Temperature Dependent Rate Coefficient for the Reaction O(3P) + NO2 → NO + O2

Article abstract of DOI:10.1021/jp983962p

The rate coefficient (k₁) for the reaction O(³P) + NO₂ → O₂ + NO was measured under pseudo-first-order conditions in O(³P) atom concentration over the temperature range 220-412 K. Measurements were made using pulsed laser photolysis of NO₂ to produce oxygen atoms and time-resolved vacuum UV resonance fluorescence detection of O atoms. The NO₂ concentration was measured using three techniques: flow rate, UV absorption, and chemical titration (NO + O₃ → NO₂ + O₂). The NO₂ UV absorption cross section at 413.4 nm was determined as a function of temperature using the chemical titration and flow methods. Including the low-temperature data of Harder et al.¹, the temperature-dependent NO₂ cross section is given by σ_413.4(T) = (9.49 - 0.00549 T) × 10^-19 cm² molecule-1. The measured rate coefficients for reaction 1 can be expressed as k₁(T) - (5.26 ± 0.60) × 10^-12 exp[(209 ± 35)/T] cm^-3 molecule^-1 s^-1, where the quoted uncertainties are 2σ and include estimated systematic errors. This result is compared with previously reported measurements of k₁.

Full text of DOI:10.1021/jp983962p

J. Phys. Chem. A 1999, 103, 877-883
877
3
Temperature Dependent Rate Coefficient for the Reaction O( P) + NO2 f NO + O2
†
,‡
Tomasz Gierczak, James B. Burkholder, and A. R. Ravishankara*
National Oceanic and Atmospheric Administration, Aeronomy Laboratory, 325 Broadway,
Boulder, Colorado 80303, and CooperatiVe Institute for Research in EnVironmental Sciences,
UniVersity of Colorado, Boulder, Colorado 80309
ReceiVed: October 5, 1998; In Final Form: December 8, 1998
3
The rate coefficient (k
1
) for the reaction O( P) + NO
2
f O
2
+ NO was measured under pseudo-first-order
3
conditions in O( P) atom concentration over the temperature range 220-412 K. Measurements were made
using pulsed laser photolysis of NO to produce oxygen atoms and time-resolved vacuum UV resonance
fluorescence detection of O atoms. The NO concentration was measured using three techniques: flow rate,
UV absorption, and chemical titration (NO + O f NO + O ). The NO UV absorption cross section at
13.4 nm was determined as a function of temperature using the chemical titration and flow methods. Including
2
2
3
2
2
2
4
1
the low-temperature data of Harder et al. , the temperature-dependent NO cross section is given by σ413.4(T)
2
-19
2
-1
)
as k
2
(9.49 - 0.00549 T) × 10 cm molecule . The measured rate coefficients for reaction 1 can be expressed
-12 3 -1 -1
1
(T) ) (5.26 ( 0.60) × 10 exp[(209 ( 35)/T] cm molecule s , where the quoted uncertainties are
σ and include estimated systematic errors. This result is compared with previously reported measurements
of k
1
.
Introduction
tions. The values of k1(298 K) reported in the above five studies
agree within 10%. However, the temperature dependence of k1
from these studies disagree significantly; the activation energies
reported from various groups fall in the range 0 to ∼-400 cal
Nitrogen oxides, NO and NO2 (collectively called NOx), play
a crucial role in atmospheric ozone chemistry: they lead to
photochemical ozone production in the troposphere and catalytic
ozone destruction in the stratosphere. In the stratosphere, NOx
chemistry affects both the ozone abundance and its vertical
profile. Of the many possible catalytic ozone destruction cycles
involving NOx, the following is the most important:
-
1
mol . The recommended value of the activation energy, 240
-
1
(
240 cal mol , has been derived from the studies of Davis et
al., Ongstad and Birks, and Geers-Muller and Stuhl. The current
recommendations suggest an uncertainty of ∼60% in the value
of k1(200 K); this large range for k1(200 K) is mostly due to
the uncertainty in the activation energy. This level of uncertainty
has significant implications in the interpretation of atmospheric
measurements of trace species and model calculated abundances
and trends of ozone. The rate coefficient has been identified as
a major source of uncertainty in stratospheric models (see, for
example, ref 9)
3
O( P) + NO f O + NO
(1)
(2)
2
2
O + NO f O + NO
2
3
2
net:
3
Here, we report the temperature dependence of k1 measured
O( P) + O f 2O
(3)
3
2
using the technique of pulsed laser photolysis with resonance
3
fluorescence detection of O( P) atoms (PP-RF). During these
Atmospheric model calculations of ozone abundances and
vertical profiles rely on the temperature-dependent rate coef-
ficients for reactions 1 and 2. Reaction 1 is the rate-limiting
step in this catalytic cycle and has been studied many times
over the past few decades. However, a careful examination of
the available data shows that there are significant discrepancies
and that a more accurate rate coefficient would be beneficial.
experiments, special emphasis was placed on the determination
of NO2 concentration and measurements of k1 at stratospheric
temperatures. Our results are compared with previous measure-
ments and a new value for stratospheric modeling is suggested.
