Kinetics of the Removal of OH(v = 1) and OD(v = 1) by HNO3 and DNO3 from 253 to 383 K

Article abstract of DOI:10.1021/jp0346413

The reaction of OH with HNO₃ involves both HO_X (HO_X ≡ OH + HO₂) and NO_X (NO_X ≡ NO + NO₂ + NO₃) species and plays two important roles in the atmosphere. The rate coefficients for the removal of OH(ν = 1) and OD(ν = 1) by HNO₃ and DNO₃ as a function of temperature from 253 to 383 K. The measured rate coefficients at 295 K ranged from 2.5 × 10^-11 cc/molecule-sec for the removal of OH(ν = 1) by HNO₃ to 6 × 10^-12 cc/molecule-sec for the removal of OH(ν = 1) by DNO₃; the rate coefficients for the like-isotope processes [removal of OH(ν = 1) by HNO₃ and removal of OD(ν = 1) by DNO₃] were 2-4 times higher than the rate coefficients for the unlike-isotope processes. All four rate coefficients showed negative temperature dependences that were too strong to be attributable only to long-range interactions between the reactions. The higher rate coefficients of the like-isotope processes (relative to the unlike-isotope processes) was attributed to faster, resonant, intramolecular vibrational energy redistribution within the reaction complexes containing the same isotopes. The yields of OH(ν = 1) were estimated to be ～ 1% and OH(ν = 2) to be ～ 0.4% of that of OH(ν = 0) from the photolysis of HNO₃ at 248 nm.

Full text of DOI:10.1021/jp0346413

7762
J. Phys. Chem. A 2003, 107, 7762-7769
Kinetics of the Removal of OH(W ) 1) and OD(W ) 1) by HNO3 and DNO3 from
2
53 to 383 K
David C. McCabe,† Steven S. Brown,†,§ Mary K. Gilles,
,‡,§
†,§,|
Ranajit K. Talukdar,†,§
⊥
,†,‡,§
Ian W. M. Smith, and A. R. Ravishankara*
NOAA Aeronomy Laboratory, 325 Broadway R/AL2, Boulder, Colorado 80305, Department of Chemistry and
Biochemistry, UniVersity of Colorado, Boulder, Colorado 80309, CooperatiVe Institute for Research in
EnVironmental Science, UniVersity of Colorado, Boulder, Colorado 80309, and School of Chemical Sciences,
UniVersity of Birmingham, Edgbaston, Birmingham BI5 2TT, United Kingdom
ReceiVed: March 12, 2003; In Final Form: July 10, 2003
We have measured the rate coefficients for the removal of OH(V ) 1) and OD(V ) 1) by HNO
as a function of temperature from 253 to 383 K. OH(V ) 1) and OD(V ) 1) were produced by photolysis of
HNO /DNO at 248 nm; laser-induced fluorescence was used to monitor the kinetics of the vibrationally
excited radicals. The measured rate coefficients at 295 K range from 2.5 × 10 cm molecule
removal of OH(V ) 1) by HNO for the removal of OH(V ) 1) by DNO
to 6 × 10 cm molecule
the rate coefficients for the like-isotope processes [removal of OH(V ) 1) by HNO and removal of OD(V )
) by DNO ] are 2-4 times higher than the rate coefficients for the unlike-isotope processes. All four rate
3 3
and DNO
3
3
-
11
3
-1 -1
s
for the
-
12
3
-1 -1
3
s
3
;
3
1
3
coefficients show negative temperature dependences that are too strong to be attributable only to long-range
interactions between the reactants. Expressed as negative activation energies, the temperature dependences
yield values of E
occurs via formation of the hydrogen-bonded, cyclic OH‚HNO
a
/R from -520 to -750 K. We suggest that the removal of the vibrationally excited radicals
3
reaction complex (or the appropriate isotopomer
of the complex) invoked to explain the unusual kinetics of the reaction of ground-state OH with nitric acid.
We postulate that dissociation of the reaction complex to regenerate nitric acid and vibrationally excited OH
or OD competes with intramolecular vibrational redistribution of the OH/OD vibrational excitation energy
within the reaction complex, leading to the observed negative temperature dependence. We attribute the higher
rate coefficients of the like-isotope processes (relative to the unlike-isotope processes) to faster, resonant,
intramolecular vibrational energy redistribution within the reaction complexes containing the same isotopes.
