Constraining the mechanism of OH + NO2 using isotopically labeled reactants: Experimental evidence for HOONO formation

Article abstract of DOI:10.1021/jp0035582

The reaction of OH with NO₂ is central to atmospheric chemistry, and its dynamics can be constrained by studying the kinetics of isotopically labeled ¹⁸OH with NO₂. Experimental data on the scrambling rate constant for oxy

Full text of DOI:10.1021/jp0035582

J. Phys. Chem. A 2001, 105, 1515-1520
1515
Constraining the Mechanism of OH + NO2 Using Isotopically Labeled Reactants:
Experimental Evidence for HOONO Formation†
Neil M. Donahue,*,‡ Ralf Mohrschladt,§ Timothy J. Dransfield, and James G. Anderson
Department of Chemistry and Chemical Biology, HarVard UniVersity, Cambridge, Massachusetts 02138
Manvendra K. Dubey|
Atmospheric and Climate Sciences, Los Alamos National Labortaory, Los Alamos, New Mexico 87545
ReceiVed: September 29, 2000; In Final Form: NoVember 10, 2000
The reaction of OH with NO2 is central to atmospheric chemistry, and its dynamics can be constrained by
studying the kinetics of isotopically labeled 18OH with NO2. This labeling opens an isotopic scrambling pathway
in the reaction coordinate for nitric acid formation, providing experimental constraints on the high-pressure
behavior of the reaction with data obtained at low pressures. This reaction, however, is complicated by the
presence of a second product isomer, peroxynitrous acid (HOONO), which does not have a scrambling pathway.
We present data for the reaction of 18OH with NO2 at room temperature between 4 and 200 Torr. The reaction
is rapid and independent of pressure. We also locate the H-atom isomerization transition state and show that
the isomerization rate constant is at least an order of magnitude faster than adduct dissociation. These results
allow us to accurately constrain the formation rate constant of HONO2, which is a factor of 5 slower than the
observed OH removal rate constant at high pressure. We conclude that the difference is due to HOONO
formation. Our conclusion is consistent with recent theoretical predictions of this branching, and also provides
the only self-consistent reconciliation of the high-pressure data with the remainder of the experimental data
set.
Introduction
It has been proposed that many of the vexing features of this
system can be explained by the presence of a moderately stable
Not only does the reaction of NO2 with OH play a central
role in atmospheric chemistry,1,2 it also has highly complicated
and confounding reaction dynamics, comprising multiple vi-
brationally excited intermediates prone to collisional stabili-
zation.3-5 Consequently, evaluation of the pressure dependent
rate constant for this reaction has proven to be difficult, despite
a data set nearly without parallel in either scope or quality.6,7
The high pressure limit has been especially troublesome, both
because the small size of the reactive phase space forces this
limit to exceedingly high pressures8,9 and because it has proven
difficult to reconcile the existing high-pressure data with data
obtained at lower pressures.10,11
These mechanistic complexities have tangible effects. They
are the root cause of substantial disagreements between pub-
lished extrapolation functions for this reaction, which in turn
profoundly influence model results throughout the troposphere
and stratosphere. The lower stratosphere is a case in point. The
uncertainty in this rate constant affects assessments of impact
of aircraft effluents on ozone,12,13 and, conversely, recent in situ
field measurements of the coupled HOx-NOx system have
pointed to inadequacies in the recommendation for this rate
constant.14 Here we shall address these issues by exploiting
isotopes to separate the elementary steps comprising the complex
reaction mechanism.
reaction product, peroxynitrous acid (HOONO).15,16 This is an
extension of an earlier suggestion that excited electronic states
of HONO2 could be responsible for the unusually broad falloff
curve of this rate constant.17 The anomalously low isotopic
scrambling rate of 18OH + NO2 also led to speculation that
HOONO could play an important role in the kinetics of the
OH + NO2 reaction,18 though this appeared to run counter to a
lack of direct experimental evidence for HOONO formation in
the gas phase.19
Because HOONO should be much less stable than the major
product, nitric acid (HONO2), it should have a substantially
smaller density of states near the reactant energy than will nitric
acid, a correspondingly shorter excited state lifetime, and a lower
third-order rate constant. However, it is entropically favored,
both because attack on either oxygen atom of NO2 will lead to
its formation and because there are several stable conformers
corresponding to several allowed approach geometries, while
HONO2 is confined to a single, planar geometry. Finally, there
is nothing about the frontier orbital structure of NO2 to suggest
that either reaction path (leading to HOONO or HONO2) should
have a significant barrier. Thus, the high-pressure limit for
HOONO formation could well be much higher than that for
nitric acid formation, though given its relative instability
HOONO could have a very short thermal lifetime.