Experimental Section
2,3
Current recommendations for reaction 1 give k1(T) ) 6.5
The accuracy of the value of k1, determined in a system where
-
12
3
-1 -1
×
10
exp(120/T) cm molecule
s
and are based on the
3
the temporal profile of O( P) atoms are measured under pseudo-
4
5
6
studies of Davis et al., Slanger et al., Bemand et al., Ongstad
first-order conditions, depends on how well the concentration
of NO2 is known. Even though NO2 is a stable gas, there are a
few difficulties associated with its handling and knowing its
concentration accurately. First, NO2 can react on the walls of
the reactor, thermally decompose (e.g., in electronic flow
meters), and be photolyzed by room light. Second, it can
undergo self-association
7
8
and Birks, and Geers-Muller and Stuhl. Other earlier studies,
which yielded lower values of k1(298 K) and positive activation
energies, have not been included in deriving the recommenda-
†
Permanent address: Department of Chemistry, Warsaw University, ul.
Zwirki i Wigury 101, 02-089 Warszawa, Poland.
*
Address correspondence to this author at NOAA/ERL, R/E/AL2, 325
Broadway, Boulder, CO 80303. E-mail: ravi@al.noaa.gov.
‡
Also affiliated with the Department of Chemistry and Biochemistry,
2
NO a N O
4
(4)
University of Colorado, Boulder, CO 80309.
2
2
1
0.1021/jp983962p CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/03/1999

878 J. Phys. Chem. A, Vol. 103, No. 7, 1999
Gierczak et al.
hν
3
NO2
9
8 O( P) + NO
(5)
3
3
3
The yield of O( P) in reaction 5 is unity. The O( P) atom
temporal profile (loss) in the presence of NO2 was governed
by the following processes:
3
O( P) + NO f O + NO
(1)
(6)
2
2
M
3
O( P) + NO
9
8 NO
2
3
O( P) + impurities f products
(7)
(8)
3
O( P) f loss
Figure 1. Experimental apparatus used to measure the rate coefficient
3
for reaction 1, O( P) + NO
fluorescence detection of O atoms. O atoms were produced by
photolysis of NO via an excimer laser. The NO concentration in the
reaction cell was determined by UV absorption, flow rate measurements,
or chemical titration (NO + O f NO + O ) as discussed in the text.
2
f NO + O
2
, utilizing atomic resonance
Reaction 6 needs only to be considered in experiments in which
NO was added to the reaction mixture, as described below.
Process 7 represents possible loss of O( P) via reactions with
2
2
3
any impurities in the carrier gas, while reaction 8 represents a
3
2
3
pseudo-first-order loss of O( P) due to diffusion and flow out
of the reaction volume. In our experiments, reactions 7 and 8
were indistinguishable and together accounted for a first-order
to make significant amounts of N2O4 if the NO2 concentrations
-29
are large and/or the temperatures are low (Keq(T) ) 5.2 × 10
-
1
3
-1
loss rate constant of ∼20 s . Note that the first-order rate
exp(6643/T) cm molecule from ref 3). The large temperature
dependence of the equilibrium constant can make even small
changes in temperature result in significant changes in the NO2
concentration at low temperatures. This factor alone has limited
previous kinetic studies to measurements above 230 K. Third,
because k1 is large, only small concentrations of NO2 which
are difficult to accurately determine, have to be used; otherwise,
3
constant for loss of O( P) due to reaction 1 was at least five
times, and on the average a few hundred times, larger than that
due to processes 7 and 8. Oxygen atom temporal profiles were
measured under pseudo-first-order conditions, i.e., [NO2] .
3
3
3
[
O( P)]o, with a ratio [NO2]/ [O( P)]o typically >2.5 × 10 (for
-2
photolyzing with a 308 nm pulse of ∼1 mJ cm ). The exact
3
ratio of [NO2]/[O( P)]o was controlled by the photolysis laser
3
the loss rate coefficients for O( P) atoms exceeds the measurable
-2
fluence and was varied over a wide range (∼0.5 to 7.5 mJ cm ).
4
-1
range (<10 s ). In this work, several methods were employed
to determine the concentration of NO2 in the reactor. Use of a
combination of these methods reduced the possible systematic
errors in NO2 concentration and, therefore, yielded a more
accurate value of k1. The details of these methods are outlined
in a separate section below.
3
The O( P) decays were represented by a simple exponential:
ln(S ) ) ln(S ) -k′t
(9)
t
o
3
where St and So are the O( P) resonance fluorescence signals at
time t and time zero (i.e., right after the photolysis) and
The pulsed photolysis resonance fluorescence (PP-RF)
apparatus and the methodology used to measure rate coefficients
k′ ) k [NO ] + k [NO] + k [impurity] + k
8
(10)
3
1
2
6
7
for O( P) atom reactions have been described in detail previ-
1
0,11
ously.
The apparatus was slightly modified for the present
Approximately 1000-5000 temporal profiles were coadded to
study, as shown in Figure 1, to directly measure NO2 via UV-
visible absorption in the reactor. The photolysis laser beam was
propagated along the same path as the light beam for absorption
measurements. The beams were interchanged by moving two
mirrors mounted on repositioning mounts. The metal reaction
cell was designed to reduce scattered light and increase the
efficiency for detecting fluorescence.12 A solar blind photo-
multiplier tube (PMT) detector was mounted orthogonal to both
the photolysis laser and resonance lamp to detect the fluores-
cence. A 1 mm thick CaF2 window mounted in front of the
PMT blocked short wavelength radiation (λ < 130 nm) from
reaching the PMT. The volume between the cell window and
the PMT was flushed with N2 to eliminate absorption of the
oxygen atom fluorescence by air. The sensitivity for detection
3
improve the signal-to-noise ratio. Measured O( P) atom temporal
profiles were fit to eq 9 using linear least squares routines to
obtain k′. Such temporal profiles were measured at various NO2
1
5
-3
concentrations over the range (0.04-4) × 10 molecule cm .