Additionally, we estimate the yields of OH(V ) 1) to be ∼1% and OH(V ) 2) to be ∼0.4% of that of OH(V
)
3
0) from the photolysis of HNO at 248 nm.
Introduction
The kinetics of reaction 1 are unusual in several ways.
Previous work has shown that the rate coefficient for this
reaction, k1, is pressure dependent at temperatures below ∼325
K, increases with decreasing temperature below ∼300 K, and
is strongly affected by isotopic substitution of deuterium for
The reaction of OH with nitric acid, reaction 1
OH + HNO f H O + NO
(1)
3
2
3
1
,2
the hydrogen in the nitric acid.
involves both HOX (HOX ≡ OH + HO2) and NOX (NOX ≡ NO
NO2 + NO3) species and plays two important roles in the
Reaction 1 is thought to proceed via formation of a hydrogen-
+
bonded OH‚HNO3 complex containing a six-membered ring,
atmosphere. It is a sink for HOX, and it converts HNO3, a
relatively unreactive reservoir for odd-nitrogen species, into
NOX. Therefore, reaction 1 affects the concentration of HOX
and the fraction of nitrogen oxides in the form of NOX (i.e., the
ratio of NOX to NOY). The concentration of HOX and the fraction
of nitrogen oxides in the form of NOX strongly affect the ozone
concentration in the lower stratosphere and in the upper
troposphere.
3
shown in Figure 1. Previous work from this laboratory has
shown that the observed kinetics of reaction 1 are compatible
with a mechanism (also shown in Figure 1) involving this OH‚
2
HNO3 complex. Two groups have reported investigations of
4,5
the OH‚HNO3 complex with ab initio methods, showing that
it is stable with respect to unbound OH and HNO3 by ∼5-8
-1
kcal mol .
The species OH‚HNO3* shown in Figure 1 represents the
nascent reactive complex OH‚HNO3 that has not been colli-
sionally deactivated. It is difficult to determine rate coefficients
for the association of reactants to form nascent reactive
complexes, such as the rate coefficient ka for the formation of
the OH‚HNO3* complex from OH and HNO3. This is particu-
larly true if high-pressure data are unavailable. Smith and co-
workers have proposed that one can estimate this type of rate
coefficient (the high-pressure limit in the case of an association
*
To whom correspondence should be addressed. E-mail: ravi@
al.noaa.gov.
†
NOAA Aeronomy Laboratory.
Department of Chemistry and Biochemistry, University of Colorado.
CIRES, University of Colorado.
‡
§
|
Current address: Chemical Sciences Division, Lawrence Berkeley
National Laboratory, MS 6R2100, 1 Cyclotron Rd., Berkeley, California,
4720.
9
⊥
University of Birmingham.
1
0.1021/jp0346413 CCC: $25.00 © 2003 American Chemical Society
Published on Web 09/06/2003

Removal of OH and OD by HNO3 and DNO3
J. Phys. Chem. A, Vol. 107, No. 39, 2003 7763
OH(V ) 1) or OD(V ) 1) in a slowly flowing mixture of HNO3
and/or DNO3 in He. The apparatus used to measure these rate
coefficients has been used many times previously in our
laboratory to measure rate coefficients for the reactions of
ground-state OH and OD and is described in detail elsewhere.9
This description focuses on the modifications to the previously
described system that were required to study the kinetics of OH-
(V ) 1) and OD(V ) 1).
HNO3 and DNO3 were used as the photolytic precursors of
OH(V ) 1) and OD(V ) 1). A mixture of HNO3 and/or DNO3
in He (both isotopomers were required for measurements of k3
-
1
2
and k4) was flowed at a linear flow velocity of ∼10 cm s
Figure 1. Mechanism for reaction 1 invoked by Brown et al. to explain
the observed kinetics of reaction 1. The mechanism involves formation
through a jacketed, temperature-controlled glass cell. The total
pressure in the reaction cell was approximately 30 Torr,
consisting mostly of He. The 248-nm output of a KrF excimer
laser was then passed through the cell (up to 60 mJ per pulse,
of the OH‚HNO
3 2
complex, which subsequently reacts to form H O
and NO
3
.
reaction) by measuring the rate coefficient for the removal of
one reactant that is excited with one quanta of vibrational
excitation by the second reactant.6-8 This approach assumes that
the vibrational excitation is only removed by a mechanism
involving the formation of the reaction complex; that the
vibrational excitation does not affect the rate coefficient for
formation of the complex; and that, if the reaction complex
dissociates to re-form the reactants, the reactant that was
originally vibrationally excited will be re-formed essentially only
in the vibrational ground state.
