Rate constants for the reaction OH+NO2+M → HNO3+M under atmospheric conditions

Article abstract of DOI:10.1016/S0009-2614(98)01283-4

We present rate constants for the title reaction, k₁, measured using a pulsed photolysis-laser induced fluorescence technique between 220 and 250 K in 20-250 Torr of N₂ and 20-50 Torr of O₂. Our measured k₁ agree with literature values at low temperatures but show that the current recommendations for atmospheric modeling overestimate k₁ by 10-30% in the falloff region below 250 K. The revised values of k₁ help to better define the role of NO₂ in the stratosphere.

Full text of DOI:10.1016/S0009-2614(98)01283-4

11 January 1999
Ž
.
Chemical Physics Letters 299 1999 277–284
Rate constants for the reaction OHqNO2 qM ™ HNO3 qM
under atmospheric conditions
a,b,c
Steven S. Brown a,), Ranajit K. Talukdar a,b, A.R. Ravishankara
a
Aeronomy Laboratory, NOAA, 325 Broadway, Boulder, CO 80303, USA
b
CooperatiÕe Institute for Research in the EnÕironmental Sciences, UniÕersity of Colorado, Boulder, CO 80309, USA
c
Department of Chemistry, UniÕersity of Colorado, Boulder, CO 80309, USA
Received 22 October 1998; in final form 3 November 1998
Abstract
We present rate constants for the title reaction, k1, measured using a pulsed photolysis–laser induced fluorescence
technique between 220 and 250 K in 20–250 Torr of N2 and 20–50 Torr of O2. Our measured k1 agree with literature
values at low temperatures but show that the current recommendations for atmospheric modeling overestimate k1 by
10–30% in the falloff region below 250 K. The revised values of k1 help to better define the role of NO2 in the
stratosphere. q 1999 Elsevier Science B.V. All rights reserved.
1. Introduction
most important reactions that converts NOx into its
reservoir is that of hydroxyl radicals with NO2.
Nitrogen dioxide is central to the chemistry of the
atmosphere. In the stratosphere it directly partici-
pates in catalytic ozone destruction, suppresses ozone
destruction due to other chemical cycles by forming
OHqNO2 qM ™ HNO3 qM .
1
Ž .
The pressure dependence of the rate constant for this
Ž
.
Ž .
reservoir species e.g., ClONO2 , HONO2 , HO2 NO2
association reaction k1 lies in the falloff region
between second- and third-order kinetics over the
pressure and temperature range encountered in the
troposphere and stratosphere. Accurate characteriza-
tion of the rate constant behavior under atmospheric
conditions thus requires extensive measurements as a
function of both pressure and temperature in N2 and
O2.
and directly affects the partitioning of compounds
within different chemical families. In the tropo-
sphere, one of its key roles is to photochemically
produce ozone. The determining factor for NO2
chemistry is the fraction of the total nitrogen oxides
present in the active form of NO and NO2 , collec-
tively referred to as NOx , compared to the fraction
Ž
.
Ž .
present in inactive forms e.g., HNO3 . One of the
Surprisingly, the database for reaction 1 is quite
limited over the temperature and pressure range
characteristic of the lower stratosphere and upper
Ž
.
troposphere 200–250 K, 10–300 mbar , and the fits
)
w x
w x
that come from the NASArJPL 1 and IUPAC 2
recommendations for atmospheric modeling are
Corresponding author. Fax: q1-303-497-5822; e-mail:
sbrown@al.noaa.gov
0009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 98 01283-4

(
)
278
S.S. Brown et al.rChemical Physics Letters 299 1999 277–284
X
w
x
heavily weighted with rate constant data at higher
temperatures. The low-temperature database for Ms
N2 consists of two studies, that of Anastasi and
to a plot of k vs. NO2 gives the effective bimolec-
ular rate constant at a fixed pressure, k1, as the
slope, and the first-order loss rate constant in the
absence of NO2 , kd, as the intercept. The reaction of
OH with HNO3 or H2O2 is the most important
contribution to kd in these experiments. Nitric acid
was the OH precursor for all rate constant measure-
ments reported here; however, we checked that the
measured rate constants were invariant to the OH
source by also using H2O2 photolysis as an OH
source at 250 and 296 K. In all cases, kd was
;150–250 sy1, while the estimated first-order loss
rate constant due to diffusion and flow out of the
reaction zone was 20–50 sy1. The linear gas flow
velocity through the 150 cm3 reaction cell was 7–8
cm sy1, enough to refresh the gas mixture in the
reaction zone every 1–2 laser shots at 10 Hz. A
factor of two variation in the linear flow velocity
produced no change in the measured rate constants.