The slope of the plots of k′ vs [NO2] yielded the bimolecular
rate coefficient for reaction 1, k1. The bimolecular rate constants
were measured at 15 different temperatures between 220 and
4
12 K.
All experiments were performed under flow conditions (i.e.,
-1
flow velocity ≈ 100 cm s ). The photolysis laser (XeCl excimer
laser, 308 nm) beam passed along the length of the absorption
cell (parallel to the gas flow). The laser was operated at a
repetition rate of 5 Hz. Therefore, the flowing gas sample was
exposed to no more than three photolysis laser pulses before
reaching the detection zone. At the highest laser fluence, about
3
8
-3
of O( P) was nominally ∼5 × 10 atom cm for a 1 s
integration; this was measured by generating a known concen-
tration of O( P) via photolysis of a known concentration of
0
.3% of NO2 was lost due to photolysis. The kinetic data was
3
not corrected for this small loss.
ozone in N2 at 248 nm using pulsed KrF excimer laser of known
The temperature of the Pyrex extension (see below) and metal
cell, which together formed the absorption cell for measuring
the NO2 concentration, were regulated by flowing either cooled
methanol or heated ethylene glycol through their jackets. The
temperature was varied over the range 220-412 K. The
1
fluence. (N2 quickly quenched O( D), formed in ozone pho-
3
tolysis, to O( P).)
3
O( P) for measuring k1 was generated by photolyzing NO2
at 308 nm (XeCl excimer laser) in the presence of N2.

OH + NO2 Reaction Rate Coefficient
J. Phys. Chem. A, Vol. 103, No. 7, 1999 879
temperature of the gas flowing through the cell and the Pyrex
extension were measured with a chromel-alumel thermocouple
under flow conditions identical to those used in measuring k1.
The temperature was measured at various locations using a
retractable thermocouple. The thermocouple was retracted during
kinetic measurements. The measured gas temperature was
estimated to be accurate to 1 K. A temperature gradient from
one end to the other was less than 8 K even at 220 K, the lowest
temperature of the experiments. This temperature gradient was
sufficiently small as not to affect the NO2 concentration via
N2O4 formation or change the effective path length. The
temperature in the region where O atoms temporal profiles were
measured was constant to better than 0.5 K and was known to
better than 0.5 K.
[NO2] Determination Methods. Three methods were used
to determine the NO2 concentration for measuring k1: (1) UV
absorption, (2) flow, and (3) chemical titration. The UV-vis
absorption cross sections of NO2 were determined using flow,
chemical titration, and absolute pressure measurements; hence,
using the UV absorption to determine NO2 concentration was
not truly independent of the other two methods. The combination
of these methods was used to reduce possible systematic errors
associated with the NO2 concentration determination. Each of
these methods is described separately below.
Flow. The rate of change of pressure, dp/dt, in a calibrated
volume was measured using a flow of pure NO2. This flow was
added to the measured flows of the other gases, mostly N2. The
pressure in the reactor and the flow tube were measured with
Figure 2. UV absorption spectra of NO
line) over the range used by the diode array spectrometer. The arrow
indicates the location of the NO feature used in the cross section
determinations (see text for details).
2 2 4
(solid line) and N O (dashed
2
1
00 Torr electronic capacitance manometers. The concentration
concentration of O3 that was added to change the NO2
concentration. In most experiments, the concentration of NO
was varied along with that of O3 to ensure that all the ozone
reacted and, yet, had the minimum amount of NO left over. At
low O3 concentrations, NO2 formed upon addition of NO could
not be measured by absorption because its absorption cross
section is low, approximately a factor of 20 lower than that of
ozone. In these cases, total conversion of O3 to NO2 was
assumed. The concentration of N2O3 was calculated, on the basis
of NO2 was calculated using the measured flow rate of NO2
and other gases, and pressure in the reactor. The N2 flow rates
were measured using calibrated electronic mass flow meters.
This method provided an accurate measurement of NO2
concentrations for levels below those measurable accurately
14
-3
using UV absorption, <1 × 10 molecule cm . This method
was most advantageous for low temperature (T < 230 K)
1
4
measurements where low concentrations of NO2, <1 × 10
-
3
molecule cm , were required to avoid interference from N2O4
formation, reaction 4. The overall accuracy of this method was
controlled by the accuracy in measuring dp/dt, (3%, N2 flow
rate, (3%, and pressure, (1%, as well as the purity of the NO2
sample, >99.9%. Thus the absolute accuracy of NO2 concentra-
tions determined from this method is estimated to be ∼5% by
assuming the above uncertainties to be uncorrelated.
3
of the known equilibrium constants, to be always less than 0.6%
of the NO2 concentration. Only for the measurements at 220 K
were the corrections for N2O4 significant, i.e., 4% at the highest
NO2 concentration used.
The accuracy of this [NO2] determination method is estimated
to be ∼5%. This estimate is determined from the uncertainty
in the O3 absorption cross sections, (2%, and the precision of
the absorption measurements, (3%.