† Part of the special issue “Harold Johnston Festschrift”.
* Corresponding author. E-mail nmd+@cmu.edu.
‡ Present Address: Departments of Chemistry and Chemical Engineering,
Carnegie Mellon University, Pittsburgh, PA, 15217.
§ Present Address: BASF Aktiengesellschaft, D-67056 Ludwigshafen,
Germany.
The major shortcoming of these suggestions remains the lack
of any experimental evidence for HOONO formation in the gas
phase. Two attempts have been made to isolate this isomer, both
of which were carried out between room temperature and
approximately 220 K.5,19 The second experiment was carried
| E-mail: dubey@lanl.gov.
10.1021/jp0035582 CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/15/2000

1516 J. Phys. Chem. A, Vol. 105, No. 9, 2001
Donahue et al.
TABLE 1: Rate Constants for 18OH + NO2 vs Pressure
out in our lab and is described in a companion paper. Neither
experiment revealed any evidence for HOONO formation at any
temperature. The possibility remains that HOONO is formed
but has a very short thermal lifetime, or that the branching ratio
is relatively small (and distributed among several conformers).
However, our recent experiment, in particular, places severe
constraints on these possibilities; we observe HONO2 production
with a 90% yield (within error of unity) with a 20 ms
experimental time scale.5
Given the complexity and importance of this system, it is
essential that we obtain additional experimental constraints on
its behavior. Isotopic scrambling is one avenue. When OH adds
to the nitrogen in NO2, forming vibrationally excited HONO2,
the hydrogen atom can be transferred to either of the other two
oxygen atoms. If this transfer is facile, the vibrationally excited
intermediate can be considered as an [H‚NO3] complex. If
the complex lifetime is sufficiently long, when it decomposes
back to reactants (which it will do at low pressure, lacking a
third body to remove the vibrational energy) the departing
oxygen atom will be randomly selected from among the three
candidates. Ordinarily, this is a null process, as the atoms are
indistinguishable. However, if the oxygen atom on the attacking
OH is isotopically labeled, the exchange reaction will be
observable. In the perfect case so far described, the disappear-
ance rate constant of the labeled radical at low pressure will be
two-thirds of the high-pressure limit for the reaction with or
without the label (assuming that mass-dependent effects associ-
ated with the labeling are negligible).
This isotopic scrambling experiment has twice been applied
to the OH + NO2 reaction in low-pressure discharge-flow
measurements of 18OH removal kinetics.20,18 The results are in
good agreement, but like virtually all aspects of this reaction
they are not easy to reconcile with other characteristics of the
system. In particular, the resulting rate constants are too low
for the simple explanation provided here to hold. Greenblatt
and Howard18 offered HOONO formation as a possible explana-
tion for the discrepancy between their low pressure (2 Torr)
scrambling data and the high-pressure data, but without any
quantitative analysis of the aggregate kinetics and pressure
falloffs. Some time ago (in 1994), we measured the rate constant
for the 18OH + NO2 reaction as a function of nitrogen pressure
from 4 to 200 Torr at 298 K. Here we shall present these data,
together with an interpretation that is consistent with all known
facts about the reaction. The general notion that both isomers
of HNO3 contribute to the high-pressure kinetics data is strongly
supported by our observations.
P Torr
k (×1012) cm3/sec
4.49
4.49
9.53
9.58
21.96
40.41
73.11
73.04
134.40
211.90
13.5 ( 1.5
8.8 ( 1.2
11.1 ( 1.0
11.5 ( 1.1
10.2 ( 0.9
10.9 ( 1.1
9.3 ( 1.0
9.3 ( 1.0
9.5 ( 0.9
8.8 ( 0.7
Figure 1. Theoretical potential energy surface for the title reaction.