w x
Smith 3 at 220 and 238 K and that of Wine et al.
w x
4 at 247 K. Both of these studies measured rate
constants in a variety of bath gases, including N2 ,
but neither made measurements with MsO2 at low
w
x
temperatures. Numerous other studies 5–10 of reac-
Ž .
tion 1 have examined rate constants at temperatures
closer to 298 K andror in the low- and high-pressure
limits, and in bath gases other than N2.
Ž .
The atmospheric significance of reaction 1 and
the lack of data under the conditions of interest, in
addition to recent discrepancies between calculated
and observed NOx concentrations in the atmosphere
w
x
11 , warrant further study of k1 at low temperatures
and sub-atmospheric pressures in N2 and O2. We
report rate constant measurements over the pressure
range 20–250 Torr in N2 and O2 bath gases and at
temperatures from 220 to 250 K and at room temper-
ature.
We measured nitric acid concentrations by absorp-
tion at 185 nm ss1.63=10
y17
cm2 14 in a 100
Ž
w x.
cm external absorption cell using a mercury lamp
and a solar blind photodiode and verified that the
nitric acid concentrations, which were typically ;5
=1014 cmy3, were consistent with the observed
2. Experiments and data analysis
w
x
The pulsed-photolysis laser-induced fluorescence
values of kd 13 . The first-order loss rate constants
due to reaction with HNO3 remained the same be-
fore and after measurement of a series of kX values.
Nitrogen and oxygen bath gases were from Scott
Specialty Gases and had stated purities of 99.9995%
and 99.99%, respectively. We used them without
further purification. We synthesized NO2 from the
reaction of NO with excess O2 and purified the
product by repeated freeze–pump–thaw cycles at 77
K. We made bulbs of 0.2–1% NO2 in He and
measured the concentration of the mixtures via ab-
sorption near 365 nm using an Hg lamp, a 100 cm
cell, and a bandpass filter centered at 365 nm with a
FWHM of "10 nm. We took the NO2 absorption
cross-section at the pair of Hg emission lines near
Ž
.
LIF apparatus has been extensively used in previ-
w
x
ous OH kinetic studies in this laboratory 12,13 . We
briefly describe the apparatus here to emphasize
some of the features and changes made to accurately
measure k1. Photolysis of either HNO3 or H2O2 at
Ž
.
248 nm KrF excimer laser produced an initial OH
11
concentration, OH , of typically 10 cmy3, and
w
x
0
we monitored the OH decay over 2–3 orders of
magnitude via LIF. We measured k1 under pseudo-
w
x
w
x
first-order conditions in OH, i.e., M 4 NO2 4
w
x
OH , with the minimum NO2 concentration ap-
0
proximately two orders of magnitude larger than
w
x
w x
OH . At a fixed pressure, M , the OH temporal
profile is given by:
0
w x
365 from the work of Wine et al. 4 : ss5.75=
w
w
x
x
OH
OH
St
sln sy k NO qkd tsykX t ,
10y19 cm2 "3% accuracy . The absorption mea-
surement agreed with a manometric measurement of
the NO2 concentration to within the cross-section
Ž
.
w
x
ln
Ž
.