Chemical Titration. This method was used to produce NO2
directly upstream of the absorption cell via the reaction
NO2 UV Absorption. In this method the concentration of NO2
was measured by UV-visible absorption. The collimated output
of a 30 W D2 lamp was passed through the cell (see Figure 1)
and was focused onto the entrance slit of a spectrograph with
a 1024 element diode array detector. The combined path through
the metal reactor and the Pyrex extension was 42.1 cm. A 100
µm entrance slit, resolution of ∼1 nm, was used for all
absorption measurements. The wavelength of the spectrometer
was calibrated using a Hg pen ray lamp and a 10 µm entrance
slit. The NO2 spectrum was recorded between 280 and 420 nm.
A spectrum of NO2 in this wavelength region is shown in Figure
O + NO f O + NO
2
(2)
3
2
where k2(298 K) ) 1.8 × 10-14 cm molecule s . Note that
3
-1 -1 3
the measured NO2 concentration is not dependent on the exact
value of this rate constant because all O3 was converted to NO2.
NO was used in excess over O3, and the reaction was run to
completion (i.e., more than five lifetimes with respect to reaction
2
) in a sidearm reactor (Pyrex, 3 cm i.d. and 20 cm long). The
residence time of the gas in the reactor was about 2 s. The loss
of O3 and formation of NO2 were monitored simultaneously by
UV absorption with the diode array spectrometer. Concentrations
of NO were determined by the dp/dt method, which was
described above for the case of NO2. The concentration of NO
2
. The prominent absorption feature at 413.4 nm was chosen
for NO2 concentration quantification. At this wavelength, the
contribution of N2O4, if present, would be minimal (see Figure
14
-3
2) and the NO cross section is near its maximum value. The
was in the range (1 to 8) × 10 molecule cm . The initial
ozone concentration was calculated using its absorption spectrum
measured using the diode array spectrometer and the known
absorption cross section at 253.7 nm. In some experiments, the
concentration of NO was held constant while varying the
2
NO2 spectrum contains diffuse structure at this resolution and
the cross sections depend on both temperature and resolution.
NO2 absorption cross sections have been previously mea-
3
1,13,14
sured and recently remeasured in several laboratories
using

880 J. Phys. Chem. A, Vol. 103, No. 7, 1999
Gierczak et al.
1
spectrometers with different resolutions. Harder et al. have
provided high-resolution measurements at 216, 230, 238, and
2
93 K. Cross sections at lower resolution were calculated from
these measurements. However, there is limited cross section data
available at temperatures above 298 K. Therefore, we measured
NO2 cross sections over the temperature range 259-385 K using
the same spectrograph used in the kinetic measurements, i.e.,
the same resolution. The measurements of Harder et al. were
used to confirm our low-temperature measurement and help
establish an empirical relationship between the cross section
and temperature.
The NO2 cross sections were determined using known
concentrations of NO2 generated by the titration reaction 11 as
described above. The absorption cross section for NO2 at 308
-
19
2
-1
nm is ∼2 × 10
cm molecule (see Figure 1).
Material. NO2 was prepared by reacting purified NO with
excess O2 which had been passed through a molecular sieve
trap at dry ice temperature. NO2 was collected in a dry ice cooled
trap and purified by trap-to-trap distillation in an excess O2 flow
until a pure white solid remained. The major impurities (<0.1%)
in our sample of NO2 are NO and O2, whose reactions with O
atoms are slow compared to reaction 1. Because k1 is so large,
impurities do not contribute significantly to its measured value,
unless the impurity levels reach a few percent. NO (C. P. Grade)
was purified by passage through the silica gel trap at dry ice
temperature. N2 (UHP, 99.9995%) was used as the carrier gas
in all experiments and was used as supplied.
Figure 3. NO
2
UV absorption cross section at 413.4 nm as a function
of temperature as determined using methods described in the text. This
work (solid circles). Harder et al.1 (solid squares, after resolution
correction). The solid line is the weighted (according to the precision
of the measured values) least-squares fit to all the data yielding
-
19
2
-1
σ413.4(T) ) (9.49-0.00549T) × 10 cm molecule
.
measured concentration of NO2 in the reactor could have no
more than a 5% uncertainty due to its cross section.
Kinetic Measurements. Rate coefficients for reaction 1 were
measured at 15 temperatures over the range 220 to 412 K. The
results of these measurements are listed in Table 1 and are
plotted in Figure 4. A least-squares fit of the data (ln k1 vs 1/T)
Results and Discussion
-
12
exp[(209 ( 15)/T] cm3
yielded k1(T) ) (5.26 ( 0.27) × 10
NO2 Cross-Section Measurements. UV-visible absorption
cross sections of NO2 were determined at five temperatures:
-
1 -1
molecule s , where the quoted errors represent 2σ precision
of the fit and σA ) AσlnA. The 298 K rate coefficient derived
3
85, 348, 323, 298, and 259 K. Absorbances at 4-6 different
-
12
3
-1 -1
from this fit is 10.6 × 10
cm molecule s , and it agrees
NO2 concentrations were measured at each temperature. Beer’s
law was obeyed at all temperatures. The cross-section values
with that measured at this temperature; therefore, we quote our
-
12
3
2
98 K value to be k1(298 K) ) (10.6 ( 0.8) × 10
cm
-
19
2
at 413.4 nm, in units of 10 cm , determined are 7.39 ( 0.07
385 K), 7.53 ( 0.04 (348 K), 7.72 ( 0.04 (323 K), 7.90 (
.06 (298 K), and 7.79 ( 0.11 (259 K). The quoted errors are
-
1
-1
molecule
s
(Table 2).