The isotopically labeled reactants are shown in the middle. The 18O is
indicated with an italicized O. Formation of HONO2 proceeds to the
right, first through H18ONO2, then to HONO18O, and finally to OH +
NO18O. Formation of HOONO proceeds to the left, where each of
several conformers can be formed directly but no isotopic scrambling
is facile. Note that the critical barrier to H-atom transfer (TS1) sits
well below the reactants (by 18 kcal/mol).
state using density functional theory (B3LYP/6-31G**). An
extensive portion of the reaction coordinate for the total system
is shown in Figure 1. These calculations are discussed in greater
detail in a related paper,5 but we shall summarize the germane
features here, in particular focusing on isotopic scrambling. We
shall then compute some rate constants based on this surface
to demonstrate that isotopic scrambling is much more rapid than
adduct dissociation and should thus be complete at low pressure.
This surface agrees well with another recently published
surface,4 though ours is the first to describe the isomerization
transition state under consideration here. Figure 1 is best read
out from the middle, where one finds the reactants. Formation
of nitric acid is shown to the right of the reactants. The barrier
to isomerization lies 18 kcal/mol below the reactant energy.
Isotopically scrambled reactants, which are a product from the
perspective of radical decay, are shown on the right-hand edge
of the figure. Formation of HOONO is also possible. This will
not have any facile isotopic scrambling pathway and will thus
behave nearly identically for 18OH and 16OH. This process is
shown to the left of the reactants (note that any conformer of
HOONO may be formed directly from OH and NO2, and only
the two most stable are shown).
Experimental Results
We measured rate constants for 18OH disappearance in the
presence of excess NO2 in a high-pressure flow system (HPFS).
All details of this experiment have been described extensively
in the literature,21 including application to the unlabeled OH +
NO2 reaction10 and to isotopic scrambling in the 18OH + H2O
reaction.22 The measurements described here were carried out
between 4.5 and 212 Torr of nitrogen at 299 ( 2 K. They are
summarized in Table 1.
Our primary objective is to assess whether isotopic scrambling
is indeed fast compared to dissociation of vibrationally excited
nitric acid. To achieve this, we need to know the unimolecular
isomerization and dissociation rates at roughly the reactant
energy. We shall present a simple calculation, using an average
dissociation rate based on the observed rate constant and a
microcanonical isomerization rate constant at approximately the
reactant energy. This is easy to understand and amply demon-
strates the facility of isomerization.
As a test of our overall accuracy, we also measured the rate
for the reaction 18OH + ethane f products over the same
pressure range. We found a pressure independent rate constant
of (2.5 ( 0.15) × 10-13 cm3 molec-1 sec-1, which agrees well
with the accepted rate for 16OH + ethane at room temperature.
Theoretical
To better constrain the H-atom transfer associated with
isotopic scrambling, we have located the pertinent transition

OH + NO2 Mechanism Constraint
J. Phys. Chem. A, Vol. 105, No. 9, 2001 1517
TABLE 2: B3LYP Frequencies for Species (in cm-1
(energies in kcal/mole)
)
mode
OH + NO2
HONO2
TS1
ν1
ν2
ν3
ν4
ν5
ν6
ν7
ν8
ν9
Ia
Ib
Ic
E
Ezp
3705
1709
1396
3717
1766
1354
1322
901
766
649
2333
1684
1349
1098
1063
840
724
614
749
581
479
1918i
Figure 2. RRKM isomerization rate vs HONO2 energy. Zero energy
is the zero-point energy of HONO2. The dissociation threshold is shown
as a horizontal dashed line at 43.1 kcal/mol, while the experimentally
constrained mean dissociation rate constant is shown as a vertical dashed
line at 8 × 108 sec-1. Clearly, isomerization is at least an order of
magnitude faster than dissociation.
0.207
0.401
0.429
-48.8
-32.3
0.213
0.376
0.493
0
10.8
-17.6
-3.8
First we shall calculate the average dissociation rate constant.
In the Lindemann-Hinshelwood model, the excited state
lifetime may be easily inferred from the observed high and low
pressure rate constants. The low pressure rate constant is
approximately
k∞
k0 ) âkII
(1)
kd
where â is the collisional efficiency of the bath gas (ap-
proximately 0.4 for nitrogen), kII is the second-order collision
rate constant (3 × 10-10 cm3 sec-1), and kd is the unimolecular
dissociation rate of the adduct. Solving for kd we find
k∞
kd ) âkII
(2)
k0
Using the JPL recommendation for the high and low-pressure
limits (the reason for this choice will become clear), we find kd
∼ 8 × 108 sec-1
.