1
2
S0
2
Ž .
w
x
uncertainty. We determined NO2 in the reaction
cell by measuring the NO2rHe flow relative to the
total flow using calibrated electronic mass flow me-
X
w
x
k sk1 NO2 qkd ,
where S refers to the OH LIF signal and kX is the
3
Ž .
w x
pseudo-first-order OH loss rate constant. A linear fit
ters. At the highest NO2 measurable using the 10

(
)
S.S. Brown et al.rChemical Physics Letters 299 1999 277–284
279
Ž
.
. w
x
SCCM standard cubic centimeters per minute flow
meter for the HerNO2 mixture, we checked the flow
measurement by absorption. The absorption and flow
flow measurement alone , NO2 is the actual con-
centration in the low-temperature reaction cell and
Keq is the equilibrium constant at the reaction cell
temperature. The solid points in Fig. 1 are the first-
w
x
measurement gave the same NO2 , although the
flow measurement had significantly higher precision
because of the small NO2 absorbances.
Dimerization of NO2 to form N2O4 is important
at low temperatures 1 . The extent of this equilib-
w
x
order OH loss rate constants vs. NO2 0, and the
Ž .
w
x
solid line is a fit to Eq. 3 with NO2 coming from
Ž .
Eq. 4 . The value of Keq is fixed at its literature
value, 1 and k1 and kd are varied to produce the
w x
w x
rium limited the maximum NO2 concentrations and
thus the maximum measurable first-order OH loss
rate constant. To demonstrate the effect of this equi-
librium and the resulting correction to obtain k1, we
fit. The open circles are the first-order OH loss rate
constants vs. the corrected NO2 concentration given
Ž .
by Eq. 4 , and the dashed line is the resulting linear
Ž .
fit to Eq. 3 . The agreement of the calculated line
w
x Ž
.
carried out measurements at high NO2 Fig. 1 . The
with the measured k1 in Fig. 1 shows that N2O4
does not react as rapidly with OH as NO2 and that
the correction for N2O4 can be carried out quite
accurately. The straight line fit in Fig. 1 reproduces
the value for k1 in Table 1 for identical temperature
w
x
maximum NO2 in the figure is more than an order
of magnitude larger than that actually used to mea-
Ž
.
sure k1 Table 1 . According to the equilibrium,
2NO2 | N2O4, the following expression gives the
NO2 concentration:
Ž
.
and pressure 220 K, 50 Torr N2 but much lower
w
x
NO2 to within 3%, a deviation smaller than the
w
2 x
4Keq
8 NO Keq q1 y1
(
s ž
x
/
0
experimental uncertainty. For the rate constants in
Table 1, the correction due to the N2O4 equilibrium
was -10% at 220 K and significantly smaller at
higher temperatures. Therefore, any error introduced
by the correction is negligibly small.
w
NO2
,
4
Ž .
w
x0
where NO2 is the measured NO2 concentration in
Ž
the absence of the N2O4 equilibrium i.e., from the
Fig. 1. Effect of the N2O4 equilibrium at 220 K on the first-order OH loss rate constant, kX. Solid points are kX vs. NO2 as measured by
w
x
Ž
.
mass flow outside the low-temperature reactor i.e., uncorrected for the N2O4 equilibrium , and the solid line is a non-linear least-squares fit
to the equation shown in the figure. Open circles are for kX vs. the calculated NO2 concentration in the reactor after accounting for N2O4
Ž .
Ž
.
,
and the dashed line is a least-squares fit to Eq. 3 .

(
)
280
S.S. Brown et al.rChemical Physics Letters 299 1999 277–284
Table 1
insignificant contributor to the observed OH loss rate
Measured OHqNO2 rate constants
constants. To verify the lack of any secondary chem-
istry arising from NO2 photolysis products, we var-
ied the excimer laser fluence by a factor of 4 and
observed that it did not influence the measured rate
constants. In experiments using H2O2 as an OH
precursor, OH regeneration can occur through the
reaction of NO with HO2 , generated from reaction of
OH with H2O2. This reaction sequence can regener-
12
T
P Torr
k 10 cm3 sy1
Ž .