(
0
Several experimental checks were conducted to ensure that
the measured rate coefficients did not depend systematically
on the experimental conditions. In three room temperature
experiments, the initial (photolytically produced) oxygen atom
concentration was varied by a factor of 20. The rate coefficients
determined in these measurements agreed to within 8%, well
within the estimated accuracy of our measurements.
The pressure was varied between 15 and 105 Torr in two
room temperature experiments. This tested for possible interfer-
ence from the third body reaction
two standard deviations of the slopes of the absorbance vs
concentration plots as determined by linear least-squares
analyses. The absolute uncertainty in the cross-section values
is estimated to be 3% from the measurements in which both O3
and NO2 were measured simultaneously. Our room temperature
cross section at 413.4 nm is ∼2% lower than that obtained from
1
Harder et al. and ∼4% lower than that reported by Vandaele
et al.14 The cross sections measured here and those of Harder
et al. at 413.4 nm are shown in Figure 3. The agreement between
the two sets of measurements is exceptionally good. Our NO2
cross sections are tied to the cross sections of ozone, which is
M
3
O( P) + NO
9
8 NO
(12)
2
3
3
known to very high accuracy (within 2%) . Thus, normalizing
-
14
3
-1 -1
where k12(15 Torr, 298 K) ∼4 × 10
cm molecule
s
other NO2 spectra to our cross sections, after properly accounting
for differences in resolution, would be reasonable. For a critical
evaluation of previous NO2 cross-section determinations readers
are referred to Harder et al.1
-
13
3
-1 -1 3
and k12(105 Torr, 298 K) ) 2.7 × 10
cm molecule s .
Higher pressures were used only at 298 K and above to test for
secondary reactions. Under the low-pressure conditions used
here (∼15 Torr), reaction 12 makes a negligible contribution
A weighted linear least-squares fit to the cross sections at
3
(
2
<1%) to the O( P) decay even at 220 K, where k12(15 Torr,
20 K) ∼1 × 10
413.4 nm measured by Harder et al. and us yields
-
13
3
-1 -1
cm molecule s . Our measured values
of k1 at 15 and 105 Torr total pressure at 298 K agree within
the measurement uncertainties (see Table 1). A single elevated
pressure measurement, 103 Torr, was also performed at 393 K.
The value of k1 determined under these conditions agreed very
well with the value calculated from our Arrhenius fit. These
results confirm that reaction 13 was not significant under our
conditions.
-19
2
σ413.4(T) ) (9.49 - 0.00549T) × 10 cm
(11)
This relationship was used to calculate the NO2 cross sections
at the temperatures used in the kinetic measurements. The cross
section at 413.4 nm was also used to normalize the diode array
spectra. On the basis of our measurements and comparison with
previous reports, we believe that the absorption cross section
of NO2 at 413.4 nm is known to better than 5% and, hence, our
The methods used for the determination of NO2 concentration
are given in Table 1. In most experiments where the chemical

OH + NO2 Reaction Rate Coefficient
J. Phys. Chem. A, Vol. 103, No. 7, 1999 881
3
TABLE 1: Experimental Conditions and Measured Rate Coefficients for O( P) + NO
2
f NO + O
P (Torr)
2 1
(k )
] methoda
+ NO
[O] × 10 atom cm-3
-11
[NO
2
] × 10 molecule cm-3
-14
k × 10 cm molecule s-1
12
3
-1
T (K)
[NO
2
412
393
383
364
348
325
311
297
296
296
296
296
298
271
256.5
247
236.5
234
227
220
O
3
0.4-2.9
1.3-12.7
4.4-37.2
0.3-3.5
4.5-35.9
0.2-1.8
0.8-12.5
10-32.2
0.4-3.7
1.7-28.7
7.6-74.3
0.4-4.6
average
0.15-1.0
3.9-37.7
2.0-16.7
0.1-1.2
15
103
15
15
15
15
15
101
15
105
15
9.09 ( 0.2
9.24 ( 0.3
8.57 ( 0.15
8.96 ( 0.2
9.39 ( 0.12
9.60 ( 0.32
10.8 ( 0.5
10.3 ( 0.2
11.0 ( 0.2
11.0 ( 0.3
10.2 ( 0.3
10.5 ( 0.5
abs
abs
O
abs
3
+ NO
2.0-16.2
0.08-0.6
2.4-36.9
4.6 - 12.3
3.1-32.6
1.9-32.5
4.0-39.2
0.1-1.1
O
3
+ NO
+ NO
abs
O
3
abs
abs
abs
dp/dt
15
b
10.7 ( 0.8
O
abs
3
+ NO/abs
1.9-18.9
1.0-7.6
0.6-2.2
0.3-3.0
0.2-2.1
0.6-2.2
0.1-2.0
2.2-21.7
4.4-33.9
0.2-0.8
0.1-1.0
0.06-0.8
0.2-0.8
0.04-0.6
15
15
15
15
15
15
15
11.6 ( 0.2
12.6 ( 0.2
12.4 ( 0.8
12.6 ( 0.3
12.7 ( 0.5
12.4 ( 0.05
13.7 ( 0.52
O
3
O
3
O
3
+ NO
+ NO
+ NO
dp/dt
+ NO
O
3
a
O
3
+ NO indicates that the NO
was measured via UV absorption at 413.4 nm. dp/dt indicates the NO
of change of pressure in a known volume as NO was flowed into it. (See text for details.) The value was obtained by correcting the values
2
concentration was measured by converting a measured amount of ozone to NO
2
via reaction with NO. abs
indicates that NO
2
2
concentration was measured by determining the time rate
b
2
measured close to 298 K for the difference in temperature using the E/R measured here and averaging them. The quoted uncertainty is 2σ of the
average.