We need to compare this dissociation rate constant with the
unimolecular isomerization rate constant to see which process
will dominate. For this we shall simply calculate the micro-
canonical rate constant near the reactant zero point energy. We
summarize the energies and vibrational frequencies for the
reactants, nitric acid, and the transition state, TS1, in Table 2.
Note that the transition state energy is -14.6 kcal/mol relative
to the reactant zero point, while the nitric acid ground-state lies
at -43.1 kcal/mol. Calculation of the microcanonical rate
constant for isomerization as a function of energy is straight-
forward. We calculate a simple sum over density of states,
neglecting anharmonicity and assuming that external rotations
are active:23
Figure 3. Experimentally determined rate constants for the reaction
18OH + NO2 f products plotted vs pressure. Literature values are also
shown, along with several functions. The straight line is a simple
average of all the data, while the sigmoid curve is a combination of
this average at low pressure and the lowest of the three falloff curves.
The falloff curves are recommended curves for the unlabeled OH +
NO2 rate constant. The lower curve is the current JPL recommendation6
with a slightly reduced high-pressure limit, the middle curve is from
our recent study,11 and the upper curve is the current IUPAC
recommendation.7 Residuals are shown relative to the simple average.
σW*(E - E0)
k(E) )
(3)
hF(E)
We count rovibrational states explicity.24 The result is shown
in Figure 2. Actual loss of the excited H18ONO2 will be twice
as rapid because of the two identical pathways (not treated in
the RRKM calculation), and the actual eigenvalue for equilibra-
and several functional curves. Our measurements are in nearly
perfect agreement with the literature values, but they extend
the pressure range to 200 Torr. Over the full range from 1 to
200 Torr, there is nothing in the data to suggest any significant
pressure dependence in the rate constant. If anything there is a
slight decrease in the observed rate constant with pressure,
though this is within the (1-σ) uncertainties of the individual
measurements. The weighted average of all the available data
is k ) 1.0 ( 0.04 × 10-11 cm3 molec-1 sec-1. At first blush
x
tion of the isomers will be faster by 6. Isomerization of the
vibrationally excited adduct is clearly at least an order of
magnitude faster than dissociation, and we may safely assume
that isotopic scrambling is rapid.
Discussion
Our data for the reaction 18OH + NO2 f products are plotted
in Figure 3, along with the other data from the literature20,18

1518 J. Phys. Chem. A, Vol. 105, No. 9, 2001
Donahue et al.
this is surprising, as the unlabeled reaction is known to be
strongly pressure dependent. Several recommendations for the
pressure dependence of that rate constant are shown as well in
Figure 3. Clearly, the rate constant for the isotopically labeled
reaction is much faster than that for the unlabeled reaction at
low pressures, but the two rate constants converge with
increasing pressure.
The observed behavior can be understood easily in the context
of the simple model presented in the Introduction; at low
pressures, isotopic scrambling causes adduct decomposition to
remove the reactant 18OH radical 2/3 of the time, but as pressure
increases a certain fraction of the vibrationally excited inter-
mediate is also collisionally stabilized, until eventually all of
the intermediates are stabilized and the labeled reaction behaves
identically to the unlabeled one (from the perspective of radical
removal). If the unlabeled reaction has a rate constant k(p), we
expect the labeled reaction to have a rate constant k′(p) )
2/3k(∞) + 1/3k(p). In this case, we use the average value
obtained above to predict a high-pressure limit of 1.5 × 10-11
cm3 molec-1 sec-1, which we use along with the JPL recom-
mendation for the falloff function, k(p). The resulting sigmoid-
shaped function (labeled “best guess”) is shown in Figure 3,
and while the data do not show the gentle increase with pressure
suggested by the function, the scatter is such that this is not
surprising. The rate constant should have increased by only 15%
at 200 Torr from the low pressure limit. The only modification
to the JPL recommendation required in this exercise is a slight
reduction in the high-pressure limit, from 1.67 to 1.5 × 10-11
cm3 molec-1 sec-1
.
