuracy of the NO2 concentration determined
by absorption is about (3.2%. This assessment combines the
following factors: an error in the 457.9 nm NO2 cross section
(about (3.1%), a possible small error in the measurement of
the total path length of the argon ion laser beam through the
absorption cells of about (0.4%, and a precision error in the
determinations of both I and I0 during the kinetic experiments
((0.5%).
The error in the NO2 concentration determined by “flow” is
estimated to be about (4%. This assessment includes a
(3.5% error in the calculation of the NO2 bulb fraction (which
itself includes a (2% error in the precision of the effective
NO2 cross section for the three Hg atomic lines at λ ∼ 366
nm and a (2.5% error in the pressure gauge used in the cross
section measurements), a (0.5% error in each one of the two
flowmeters, and a (2% error in the pressure gauge used in
the kinetic experiments. Despite our best efforts in deriving a
more accurate equilibrium constant for reaction 1 at low
temperatures, there is still considerable uncertainty in its
determination. Hence, at 221 K, there is an additional (4%
error associated with the choice of K10 used to correct the
NO2 concentration for the presence of N2O4 in the reaction cell.
When all the above errors are propagated, the overall accuracy
of each individual rate coefficient is approximately (6%, and
changes very little over the temperature range of our study
(221-425 K).
4. Comparison of Reported Rate Coefficients with Lit-
erature Values. Figure 4 compares the results of our study of
reaction 1 with those of Gierczak et al.9 and with the 1997
NASA panel7 recommendation. The k1(298 K) value of this
study agrees within 1% with the result of Gierczak et al.,9 while
the low-temperature rate coefficients are somewhat faster than
those reported by Gierczak et al.9 For example, at T ) 220 K
the rate coefficient obtained from our Arrhenius expression is
7.4% faster than the one reported by Gierczak et al.9 Nonethe-
less, the results of this study and of Gierczak et al.9 are in
agreement that the rate coefficient for the O + NO2 reaction is
faster at stratospheric temperatures than previously thought.7
For a critical evaluation of previous O + NO2 rate coefficient
measurements, readers are referred to Gierczak et al.9 and Sander
et al.17
Inspection of Figure 4 shows that the temperature dependence
of k1 reported in this study is somewhat more pronounced than
the temperature dependence of k1 reported by Gierczak et al.9
One possible explanation for this difference concerns the fact
that Gierczak et al.9 measured the NO2 concentration at the
temperature of the kinetic experiment and used measured
temperature-dependent absorption cross sections (measured
during the course of their study) to convert measured absor-
bances to NO2 concentrations. On the other hand, all absorption
measurements in this study were at room temperature, and the
ideal gas equation of state was employed to convert measured
NO2 concentrations to the reaction cell temperature. Also, the
only difference between our value for k1(221 K) and the one
reported by Gierczak et al.9 appears to be the approach employed
to correct for NO2 dimerization, i.e., they used the recom-
mended7 low-temperature equilibrium constant whereas we used
the approach described above.
5. Implications for Atmospheric Chemistry. Incorporation
of the results of this study in models of stratospheric chemistry
would lead to somewhat lower ozone levels than would be
obtained using the expression for k1(T) currently recommended
by the NASA panel.17 This is true even though the current
recommendation heavily weights the results of Gierczak et al.9
The biggest impact of our reported rate coefficients is expected
in the 23-40 km altitude regime, where temperatures are
relatively low and the O + NO2 catalytic cycle dominates odd-
oxygen destruction1. Some of the consequences of increasing
the value of k1 at stratospheric temperatures, together with
updated rate coefficients for the reactions OH + HNO3 (k16)
and OH + NO2 + M (k17), are discussed in a recent modeling
study by Portmann et al.18 These investigators have found that
an increase in k1, combined with the updates in k16 and k17,
leads to lowering, by about one kilometer, the altitude at which
NOx-catalyzed O3 removal dominates. In addition, a faster value
for k1 at stratospheric temperatures leads to a decrease of a few
percent in total column O3. The biggest impact is at higher
latitudes during the summer months, while no change in the O3
column is found in the tropics.18
Acknowledgment. This research was supported by the
National Aeronautics and Space Administration-Upper Atmo-
sphere Research Program through Grant NAG5-8931.