K
Ž
.
Ž
.
MsN2
MsO2
220
230
240
20
30
50
3.30"0.26
4.52"0.28
6.02"0.37
7.81"0.47
8.69"0.58
11.37"0.75
13.28"0.99
2.44"0.16
3.53"0.21
4.92"0.30
75
100
170
250
w
x
ate at most 1% of OH , and much less under most
experimental conditions. The regeneration would be
0
20
30
50
2.86"0.20
3.98"0.24
5.20"0.32
6.91"0.42
7.56"0.48
10.54"0.65
12.15"0.89
2.47"0.15
3.28"0.20
4.26"0.25
Ž
.
significant if there were appreciable )0.1% NO
impurity in the NO2 , but the observation of identical
rate constants using HNO3 and H2O2 precursors
indicates that this is not the case.
75
100
170
250
20
30
50
2.49"0.15
3.24"0.20
4.84"0.29
5.87"0.36
7.06"0.44
9.12"0.62
10.23"1.11
1.87"0.11
2.59"0.16
3.53"0.21
3. Results and discussion
75
Table 1 presents the OHqNO2 effective bi-
100
170
250
Ž
.
molecular rate constants k1 in 20–250 Torr of N2
and 20–50 Torr of O2 at temperatures between 220
and 296 K. Each rate constant is an average of 2–3
individual measurements of k1. The quoted errors
are the averages of the 2s precision from the slopes
of the kX plots combined with the uncertainties in the
NO2 concentration measurements arising from the
3% uncertainty in the 365 nm NO2 absorption
cross-section and the 5% uncertainty associated with
mass flow meter calibrations. The temperatures in
Table 1 are accurate to "0.5 K, and the uncertainty
in the pressures varies between "1 and 0.4% from
20 to 250 Torr.
250
296
20
40
70
135
250
2.33"0.15
3.62"0.22
4.90"0.34
7.48"0.46
9.26"0.79
1.86"0.11
2.90"0.17
20
50
100
200
1.12"0.08
2.11"0.14
3.24"0.21
4.48"0.73
0.93"0.08
1.72"0.12
The error bars are at the 95% confidence limit and include the
estimated systematic errors described in the text.
Fig. 2 compares our measured rate constants at
220, 240, 250 and 296 K to the literature data
available at these temperatures and to the current
Additional potential error sources include the re-
Ž
w x
action of OH with the photolysis products of NO2
O
NASArJPL recommendation
1 for the pressure
.
.
Ž
and NO . The NO2 absorption cross-section has a
minimum at the photolysis laser wavelength, 248
nm, having a room-temperature value of 1=10y20
falloff of k1. At 296 K Fig. 2a , our data in N2
agree within the experimental uncertainty with all
other measurements, including those of Anastasi and
cm2 15 . At a typical maximum NO2 concentration
Smith 3 , Wine et al. 4 and the recent, extensive,
room-temperature measurements of Donahue et al.
w
x
w x
w x
2=10 cmy3 and 248 nm excimer laser fluence
14
Ž
Ž
.
y2
.
w
x Ž
7.5 mJ cm , the concentration of photolysis prod-
10 . For clarity we have displayed only the fit
ucts is approximately O s NO s3.5=10 cmy3
.
resulting from the latter study as the dashed line.
10
w x
w
x
.
Reaction with NO2 converts the O atoms rapidly into
NO; however, the resulting NO concentration is still
Our limited 298 K measurements are intended only
to show the agreement with the large room-tempera-
ture literature data set. Interestingly, however, the
w
x
less than OH , and reaction with NO remains an
0

(
)
S.S. Brown et al.rChemical Physics Letters 299 1999 277–284
281
Ž .
Ž .
Ž .
Ž .