2
beam size of ∼1.5 cm ) in a direction perpendicular to the gas
flow to photolyze a fraction (e0.1%) of the HNO3 (DNO3). A
small fraction of the OH (OD) produced by photolysis is
vibrationally excited.
This vibrationally excited OH(V ) 1) or OD(V ) 1) was
monitored via laser-induced fluorescence. The second harmonic
of a tunable dye laser pumped by the second harmonic of a
pulsed Nd:YAG laser (532 nm) was passed through the reaction
cell, perpendicular to both the gas flow and the photolysis laser.
2
+
This probe laser excited the Q1(1) line of the A Σ (V′ ) 0) r
To further investigate the mechanism of the reaction of OH
with HNO3, we have measured the rate coefficients for the
removal of OH(V ) 1) and OD(V ) 1) by HNO3 and DNO3 as
a function of temperature from 253 to 383 K
2
10
X Π(V′′ ) 1) band (λair ) 345.85 nm for OH, 334.18 nm for
1
1
2
+
2
OD ); fluorescence in the A Σ (V′ ) 0) f X Π(V′′ ) 0) band
at ∼308 nm was detected with a photomultiplier tube (PMT).
The fluorescence passed through a band-pass filter (peak
transmission at 307.5 nm, fwhm ) 10 nm) between the reaction
region (where the photolysis and probe laser beams intersect)
and the PMT; this band-pass filter rejected scattered light from
the lasers. Temporal profiles of OH(V ) 1) or OD(V ) 1) were
generated by varying the delay time between the photolysis and
probe lasers.
OH(V ) 1) + HNO f loss of OH(V ) 1)
(2)
(3)
3
OH(V ) 1) + DNO f loss of OH(V ) 1)
3
OD(V ) 1) + HNO f loss of OD(V ) 1)
(4)
(5)
3
The yield of OH(V ) 1) from the 248-nm photolysis of HNO3,
relative to the yield of OH(V ) 0), was estimated by comparing
the LIF signals (back-extrapolated to zero delay between the
photolysis and probe lasers) from exciting the Q1(1) lines of
OD(V ) 1) + DNO f loss of OD(V ) 1)
3
These rate coefficients should be approximately equal to ka.
Measuring the various isotopomeric combinations allows for
the investigation of the effect of vibrational resonance between
the reactants on these rate coefficients.
Use of rate coefficients for the loss of vibrationally excited
reactants as an estimate of the rate coefficient for formation of
a reactive complex has primarily been used to estimate k , the
rate coefficient for an association reaction between two radicals
in the limit of high pressure. In these cases, a covalent bond is
formed between the two reactants, with bond energies greater
2
+
2
the A Σ (V′ ) 0) r X Π(V′′ ) 1) band (λair ) 345.85 nm) and
2
+
2
10
the A Σ (V′ ) 0) r X Π(V′′ ) 0) band (λair ) 307.84 nm).
The signals were recorded in back-to-back measurements, with
the concentration of HNO3 held constant.
The concentration of HNO3 and/or DNO3 was several orders
of magnitude larger than the concentration of OH(V ) 1) or
OD(V ) 1); therefore, the removal rate of the vibrationally
excited radicals was pseudo-first-order in their concentration.
However, the temporal profiles of OH/OD(V ) 1) were not
single exponential because of the formation of OH/OD(V > 1)
in the 248-nm photolysis of HNO3/DNO3. [Photolysis of HNO3
by 248-nm radiation to produce OH and NO2 is exothermic by
∞
6
-
1
than roughly 40 kcal mol . When the OH‚HNO3 complex
forms, the bonds formed are hydrogen bonds, which are weaker
OH‚HNO3 is bound by 5-8 kcal mol-1 4,5). In this paper, we
(
also discuss whether the rate coefficients for reactions 2-5 are
equivalent to ka, given the differences between reaction 1 and
typical association reactions. Last, we report measurements of
the yields of OH(V ) 1) and (V ) 2), relative to the yield of
OH(V ) 0), from the photolysis of HNO3 at 248 nm.