3
Figure 4. The temperature dependence of the rate coefficient for O( P)
Figure 5. Plot of k′ verses [NO
has been corrected for the formation of N
2
] at 220 K. The NO
2
concentration
+
NO
2
2 1
f NO + O (k ) measured in this work. The solid line is the
2 4
O
and contributions to the
weighted (according to the precision of the measured rate constants)
least-squares fit.
O atom loss due to the O + NO reaction. The solid line is a weighted
(according to the precision of the measured k′) least-squares fit to the
-
12
3
data yielding a rate coefficient for reaction 1 of 13.7 × 10
cm
titration method was used, the NO concentration was only in
-1 -1
molecule
and
s .
3
slight excess to limit contributions to the O( P) atom decay from
the reaction
M
3
O + N O f products
(14)
O( P) + NO
9
8 NO2
(6)
2
3
-
12
cm molecule s-1 and k14 <
3
-1
at 199 K to be k13 < 2 × 10
.5 × 10 cm molecule s , respectively. Hence, the
-
14
3
where k6 ∼ 4 × 10
at 298 K and 15 Torr. The measured
-
13
3
-1 -1
4
pseudo-first-order rates were corrected for reaction 6 using the
measured NO concentration. The corrections ranged from 50
to 130 s , which are small (less than 3%) compared to the
first-order loss rate constants due to reaction 1. When a constant
concentration of NO was used, the contribution to measured k′
was an intercept in the k′ vs [NO2] concentration plots.
contributions to the measured O atoms decay rate constants of
reactions 13 and 14 are at least 7 and 30 times smaller than the
contribution from reaction 1 at 199 K. At warmer temperatures,
they are negligibly small. Therefore, reaction 13 and 14 should
not significantly influence the measurement of k1. However, in
experiments at temperatures below 230 K where the NO2
concentration was deduced from flows and chemical titration
rather than measured by UV absorption, we corrected the NO2
concentration for the formation of N2O4. Figure 5 shows a plot
of k′ vs [NO2] at 220 K, where the deduced [NO2] was corrected
-
1
8
Geers-Muller and Stuhl have estimated upper limits for the
rate coefficients for the reactions
O + N O f products
(13)
2
4

8
82 J. Phys. Chem. A, Vol. 103, No. 7, 1999
Gierczak et al.
ref
3
TABLE 2: Summary of Rate Coefficient Measurements for O( P) + NO
2
2 1
f NO + O (k )
k(298 K)a
Aa
-E/R (K)
T Range (K)
experimental techniqueb
9
.12 ( 0.46
8.70 ( 0.65
3.72 ( 0.30
5.53
6.58 ( 0.52
5.21 ( 0.50
6.5
13 ( 21
267 ( 30
154
142 ( 23
202 ( 27
120 ( 120
209 ( 35
210 ( 30
230-339
298-1055
240-296
224-354
233-357
FP-RF
4
6
5
7
8
2,3
9
9
.5 ( 1.1
DF-RF/MS
FP-CL
c
.3 ( 1.4
1
1
0.3 ( 0.9
DF-CL
FP-CL
d
0.3 ( 0.2
e
9
.7
1
1
0.6 ( 0.8
0.6
5.26 ( 0.6
5.22
220-412
PP-RF
this work
e
rec.f
a
Units are 10-12 cm molecule s-1. b FP: flash photolysis. DF: discharge flow. RF: resonance fluorescence. MS: mass spectrometry. CL:
3
-1
chemical luminescence. PP: pulse photolysis. k(296 K). d Calculated from value measured at 301 K using E/R ) -202 K
c
-1 e
.
Uncertainty of
f
f(298) ) 1.1. Uncertainties at other temperatures are calculated using f(T) ) f(298) exp|∆E/R{(1/T) - (1/298)}|. Recommended based on results
of this study and refs 7 and 8.
3
for the formation of N2O4 using the equilibrium constant. These
corrections were small and amounted to a change in rate constant
of ∼7%. Therefore, even if the equilibrium constant were
somewhat uncertain, the rate coefficients would not change
much.
manometrically prepared stock mixtures. Considering these
differences the agreement with our results is surprisingly good.
Due to the limitations in the precision of their measurements
the weak temperature dependence of k1 eluded their detection.
It is very important to note that Davis et al.’s experiments were
excellent for that period, and we have learned much since then
as to how to handle gases such as NO2.