Fig. 2. Comparison of our measurement of the pressure falloff of k1 with literature data at: a 296 K; b 250 K; c 240 K; and d 220 K.
rate constants in O2 between 20 and 50 Torr are 20%
smaller than those in N2 , despite the similar molecu-
lar weights of the two gases. The only literature data
point with which we can compare the lower effi-
ciency of O2 is the 25 Torr point of Anastasi and
Smith at room temperature, where the rate constant
in O2 is similarly smaller relative to N2. Burrows et
there is no apparent mechanism for OH regeneration
in O2. The 248 nm photolysis light is not energetic
enough to photolyze ground state O2 , and the O2
y24
Ž
absorption cross-section is too small -10
cm2
.
to yield significant quantities of O atoms even at the
w
x w x
highest O2 1 . Furthermore, the lack of any mea-
surable dependence of k1 in O2 on the photolysis
laser fluence ruled out any multiphoton dissociation
of O2. Photolysis of HNO3 at 248 nm produces
w x
al. 8 measured the third-order, low-pressure limit-
ing rate constant to be very slightly smaller in O2
than in N2 , but not outside the uncertainties in the
rate constants.
w
x
-0.2% H atoms 16 , so OH regeneration via pro-
duction of HO2 is not significant.
The smaller rate constants with MsO2 do not
appear to result from an experimental artifact since
Our measurements at 250 K also agree to within
the combined uncertainty with those of Wine et al.

(
)
282
S.S. Brown et al.rChemical Physics Letters 299 1999 277–284
Table 2
rate constants in N2. This trend continues as the
temperature decreases, as Fig. 2c shows. At 240 K,
our data for MsN2 agree with Anastasi and Smith
but differ from the NASArJPL recommendation by
10–20% at the low end of the pressure range. At 220
K, the agreement between our measurements and
Anastasi and Smith is not as good, although our data
lie within their "18% error limits. Our measure-
ments for MsN2 are 15–25% smaller than the
NASArJPL recommendation.
Ž
Ž . Ž ..
Fit parameters for rate constant data Eqs. 5 and 6
k0300
n
2.47=10y30
2.97
k`300
m
1.45=10y11
2.77
1
0.70
0.94
Ž
.
.
b N2
Ž
b O2
Ž
.
b air
These parameters fit the data in Table 1 to within "8% at the
95% confidence level.
The present study greatly increases the database
w x
Ž
.
Ž .
4 at 247 K Fig. 2b . Once again the rate constants
for reaction 1 in the falloff region relevant for the
in O2 are ;20% smaller than those in N2. The
atmosphere below 250 K. To provide a rate constant
expression that is useful for lower stratospheric
chemical models, we have fit our data for MsN2
w x
NASArJPL recommendation 1 at this temperature
appears to be a slight overestimate of the measured
Ž
Ž .. Ž .
of our data and selected literature data solid lines to the current NASA
Fig. 3. Comparison of the pressure falloff in air Eq.
5
Ž
.
recommendation dashed lines at four different temperatures. Data from this work for MsN2 and O2 appear as closed and open points,
respectively. The N2 and O2 data are plotted against bi M , where bi is the collisional efficiency for each gas relative to air, which is taken
w
x
as unity in this plot.

(
)
S.S. Brown et al.rChemical Physics Letters 299 1999 277–284
283
Ž .
and O2 in the falloff region along with data for
w x
high- or low-pressure limits for reaction 1 , but we
do expect it to be valid over the pressure and temper-
ature range of interest in the lower and middle
atmosphere.
MsN2 from Wine et al. 4 , Donahue et al. at 298
w
x
K 10 , and Anastasi and Smith at 238 K and above
w x
3 . The fit does not include any data in the high- or
low-pressure limits but does encompass the tempera-
ture range 220–358 K. The fit function is the semi-
The amount of NOx present in the atmosphere
depends on the relative rates for formation and de-
struction of reservoir species such as HNO3. Reac-
w x
empirical formula to describe the pressure falloff 1 .