-
1
∼
65 kcal mol , so production of OH in vibrational states up
to (V ) 7) is thermodynamically allowed.] The formation of
OH(V > 1) was confirmed by the detection of the LIF signal of
2
+
2
OH(V ) 2) at the Q1(1) line of the A Σ (V′) 1) r X Π(V′′ )
1
0
2
) band (λair ) 350.90 nm). This signal was weak, but its
temporal behavior appeared to be single-exponential, indicating
Experimental Section
that 248-nm photolysis of HNO produces little OH(V > 2) and,
3
if any OH(V > 2) is formed, it has a negligible influence on the
Rate coefficients k2-k5 were measured using pulsed laser
photolysis of HNO3 or DNO3 to produce OH(V ) 1) or OD(V
measured temporal profiles. Thus, for reaction 2, where OH(V
) 1) is monitored with HNO3 present, the temporal behavior
of OH(V ) 1) is governed by three reactions, assuming that
)
1) and pulsed laser-induced fluorescence (LIF) detection of

7764 J. Phys. Chem. A, Vol. 107, No. 39, 2003
McCabe et al.
temporal behavior of the OH(V ) 2) signal detected via
excitation at 350.90 nm. Back-to-back temporal profiles of OH-
(V ) 1) and OH(V ) 2) were recorded, with the concentration
of HNO3 held constant. The fitted rate coefficient for the rise
in the temporal profile of OH(V ) 1) had a large uncertainty,
but it was consistent with the first-order rate coefficient for the
loss of OH(V ) 2). The OH(V ) 2) signal at 350.90 nm was
also compared to the OH(V ) 1) signal at 345.85 nm to estimate
the relative photolysis yields of these two states. However, this
measurement is quite rough, because the two bands access
different vibrational states of the electronically excited OH-
2
+
(
A Σ ) species and the detection efficiency of our system is
not the same for these different states.
Accurate LIF measurements at short reaction times (<2.5 µs)
were not possible because of an interfering fluorescence detected
through the 308-nm band-pass filter. The fluorescence was
observed immediately after the photolysis pulse (without the
probe laser) when nitric acid (HNO3 or DNO3) was present in
the reaction cell. The strength of the fluorescence signal was a
nonlinear function of the photolysis laser power, and we attribute
2
+
the fluorescence to OH(A Σ ) produced by multiphoton pho-
tolysis of nitric acid. Fluorescence around 308 nm immediately
after 248-nm photolysis of HNO3 has been previously noted
Figure 2. Temporal profiles of OH(V ) 1) at two nitric acid
concentrations on linear (top panel) and logarithmic (bottom panel)
scales, showing the non-single-exponential behavior of the signal at
early times, due to reaction 6a. Fits to eq I are also shown.
1
2
by MacLeod et al., who also described a very similar
fluorescence from pernitric acid (PNA, HOONO2). In the case
2
+
vibrational relaxation by the He bath gas is negligible
of PNA, they attributed the fluorescence to OH(A Σ ) produced
2
+
by three-photon photolysis of PNA. The OH(A Σ ) fluorescence
from photolysis of HNO3 restricted the temporal profiles to
measurements made after 2.5 µs, which limited the precision
of the fitted value of k6′. [Large nitric acid concentrations,
needed to maintain adequate OH(V ) 1) signal, led to a 10-
(6a)
3
OH(V ) 2) + HNO f OH(V ) 1) + HNO3
f loss of OH(V ) 2) not
producing OH(V ) 1) (6b)
6
0% loss of OH/OD(V ) 2) within 2.5 µs.]
The decay rate coefficient k2′ was measured at various
OH(V ) 1) + HNO f loss of OH(V ) 1) (2)
3
concentrations of HNO3. A weighted linear least-squares fit of
k2′ vs [HNO3] yielded the bimolecular rate coefficient k2. For
reaction 3, OH(V ) 1) was monitored with both HNO3 and
DNO3 present. In this case, OH(V ) 2) was quenched to OH(V
Using the pseudo-first-order rate coefficients k6a′ ≡
k6a[HNO3], k6b′ ≡ k6b[HNO3], k6′ ≡ k6a′ + k6b′, and k2′ ≡
k2[HNO3], the concentration of OH(V ) 1) at time t is given by
)
1) by both isotopomers, and OH(V ) 1) was lost via reactions
[
OH(V ) 1)] ) [OH(V ) 1)] exp(-k ′t) +
t
0
2
with both isotopomers, reactions 2 and 3. The temporal profiles
were fit to eq I, with k2′ replaced with k′ ) k2[HNO3] +
k3[DNO3]. The concentration of HNO3 was held constant while
that of DNO3 was varied; the slope of the linear fit of k′ vs
[DNO ] yielded the bimolecular rate coefficient k .