The fits of k′ vs [NO2], and hence the obtained values of k1,
were precise. However, this is not a true representation of the
reproducibility between measurements. The expression for k1
given above does include, to a first approximation, the system-
atic errors in the NO2 concentration measurements because
multiple methods were used. Yet, we include other estimated
systematic errors, derived for each NO2 concentration method,
and suggest the following expression for k1 as a function of
temperature:
6
Bemand et al. studied reaction 1 under pseudo-first-order
conditions using a discharge flow tube and detected O atoms
via resonance fluorescence (in an excess of NO2) and NO2 via
mass spectrometry (in excess of O atoms) to determine k1 at
sixteen temperatures between 298 and 1055 K. In experiments
with NO2 in excess, the NO2 concentration was varied over the
13
-3
range (0.4-4) × 10 molecule cm . NO2 concentrations were
determined using flow measurements. Under these conditions,
N2O4 formation was negligible. In experiments with O( P) atoms
in excess, the O( P) atom concentration was varied over the
-12
3
k1(T) ) (5.26 ( 0.60) × 10
×
3
exp[(209 ( 35)/T] cm molecule s-1
3
-1
1
2
-3
range (4 to 26) × 10 atom cm . The rate coefficient data
measured at each temperature shows considerable scatter with
values differing by as much as 50%. Fitting the data given in
their Table 1 (weighting the 298 K value by a factor of 20, the
number of 298 K values measured by them) yields k1(T) ) (3.68
( 1.0) × 10 exp[(274 ( 80)/T] cm molecule s , k1(298)
) 9.3 × 10 cm molecule s , where the error limits are
σ precision of the fit. This fit yields values of k1 which are
There are a number of relatively recent kinetic studies4-8 of
reaction 1 with which to compare and contrast our results. An
overview of our results and those from previous works is given
in Table 2. The recent report of k1(298 K) ) (9.3 ( 1.0) ×
-12
3
-1 -1
-
12
3
-1 -1
15
1
0
cm molecule
s
reported by Paulson et al. is not
-
12
3
-1 -1
4
included in the table. Davis et al. studied reaction 1 using
broadband flash photolysis of NO2 to produce O atoms and
vacuum UV resonance fluorescence to detect them. They
measured k1 over the temperature range 230 to 339 K. The range
of NO2 concentrations used was sufficiently low that corrections
for N2O4 were small even at the lowest temperatures used. They
determined the NO2 concentrations using manometrically
prepared gas mixtures in Ar, N2, O2, and CO2 used as the buffer
gas. (Their stock mixtures necessarily had much higher con-
centrations of NO2 than those present in the reactor.) The details
of the mixture preparation methods was not provided (i.e., the
corrections for N2O4 in the stock mixture are not known and
could have led to small overestimation of NO2 and, hence, the
1
systematically lower than our measurements, ∼6%, over our
temperature range. However, these differences are small com-
pared to the accuracy of the measurements reported by Bemand
et al. Also, the possibility of thermal decomposition of NO2 in
their system at high temperatures could have led to the higher
negative temperature dependence that they reported.
5
Slanger et al. measured k1 using pulsed photolysis of O2 at
3
1
47 nm to generate O( P) atoms in the presence of excess NO2.
3
O( P) atoms were detected by reacting it with NO and measuring
the chemiluminescence from the NO2 product. NO2 concentra-
1
3
-3
tions in the range (0.6-6) × 10 molecule cm were used.
-
12
NO concentrations were determined by UV absorption of a
lower values of k1). They report k1 ) (9.12 ( 0.44) × 10
2
3
-1 -1
2 2
NO /N mixture prior to being diluted in the reaction gas
mixture. Small corrections were applied to account for N2O4
formation. Measurements were made at 296 and 240 K yielding
cm molecule independent of temperature. The quoted
s
error limit is 1σ of the measurement precision and does not
include possible systematic errors. This value is about 10%
smaller than our room temperature value but within the
combined 2σ uncertainty limits. The experimental technique
used by Davis et al. was similar to that used in our work, but
with a few significant differences. In our experiments the gas
mixture was flowed through the reactor, a laser was used to
photolyze NO2 at a single wavelength, the mixture was exposed
to at most two photolysis pulses, and NO2 was directly measured
in the reactor in most cases. Davis et al. used a static mixture,
broadband photolysis, and exposure of the mixture to multiple
photolysis pulses, and deduced the concentration of NO2 using
-
12
3
-1 -1
k1 values of (9.3 ( 1.4) × 10 cm molecule
s
and (10.5
-12
3
-1 -1
(
1.6) × 10 cm molecule s , respectively. These results
indicate a small negative temperature dependence and are
systematically less than our results but are within the combined
uncertainties of the two data sets.
7
Ongstad and Birks used a discharge flow tube reactor and
also detected O atoms by the chemiluminescence method. They
measured k1 at six temperatures in the range 224 to 354 K at
2.5 Torr total pressure. Their NO2 concentrations were calculated
from measured flow rates of manometrically prepared gas

OH + NO2 Reaction Rate Coefficient
J. Phys. Chem. A, Vol. 103, No. 7, 1999 883
-
12
k1(T) ) (5.22 ( 0.50) × 10
×
exp[(210 ( 26)/T] cm molecule s-1
3
-1
where the errors are 2σ of the fit and σA ) A σlnA. The fit is
also shown in the figure. This expression may be used for
modeling studies. It should be noted that this lies beyond the
1
σ error bounds indicated by 1997-NASA/JPL evaluation. The
numbers that can be used for modeling studies are listed in the
last row of Table 2.
Atmospheric implications. The rate coefficient obtained in
this study is approximately 20-30% higher than that derived3
in the current kinetic evaluations at stratospheric temperatures.
Given that reaction 1 is the rate-limiting step in the major NOx
catalyzed ozone destruction cycle (see Introduction), this change
in the rate coefficient will have a significant impact on the
calculated stratospheric ozone abundance. Further, increases in
k1 will alter the calculated ozone depletion due to chlorine and
the impact of aircraft emissions on ozone levels. We have
discussed some these consequences in a separate paper16 which
reexamines the role of NOx in the stratosphere in light of the
changes in the rate coefficients for the reactions of OH with
HNO3 ref 17 and NO2 (ref 18) as well as k1.