Ž .
tion 1 is the most important gas-phase reaction that
w x
k0 T b M
Ž .
converts NOx into HNO3. Additional reactions in-
volving formation and heterogeneous hydrolysis of
ks
ž 1qk
/
w x
T b M rk
T
`Ž .
0 Ž .
w
x
10Žk0ŽT .b wMxr k`.x24y1
N2O5 also contribute 17 . The results of this work,
combined with recent measurements of the rate con-
=0.6Ä1qwlog
,
5
Ž .
yn
ym
w
x
T
T
stant for reaction of OH with HNO3 13 , an impor-
tant gas-phase process for NOx regeneration from
the nitric acid reservoir, have increased the calcu-
lated ratio of NOx to total odd nitrogen species and
have brought atmospheric models into better agree-
k0 T sk300
,
k` T sk300
.
Ž .
Ž .
0
`
ž 300 /
ž 300 /
6
Ž .
Ž .
5
Eq.
is identical to the fit function from the
w x
w
x
NASArJPL recommendation 1 except for the pa-
rameter b, which we have introduced in order to
account for the difference in collisional stabilization
efficiency between N2 and O2. We first fit the much
larger data set for MsN2 , fixing b at unity. We
then fit the limited data for MsO2 using the N2
coefficients but allowing b for O2 to vary. Table 2
lists the fit parameters. We find the collisional effi-
ciency for O2 relative to N2 to be 0.70"0.04. The
collisional efficiency of air as a third body is the
ment with observations 11 . The new OHqNO2
and OHqHNO3 rate constants also influence calcu-
lated ozone loss rates. In conjunction with new
measurements of rate constants in the NOx catalytic
w
x
ozone destruction 18 , they have resulted in a
reevaluation of the role of NOx in the lower strato-
w
x
sphere 19 .
Note added in proof: While this manuscript was in
preparation, we became aware of rate constants for
this reaction in 50–150 Torr of N2 from 212.5 to
Ž
.
Ž
.
average of b N2 and b O2 weighted by their
Ž
.
relative abundances 0.80 and 0.20, respectively , or
b air s0.94"0.01. Table 2 presents all of the
coefficients resulting from the fit.
w
x
310 K measured recently by Dransfield et al. 20 .
Their data, within the quoted uncertainties, agree
with our fits.
Ž
.
Fig. 3 compares the pressure falloff curves from
our fit for dry air as the third body to the current
w x
NASArJPL recommendation 1 at 200, 220, 240
Acknowledgements
and 296 K. Our data for MsN2 and O2 also appear
with their associated number densities scaled to a
We thank Jim Burkholder for providing samples
of NO2 and assisting in checking their purity. SSB
thanks the National Research Council for a postdoc-
toral research fellowship during this work. This work
was funded in part by the NOAA Climate and
Global Change research program.
Ž
collisional efficiency of unity for air i.e., with
Ž
.
Ž
.
b N2 s1r0.94s1.06 and b O2 s0.70r0.94s
0.74 . The NASArJPL recommendation agrees
.
closely with our fit at room temperature, but overes-
timates the rate constants by 33% at 10 Torr and 200
K and by 14% at 100 Torr at this temperature. The
solid line at 200 K in Fig. 3 is an extrapolation since
there are no measurements at this temperature. The
fit presented in Table 2 and Fig. 3 is an empirical
one intended to be useful for atmospheric modeling
of the upper troposphere and lower stratosphere. We
do not anticipate that it will accurately recover the
References
w x
1
W.B. DeMore, S.P. Sander, D.M. Golden, R.F. Hampson,
M.J. Kurylo, C.J. Howard, A.R. Ravishankara, C.E. Kolb,

(
)
284
S.S. Brown et al.rChemical Physics Letters 299 1999 277–284
M.J. Molina. Chemical Kinetics and Photochemical Data for
A.R. Ravishankara, R. Portman, P.A. Newman, M.K.W. Ko,
Ž .
Geophys. Res. Lett. 1998, submitted .