k ′[OH(V ) 2)]
0
6
a
[
exp(-k ′t) - exp(-k ′t)] (I)
6
2
k ′ - k ′
2
6
where [OH(V ) 1)]0, etc., are the concentrations of those
reactants immediately after photolysis. The measured temporal
profiles were fit to this form, with the assumption that k6a′ )
k6′. We made this assumption because k6a′ and [OH(V ) 2)]0
are not independent variables in eq I; this assumption does not
affect the measured values of k2-k5. The temporal profiles were
fit well by eq I, indicating that, if OH(V > 2) is formed, its
concentration is not large enough to influence our measured
values of k2-k5. It is possible that reaction 6b is significant, so
that k6a′ < k6′. Therefore, the fitted values of [OH(V ) 2)]0 are
lower limits. Figure 2 shows representative temporal profiles
of OH(V ) 1) at two concentrations of HNO3, on linear and
logarithmic scales. The biexponential nature of these temporal
profiles is apparent in this figure.
It must be noted that, with biexponential temporal profiles
such as those shown in Figure 2, the rate coefficients k2′ and
k6′ cannot be independently assigned as the larger of the two
will always govern the rise of the signal and the smaller of the
two will always govern the decay of the signal, regardless of
whether the faster reaction produces or consumes the monitored
species. In this system, we assign the rise to k6′ because
vibrational quenching is expected to be more rapid for higher
vibrational levels. This assignment was confirmed by the
3
3
The temporal behavior of OD(V ) 1) showed the same
biexponential behavior, and the measured temporal profiles of
this species were analyzed using the same method as employed
for OH(V ) 1) to obtain k and k . Figure 3 shows typical plots
4
5
of pseudo-first-order rate coefficients k ′-k ′ obtained at ∼296
2
5
K versus the concentration of HNO or DNO .
3
3
The ratio of the fitted values of k ′ and k ′ was used to
6
2
estimate the rate coefficient k (≡ k + k ) relative to k ; similar
6
6a
6b
2
ratios of the first-order rate coefficients for the rise and loss of
OH/OD(V ) 1) were used to estimate the rate coefficients for
reactions 7-9
OH(V ) 2) + DNO f loss of OH(V ) 2)
(7)
(8)
(9)
3
OD(V ) 2) + HNO f loss of OD(V ) 2)
3
OD(V ) 2) + DNO f loss of OD(V ) 2)
3
As noted above, the precision of the fitted value for the rate
coefficient for the rise of OH/OD(V ) 1) was quite poor, so
only rough estimates of the rate coefficients k6-k9 were possible.

Removal of OH and OD by HNO3 and DNO3
J. Phys. Chem. A, Vol. 107, No. 39, 2003 7765
2
+
obtained by probing the Q1(1) lines of the A Σ (V′ ) 0) r
2
2
+
2
X Π(V′′ ) 1) (λair ) 345.85 nm) and A Σ (V′ ) 1) r X Π(V′′
2) (λair ) 350.90 nm) OH bands. The ratio of the LIF signals
was normalized by the transition probabilities for the two
)
1
3
excitation lines and the laser powers. To account for the
different vibrational states of OH(A Σ ) accessed by the two
2
+
excitation wavelengths, we corrected the signals by assuming
2
+
that fluorescence from OH(A Σ , V ) 0) occurs in the Q1(1)
2
+
2
10
line of the A Σ (V′ ) 0) f X Π(V′′ ) 0) (λair ) 307.84 nm)
2
+
band and that fluorescence from OH(A Σ , V ) 1) occurs in
2
+
2
the corresponding line of the A Σ (V′ ) 1) f X Π(V′′ ) 1)
1
0
(
λair ) 313.46 nm) band. Therefore, the LIF signals were
further normalized by the transition probabilities of these
transitions and by the transmission of the band-pass filter on
the PMT at these two wavelengths. The resulting quantum yield
of OH(V ) 2) is roughly 30% of that of OH(V ) 1), i.e., an
absolute quantum yield of roughly 0.003.