Figure 6. Comparison of recent reaction 1 rate coefficient measure-
ments and current value recommended for stratospheric model calcula-
tions (dashed line). The heavy solid line is a weighted least-squares fit
to the data from this work (solid squares) and that of Ongstad and
7
8
Birks (open circles) and Geers-Muller and Stuhl (solid bow tie)
-
12
yielding a value of k
1
(T) ) (5.22 ( 0.50) × 10 exp[(210 ( 30)/T]
3
-1 -1
cm molecule
s .
Acknowledgment. We thank J. Harder for useful discussions
and spectroscopic calculations. This work was supported in part
by NASA Upper Atmospheric Research Program.
mixtures. NO2 concentrations were varied over the range (1-
12
-3
1
0) × 10 molecule cm . These low NO2 concentrations made
corrections for N2O4 formation negligible at all temperatures.
-
12
They report k1(T) ) (6.58 ( 0.52) × 10
exp[(142 ( 23)/T]
3
-1 -1
-11
3
-1
References and Notes
cm molecule s , k1(298) ) 1.06 × 10
cm molecule
-1
s , where the error limits are 1σ and represent the precision of
the measurements only. Their room temperature value is in
excellent agreement with the present results while the temper-
ature dependence, E/R, is smaller but lies within the combined
(1) Harder, J. W.; Brault, J. W.; Johnston, P. V.; Mount, G. H. J.
Geophys. Res. 1997, 102, 3861.
(2) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F.; Kerr, J.
A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1997, 26, 521.
(3) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
2
σ uncertainty limits.
Geers-Muller and Stuhl used pulsed H2 laser (∼160 nm)
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
3
photolysis of NO to produce O( P) atoms in the presence of
NO2. They measured the temporal profile of O( P) using the
3
(
30.
4) Davis, D. D.; Herron, J. T.; Huie, R. E. J. Chem. Phys. 1973, 58,
chemiluminescence method noted above and obtained k1 at five
temperatures over the range 233-357 K using NO2 concentra-
5
(5) Slanger, T. G.; Wood, B. J.; Black, G. Int. J. Chem. Kinet. 1973,
14
-3
tions in the range (0.5-4) × 10 molecule cm . NO2
concentrations were calculated using mass flow rates measured
with calibrated flow controllers. Corrections to the NO2
concentration due to N2O4 formation were less than 4% under
5, 615.
(6) Bemand, P. P.; Clyne, M. A. A.; Watson, R. T. J. Chem. Soc.,
Faraday Trans. 2 1974, 70, 564.
7) Ongstad, A. P.; Birks, J. W. J. Chem. Phys. 1984, 81, 3922.
(8) Geers-Muller, R.; Stuhl, F. Chem. Phys. Lett. 1987, 135, 263.
9) Dubey, M. K.; Smith, G. P.; Hartley, W. S.; Kinnison, D. E.;
Connell, P. S. Geophys. Res. Lett. 1997, 24, 2737.
10) Nicovich, J. M.; Wine, P. H.; Ravishankara, A. R. J. Chem. Phys.
1988, 89, 5670.
11) Vaghjiani, G. L.; Ravishankara, A. R. Int. J. Chem. Kinet. 1990,
2, 351.
12) Goldfarb, L. The photochemistry and kinetics of chlorine com-
(
1
5
-3
these conditions. The addition of NO, 8 × 10 molecule cm ,
(
in the detection region led to the formation of N2O3 via
(
NO + NO + M T N O + M
(15)
2
2
3
(
2
However, corrections to the NO2 concentration were less than
(
1
.4% under these conditions. They report k1(T) ) (5.21 ( 0.50)
pounds important to stratospheric mid-latitude ozone destruction. Ph.D.
Thesis, University of Colorado, 1997.
-12
3
-1 -1
×
10 exp[(202 ( 27)/T] cm molecule s , k1(298) ) 1.03
-
11
3
-1 -1
×
10 cm molecule s , where the quoted error limits are
(
13) Harwood: M. H.; Jones, R. L. J. Geophys. Res. 1994, 99, 22955-
22964.
(14) Vandaele, A. C.; Hermans, C.; Simon, P. C.; Carleer, M.; Colin,
3
σ and include estimated systematic errors. This value is in
excellent agreement with our results with k1 at 233 K only 5%
less than the value derived from the present results.
R.; Fally, S.; Merienne, M. F.; Jenouvrier, A.; Coquart, B. J. Quant.
Spectrosc. Radiat. Transfer 1998, 59, 171.
From the above description, it is clear that there are reasons
to suspect that the data of Davis et al. may not be highly
accurate, especially at low temperatures. The data of Ongstad
and Birks and of Geers-Muller and Stuhl are quite accurate.
Figure 6 shows a plot of the data obtained by these two groups,
along with those from our study. The data from these three
studies were fit to the Arrhenius expression using an unweighted
linear least squares routine (ln k1 vs 1/T) to obtain:
(
15) Paulson, S. E.; Orlando, J. J.; Tyndall, G. S.; Calvert, J. G. Int. J.
Chem. Kinet. 1995, 27, 997.
(16) Brown, S.; Gierczak, T.; Portmann, R. W.; Talukdar, R. K.;
Burkholder, J. B.; Ravishankara, A. R. Geophys. Res. Lett. Submitted for
publication.
(
17) Brown, S.; Talukdar, R. K.; Ravishankara, A. R. J. Phys. Chem.,
998. Submitted for publication.
18) Brown, S.; Talukdar, R. K.; Ravishankara, A. R. Chem. Phys. Lett.
1999, 299, 217.
1
(