This would settle the matter if the JPL recommendation
described the data at high pressures, but it does not. The other
two functions shown in Figure 3 are more representative of the
high pressure data. The middle curve, from our earlier work11
is based on all data except the very high pressure data of Forster8
and Fulle,9 while the upper curve is the IUPAC recommenda-
tion,7 which reasonably fits those high-pressure data but
overestimates the observed rate constants at intermediate
pressures. One possibility is that HOONO formation is respon-
sible for this evident increase in the high-pressure rate constant.
Figure 4. Falloff curves for the unlabeled OH + NO2 reaction.
Literature data are plotted in gray (see Dransfield).11 The falloff curves
show are for HONO2 formation (cyan), HOONO formation (magenta),
and the aggregate rate constant (green). These curves are directly
products of the data. The HONO2 formation rate constant is determined
from the isotopically labeled rate constants, while the HOONO
formation rate is left as a difference so that the aggregage rate matches
the available high pressure data. These curves are very similar to those
recently proposed by Golden and Smith.13 However, the constraints
are independent; our branching is constrained by the 18OH kinetics,
while theirs is constrained by a hindered-Gorin RRKM calculation.
If we accept this postulate, we can assume that the difference
between the high-pressure limit inferred from our scrambling
data (1.5 × 10-11 cm3 molec-1 sec-1) and the high pressure
limit of the data is due to HOONO formation. The high pressure
limit for OH removal has been measured in two independent
ways: by extension of kinetics measurements to extremely high
pressures8,9 (k∞ ) 7.5 × 10-11 cm3 molec-1 sec-1), and by
measurements of the quenching rate constant for vibrationally
excited OH, which should preferentially lose its quantum of
vibration to NO2 during a collision25 (k∞ ) 4.8 ( 0.8 × 10-11
cm3 molec-1 sec-1). These results are in reasonable agreement;
certainly there is consensus that the high-pressure limit is very
fast. Because the high-pressure kinetics data are more direct
and because they extend well into the falloff region and can
thus be compared to a falloff function over some range, we
shall use those data to constrain the high-pressure value, giving
a high-pressure limit for HOONO formation of 6 × 10-11 cm3
shown in Figure 4, along with the available data. Almost all of
the measured rate constants fall between the curve for HONO2
formation and the curve for the total rate constant, which is
consistent with the suggestion that experimental time scales are
influencing the kinetics observations.
Our measurements and analysis strongly support the results
of both recent theoretical predictions,15,16 adding the additional
constraint of isotopic labeling to the problem. They also confirm
the suggestion by Greenblatt and Howard18 that the unusually
low rate constant for the scrambling reaction is due to the
HOONO formation channel. We stress that our analysis is based
entirely on experimental results; we have used our calculated
isomerization rate constant to help build the case that isomer-
ization is facile, but it in no other way enters the calculation.
The critical data are the pressure-dependent rate constants for
the isotopic scrambling reaction. Because the scrambling rate
constant is blind to HOONO formation, we can infer the high-
pressure limit for HONO2 formation from these results, provided
that isomerization indeed far outpaces decomposition for the
vibrationally excited intermediate. With a second assumption
that HONO2 formation dominates at low pressure, we can
completely describe the HONO2 formation rate constant and
then, by subtraction, the HOONO formation rate constant.
molec-1 sec-1
.
This leaves the low pressure limit of this channel as the sole
remaining parameter to constrain. The data are reasonably
reproduced with a low-pressure rate constant for HOONO
formation of 4 × 10-31 cm6 molec-1 sec-1. This may be
compared with our low-pressure rate constant for HONO2
formation of 2.4 × 10-30 cm6 molec-1 sec-1. The separate
formation rate constants and the sum of both channels are all

OH + NO2 Mechanism Constraint
J. Phys. Chem. A, Vol. 105, No. 9, 2001 1519
Our results are quantitatively closer to those of Golden and
Smith16 than those of Matheu et al.15 Indeed, the consistency
of the curves plotted in Figure 4 is striking; they are within
20% of each other over the entire pressure range. While the
master equation simulations of Matheu et al.15 do not produce
simple Lindemann-Hinshelwood type rate expressions, in
general terms their results suggest a much shorter lifetime for
HOONO and a correspondingly smaller third-order rate constant
for this channel. Thus we conclude that our results substantially
confirm those of Golden and Smith,16 with the only interde-
pendence being that each result is in part constrained by the
data for the unlabeled reaction.