The yield of OH(V ) 2), relative to that of OH(V ) 1), was
estimated a second way by comparing the fitted values of the
initial concentrations, [OH(V ) 2)]0 and [OH(V ) 1)]0, from
the temporal profiles (eq I). Assuming that k6 ) k6a, comparison
of the fitted values indicates that the yield of OH(V ) 2) from
248-nm photolysis of nitric acid is approximately 40% of that
of OH(V ) 1), i.e., a quantum yield of 0.004. However, because
k6 might well be greater than k6a, this ratio of yields is actually
a lower limit. Given that estimates of the OH(V ) 2) yield from
the fits of the temporal profiles and the comparison of LIF
signals are both rather imprecise, the agreement is good. The
agreement in the relative yields of OH(V ) 2) estimated by these
two methods suggests that k6 is indeed approximately equal to
k6a, i.e., that OH(V ) 2) only loses one quantum of vibrational
energy to nitric acid per “reaction”.
Figure 3. Pseudo-first-order rate coefficients k
concentration of HNO or DNO
at ∼296 K. Circles, k
′; squares, k ′; triangles, k ′. The slopes of these plots are the second-
order rate coefficients k -k ) (5.0
2
′-k
5
′ versus the
3
3
2
′; diamonds,
k
3
4
5
-
11
2
5
: k
2
) (2.55 ( 0.12) × 10 , k
3
-
12
-12
-11
(
0.5) × 10 , k
4
) (7.8 ( 0.5) × 10 , k ) (1.94 ( 0.07) × 10
5
3
-1 -1
(
all in units of cm molecule
s
). The large intercepts for k
3
and k
4
are due to the removal of OH(V ) 1) or OD(V ) 1) by HNO
3
or DNO ,
3
respectively (i.e., due to the reaction of the vibrationally excited radical
with its photolytic precursor).
Anhydrous HNO3 and DNO3 were prepared by vacuum
distillation of the nitric acid formed by the addition of H2SO4
or D2SO4 to NaNO3. HNO3 and DNO3 concentrations were
measured by absorption, at room temperature, using the 213.9-
nm line from a Zn lamp and 100-cm absorption cells im-
mediately upstream of the LIF reactor [σ213.9 nm(HNO3) ) (4.52
The relative quantum yields of the vibrational states of OH
from 248-nm photolysis are somewhat surprising: Only about
1% of the OH is vibrationally excited, and the yield of OH(V
-
19
2 1
(
0.19) × 10
cm ]. The absorption cross section of DNO3
at this wavelength was obtained by measuring the absorption
as a function of the pressure of DNO3 and was found to be 4.5
) 2) is ∼30-40% of that of OH(V ) 1). This might be due to
-
19
2
15,17
×
10
cm . The nitric acid/He mixture flowed through only
efficient multiphoton photolysis of HNO ,
although the ratio
3
glass tubing and Teflon fittings before the reaction cell to prevent
the decomposition of nitric acid on surfaces such as stainless
steel.
[OH(V ) 1)] /[OH(V ) 2)] , as calculated from the fits to eq I,
0
0
did not vary with the photolysis laser power as the fluence was
2
varied from 2.7 to 21 mJ/cm . The ratio also did not vary
significantly when N2 (∼2 Torr), which efficiently quenches
Results and Discussion
1
1
O( D), was added to the reaction mixture, suggesting that O( D)
(
conceivably formed in the photolysis of HNO3) is not an
Yields of OH(W ) 1) and OH(W ) 2) from 248-nm
Photolysis of HNO3. The ratio of the LIF signals of OH(V )
important precursor of vibrationally excited OH. The directly
observed single-exponential temporal profile of OH(V ) 2)
suggests that, if OH(V > 2) is produced by HNO3 photolysis at
this wavelength, it is produced only in small amounts relative
to OH(V ) 2).