12 G.L. Vaghjiani, A.R. Ravishankara, J. Phys. Chem. 93 1989
Use in Stratospheric Modeling, Jet Propulsion Lab., Pasadena,
CA, 1997, JPL Publ. 97-4.
R. Atkinson, D.L. Baulch, R.A. Cox, R.F. Hampson Jr., J.A.
Kerr, M.J. Rossi, J. Troe, J. Phys. Chem. Ref. Data 26
w
w
x
Ž
.
w x
2
1948.
x
13 S.S. Brown, R.K. Talukdar, A.R. Ravishankara, J. Phys.
Ž
.
Ž .
Chem. 1998, submitted .
1997 521.
w x
w
x Ž .
14 F. Biaume, J. Photochem. 2 1973r1974 139.
3
C. Anastasi, I.W.M. Smith, J. Chem. Soc., Faraday Trans. 2
Ž
.
w
x
15 W. Schneider, G.K. Moortgat, G.S. Tyndall, J.P. Burrows, J.
72 1976 1459.
w x
4
Ž .
Photochem. Photobiol. A 40 1987 195.
P.H. Wine, N.M. Kreutter, A.R. Ravishankara, J. Phys.
Ž
.
w
x
Chem. 83 1979 3191.
16 A.A. Turnipseed, G.H. Vaghjiani, J.E. Thompson, A.R. Rav-
w x
5
Ž .
ishankara, J. Chem. Phys. 96 1992 5887.
J.G. Anderson, J.J. Margitan, F. Kaufman, J. Chem. Phys. 60
Ž
.
w
x
1974 3310.
17 D.W. Fahey, S.R. Kawa, E.L. Woodbridge, P. Tin, J.C.
Wilson, H.H. Jonsson, J.E. Dye, D. Baumgardner, S. Bor-
rman, D.W. Toohey, L.M. Avallone, M.H. Proffitt, J. Margi-
tan, M. Loewenstein, J.R. Podolske, R.J. Salawitch, S.C.
Wofsky, M.W.K. Ko, D.E. Anderson, M.R. Schoeber, K.R.
w x
Ž
.
6
C.J. Howard, K.M. Evenson, J. Chem. Phys. 61 1974 1943.
w x
7
K. Erler, D. Field, R. Zellner, I.W.M. Smith, Ber. Bunsen-
Ž
.
ges. Phys. Chem. 81 1977 22.
w x
8
J.P. Burrows, T.J. Wallington, R.P. Wayne, J. Chem. Soc.,
Ž
.
Ž . Ž .
Chan, Nature London 363 1993 509.
Faraday Trans. 2. 79 1983 111.
w x
9
w
x
18 T. Gierzack, J.B. Burkholder, A.R. Ravishankara, J. Phys.
R.M. Forster, H.F. Frost, H. Hamann, H. Hippler, Schelpe-
Ž
.
Ž .
Chem. 1998, submitted .
greli, J. Troe, J. Chem. Phys. 103 1996 2949.
w
w
x
w
x
10 N.M. Donahue, M.K. Dubey, R. Mohrschladt, K.L. Demer-
19 S.S. Brown, T. Gierczak, R. Portman, R.K. Talukdar, J.B.
Ž
.
Ž
Burkholder, A.R. Ravishankara, Geophys. Res. Lett. 1998,
jian, J.G. Anderson, J. Geophys. Res. 102 1997 6159.
x
.
11 R.-S. Gao, D.W. Fahey, L.A. DelNegro, S.G. Donnelly, E.R.
Keim, L. Teverovski, P.O. Wennberg, T.F. Hanisco, M.H.
Proffitt, J.C. Wilson, J.W. Elkins, R.M. Stimpfle, R.C. Co-
hen, C.T. McElroy, T.P. Bui, R.J. Salawitch, S.S. Brown,
submitted .
w
x
20 T.J. Dransfield, M.M. Sprengnether, K.K. Perkins, N.M.
Ž
Donahue, J.G. Anderson, Geophys. Res. Lett. 1998, submit-
.
ted .