One problem remains. The hindered-Gorin RRKM calcula-
tions of Golden and Smith16 show a long lifetime for HOONO
at 230 K. In a companion paper we report a product study of
the unlabeled reaction revealing no sign of HOONO at 230 K
and 375 torr.5 It is difficult to reconcile this with the evidence
presented here, which suggests that the yield of HOONO should
be up to 30%, unless the HOONO lifetime is less than the
experimental time scale of 0.2 s even at 230 K. We should note
that any HOONO produced would be dispersed among several
conformers, so individual IR absorption bands may be quite
weak. It is also possible that we have overestimated the HOONO
formation rate; using vibrational deactivation to constrain the
high pressure limit for this channel reduces the rate by nearly
40%. A combination of these effects could render the HOONO
itself quite difficult to observe and the difference of the HONO2
branching ratio from unity difficult to discern.
A second possibility could explain the isotopic scrambling
data: if H-atom transfer is not facile but is instead in competition
with adduct decomposition, the observed low-pressure limit of
the title reaction would indeed be lower than the expected 2/3
of the high-pressure limit. While a model based on incomplete
scrambling can explain the 18OH kinetics, the general consis-
tency of our current explanation with the other available facts
strongly favors the simpler explanation we have already
presented.
Several additional experimental constraints can and should
be brought to bear on this problem. First, temperature-dependent
data for the isotopic scrambling would allow us to extend this
analysis away from room temperature and down to temperatures
relevant to the upper troposphere and lower stratosphere.
Second, measurements with 18OD would quickly reveal whether
H-atom transfer is truly facile in this system. If it is, the primary
isotope effect should be only modest. Third, measurement of
the formation rate constant for 16OH would directly constrain
the scrambling rate constant. Fourth, a product study of the
reaction should reveal a statistical (2:1) distribution of nitric
acid isomers at low pressure that starts to become biased strongly
toward H18ONO2 at approximately 200 Torr. Fifth, kinetics
measurements of the 16OH + NO2 reaction at the appropriate
time scale should reveal biexponential behavior in the OH
decays if these conjectures are correct. Except at very high
pressure, the effect would be subtle because the two eigenvalues
would be the sum of the rate constants and the HONO2
formation rate constant alone, which do not differ by a factor
of 2 until 2000 Torr. Finally, direct observations of the various
HOONO conformers, and in particular their thermal lifetimes,
would close the final loophole in these calculations.
OH, NO2, and HONO2 is in a rough diurnal steady state. The
process described here should work to enrich NO2 in oxygen
isotopes at the expense of OH (and HONO2). Consequently,
18OH should be depleted in the stratosphere, and 18ONO
enriched, unless other feedbacks diminish the effect. Because
of the many radical feedback loops in the stratosphere, the
ultimate fractionation can only be predicted by a complete
photochemical model.
Our results suggest that, while HOONO is probably formed
in significant quantities in the atmosphere, it is probably lost
just as quickly through thermal decomposition. These results
strongly support the conclusion of Golden and Smith16 that the
rate constants appropriate to atmospheric chemistry modeling
(i.e., for HONO2 formation) are those generally following the
lower bound of the available rate constant data (the current JPL
recommendation). In fact, the HONO2 formation rate constant
presented here is 5% lower than the current JPL recommendation
at room temperature and pressure. The chain termination rate
due to OH + NO2 is thus probably overestimated in current
photochemical models; this will significantly affect ozone
assessments and the interpretation of field data.
Conclusions
We have presented a consistent explanation for the unusual
kinetics of the OH + NO2 reaction. Our experimental data on
the scrambling rate constant for oxygen isotopes show that the
pressure dependence (none) for this process is quantitatively
consistent with facile H-atom transfer in the vibrationally excited
complex. This in turn confirms that the high-pressure limit for
HONO2 formation is well below the observed high pressure
limit for OH disappearance, leaving HOONO formation as the
only likely explanation for those observations. These results
support recent predictions that arrived at the same general
conclusion15,16 using master equation and RRKM computations
on the theoretical surface constrained by the kinetics of the
unlabeled reaction. Continued failure to observed HOONO
under any experimental conditions5 remains a perplexing
problem, but this result may be due to a difference between the
thermal lifetime of HOONO and the experimental time scale
of the product studies.