The study by Gierczak et al.9 also used the technique of LFP-
RF to study the kinetics of reaction 1. In their study, oxygen
atoms were produced by 308 nm laser flash photolysis of NO2
using a XeCl excimer laser. As in this study, kinetic information
was obtained by monitoring the temporal profile of oxygen
atoms under pseudo-first-order conditions with NO2 in large
excess. To minimize systematic error in the NO2 concentration
measurement, three independent methods were employed: flow
rate measurements, absorption, and chemical titration (NO +
O3 f NO2 + O2). In the absorption method, the NO2 con-
centration was measured directly in the reactor using UV-
visible photometry at 413.4 nm (a D2 lamp was the light source).
This can be compared with our approach where the NO2
concentration was measured upstream and downstream of the
reaction cell using the output of an argon ion laser at 457.9
nm. As discussed above, excellent agreement was found between
our upstream and the downstream NO2 concentration measured
by absorption. Our approach to obtaining NO2 concentrations
seems to result in better precision than was obtained by Gierczak
et al.9 although, of course, their measurement of NO2 in situ in
the reaction cell is commendable as an approach for minimizing
systematic errors.
References and Notes
(1) McElroy, M. B.; Salawitch, R. J.; Minschwaner, K. Planet. Space
Sci. 1992, 40, 373.
(2) Davis, D. D.; Herron, J. T.; Huie, R. E. J. Phys. Chem. 1973, 58,
530.
(3) Slanger, T. G.; Wood, B. J.; Black, G. Int. J. Chem. Kinet. 1973,
5, 615.
(4) Bemand, P. P.; Clyne, M. A. A.; Watson, R. T. J. Chem. Soc.,
Faraday Trans. 2 1974, 70, 564.
(5) Ongstad, A. P.; Birks, J. W. J. Chem. Phys. 1984, 81, 3922.
(6) Geers-Muller, R.; Stuhl, F. Chem. Phys. Lett. 1987, 135, 263.
(7) 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. Evaluation No. 12; Jet Propulsion Laboratory: Pasadena, CA,
1997.
(8) Dubey, M. K.; Smith, G. P.; Hartley, W. S.; Kinnison, D. E.;
Connell, P. S. Geophys. Res. Lett. 1997, 24, 2737.
(9) Gierczak, T.; Burkholder, J. B.; Ravishankara, A. R. J. Phys. Chem.
A 1999, 103, 877.
(10) Thorn, R. P.; Nicovich, J. M.; Cronkhite, J. M.; Wang, S.; Wine,
P. H. Int. J. Chem. Kinet. 1995, 27, 369, and references therein.
(11) Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Chemistry of the Upper and
Lower Atmosphere; Academic Press: New York, 2000; p 146.
(12) White, J. U. J. Opt. Soc. Amer. 1942, 32, 285.

Stratospheric Reaction O(3P) + NO2 f O2 + NO
J. Phys. Chem. A, Vol. 105, No. 42, 2001 9703
(13) Harder, J. W.; Brault, J. W.; Johnston, P. V.; Mount, G. H. J.
Geophys. Res. 1997, 102, 3861.
(14) Schneider, W.; Moortgat, G. K.; Tyndall, G. S.; Burrows, J. P. J.
Photochem. Photobiol. A 1987, 40, 195.
(15) Burkholder, J. B.; Ravishankara, A. R. J. Phys. Chem. A 2000,
104, 6752.
(16) Wollenhaupt, M.; Crowley, J. N. J. Phys. Chem. A 2000, 104, 6429.
(17) Sander, S. P.; Friedl, R. R.; DeMore, W. B.; Golden, D. M.; Kurylo,
M. J.; Hampson, R. F.; Huie, R. E.; Moortgat, G. K.; Ravishankara, A. R.;
Kolb, C. E.; Molina, M. J. Chemical Kinetics and Photochemical Data for
Use in Stratospheric Modeling. Evaluation No. 13; Jet Propulsion Labora-
tory: Pasadena, CA, 2000.
(18) Portmann, R. W.; Brown, S. S.; Gierczak, T.; Talukdar, R. K.;
Burkholder, J. B.; Ravishankara, A. R. Geophys. Res. Lett. 1999, 26, 2387.