Acknowledgment. This work was supported by grants
ATM-9414843 and ATM-9977992 from NSF. Laboratory work
by R.M. and M.K.D. provided the grist for this study. The
authors also thank Jesse Kroll for assistance with the RRKM
calculations. M.K.D. thanks LANL-NMRPI for funds to con-
tinue his work on atmospheric ozone photochemistry.
References and Notes
(1) Johnston, H. Science 1971, 173, 517.
(2) Logan, J. A.; Prather, M. J.; Wofsy, S. C.; McElroy, M. B. J.
Geophys. Res. 1981, 86, 7210.
(3) McGrath, J. P.; Rowland, F. S. J. Phys. Chem. 1994, 98, 1061.
(4) Chakraborty, D.; Park, J.; Lin, M. C. J. Chem. Phys. 1998, 231,
39.
(5) Dransfield, T. J.; Donahue, N. M.; Anderson, J. G. J. Phys. Chem.
A 2001, 105, 1507.
(6) 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 stratopsheric modeling. Technical Report 00-3, Jet Propulsion
Laboratory, 2000.
(7) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr,
J. A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1997, 26, 1329.
(8) Forster, R.; Frost, M.; Fulle, D.; Hamann, H. F.; Hippler, H.;
Schlepegrell, A.; Troe, J. J. Chem. Phys. 1995, 103, 2949.
(9) Fulle, D.; Hamann, H. F.; Hippler, H.; Troe, J. J. Chem. Phys. 1998,
108, 5391.
Atmospheric Implications
The title reaction may play an important role in the
stratosphere by causing strong mass-independent fractionation
of OH. Under many conditions, the NOx-NOy system including

1520 J. Phys. Chem. A, Vol. 105, No. 9, 2001
Donahue et al.
(10) Donahue, N. M.; Dubey, M. K.; Mohrschladt, R.; Demerjian, K.
L.; Anderson, J. G. J. Geophys. Res. 1997, 102, 6159.
(11) Dransfield, T. J.; Perkins, K. K.; Donahue, N. M.; Anderson, J.
G.; Sprengenther, M. M.; Demerjian, K. L. Geophys. Res. Lett. 1999, 26,
687.
(12) Stolarski, R. S. et al. Scientific assessment of the atmospheric effects
of stratospheric aircraft. Technical Report Refn. Pub. 1381, NASA, 1995.
(13) Dubey, M. K.; Smith, G. P.; Hartley, W. S.; Kinnison, D. E.;
Connell, P. S. Geophys. Res. Lett. 1995, 24, 2737.
(14) Perkins, K. K.; et al. J. Phys. Chem. A 2001, 105, 1521.
(15) Matheu, D. M.; Green, W. H. Int. J. Chem. Kinet. 2000, 32, 245.
(16) Golden, D. M.; Smith, G. P. J. Phys. Chem. 2000, 104, 3991.
(17) Howard, M. J.; Smith, I. W. M. In Jennings, K. R.; Cundall, R. B.;
and Margerum, D. W., Eds.; Progress in Reaction Kinetics; 1983; Vol. 12,
p 55.
(18) Greenblatt, G. D.; Howard, C. J. J. Phys. Chem. 1989, 93, 1035.
(19) Burkholder, J. B.; Hammer, P. D.; Howard, C. J. J. Phys. Chem.
1987, 91, 2136.
(20) Dransfeld, P.; Lukacspaal, J.; Wagner, H. G. Z. Natur. 1986, 41,
1283.
(21) Donahue, N. M.; Clarke, J. S.; Demerjian, K. L.; Anderson, J. G.
J. Phys. Chem. 1996, 100, 5821.
(22) Dubey, M. K.; Mohrschladt, R.; Donahue, N. M.; Anderson, J. G.
J. Phys. Chem. 1997, 101, 1494.
(23) Kroll, J. H.; Shahai, S.; Anderson, J. G.; Demerjian, K. L.; Donahue,
N. M. J. Phys. Chem., submitted.
(24) Stein, S. E.; Rabinovitch, B. S. J. Chem. Phys. 1973, 58, 2438.
(25) Smith, I. W. M.; Williams, M. D. J. Chem. Soc. Faraday Trans.
1985, 81, 1849.