Rate Coefficients k2-k5. The rate coefficients for removal
of OH(V ) 1) and OD(V ) 1) by HNO3 and DNO3 and the
temperature dependence of these rate coefficients are reported
in Table 1 and displayed in Figure 4. The figure also shows the
previously reported values of k2 and k5 at room temperature;7
the agreement between the previously reported measurements
and values obtained in the present work is excellent. All of these
rate coefficients are much larger (by a factor of 50-1200) than
the rate coefficients for the reactions of OH(V ) 0) and OD(V
) 0) with HNO3 and DNO3 under similar experimental
0
) and OH(V ) 1) (back-extrapolated to time zero), normalized
13
by the ratios of the transition probabilities and the laser
fluences, gives the ratio of the yield of OH(V ) 1) to that of
OH(V ) 0). The ratio is ∼1%. The only reported yield of OH
(
2
2
V ) 1), relative to OH (V ) 0), from photolysis of HNO3 at
48 nm is an upper limit of 5%,14 while the relative yield at
41 nm has been reported to be 2%,15 in fair agreement with
our rough measurement. The quantum yield for OH production
1
6
from nitric acid photolysis at 248 nm is 1, or close to 1, so
the absolute quantum yield for OH(V ) 1) is about 1%. The
OH(V ) 1) LIF signal was a linear function of the photolysis
laser power, suggesting that multiple-photon processes were not
significant. Although we did not measure the yield of OD(V )
1
) from photolysis of DNO3, the signal levels for OD(V ) 1)
1,2
were comparable to those for OH(V ) 1) under similar
conditions, suggesting that the yield of OD(V ) 1) from 248-
nm photolysis of DNO3 is similar to that of OH(V ) 1) from
HNO3.
conditions. It is unlikely that this difference in rate coefficients
for removal of the vibrationally excited and vibrationally
unexcited radicals can be explained by an increase in the
reaction rate coefficients due to the vibrational excitation of
the OH or OD radical. Therefore, the primary removal process
for OH/OD(V ) 1) is likely relaxation of the vibrationally
To estimate the quantum yield of OH(V ) 2) relative to that
of OH(V ) 1), we compared the LIF signals of the two species

7
766 J. Phys. Chem. A, Vol. 107, No. 39, 2003
McCabe et al.
5
to Arrhenius
TABLE 1: Measured Rate Coefficients for the Loss of OH(W
1) or OD(W ) 1) by HNO and DNO
TABLE 2: Parameters for Fits of k
2
-k
)
3
3
Expression and Expression IIa
k(T) ) Ae-
Ea/RT
k(T) ) k(298)(T/298)-n
k(298)c
temperature [HNO
3
] or
a,b
k ( 2σc,d
reaction
(K)
3
[DNO ]
reaction
Ab
E
a
/R (K)
n
OH(V ) 1) + HNO
OH(V ) 1) + DNO
OD(V ) 1) + HNO
3
3
3
(k
(k
(k
2
3
4
)
)
)
252.3
4.5-30.7
5.3-31.1
7.3-43.9
3.5-32.4
4.2-25.0
2.3-18.5
31.3 ( 1.5
26.1 ( 1.6
25.5 ( 1.2
24.5 ( 1.3
17.6 ( 1.3
15.8 ( 1.5
OH(V ) 1) + HNO
OH(V ) 1) + DNO
OD(V ) 1) + HNO
OD(V ) 1) + DNO
3
3
3
3
40 ( 20 -520 ( 140 23.2 ( 1.3 1.8 ( 0.5
4.7 ( 3.4 -750 ( 200 5.8 ( 0.5 2.6 ( 0.8
7.2 ( 3.5 -700 ( 140 7.8 ( 0.4 2.4 ( 0.4
19 ( 15 -660 ( 220 17.9 ( 1.6 2.2 ( 0.8
2
2
2
3
3
71.0
95.3
95.5
30.9
73.1
a
Error bars are 2σ representations of the uncertainty of the fits.
b
-13
3
-1 -1
c
-12
3
-1
Units are 10 cm molecule
s . Units are 10 cm molecule
253.0
8.6-51.7
5.8-40.8
4.9-32.3
19.5-60.2
4.7-39.9
6.1-33.3
9.1 ( 0.4
6.9 ( 0.5
6.5 ( 0.5
5.0 ( 0.5
4.4 ( 0.4
3.8 ( 0.4
s-
1
.
2
2
2
3
3
76.5
95.6
95.7
33.7
63.0
The rate coefficients for the removal of OH/OD(V ) 2) by
HNO3/DNO3, k6-k9, are roughly 3-6 times larger than the
corresponding rate coefficients k2-k5. This ratio is roughly the
same for each isotopic combination. However, the ratio did
increase somewhat with increasing temperature, so the rate
coefficients k6-k9 appear to show a weaker negative temperature
dependence (see below) than the rate coefficients k2-k5.
Any proposed mechanism for these processes must explain
three aspects of the measured rate coefficients. First, all of the
rate coefficients correspond to quite facile loss of OH/OD(V )
252.7
22.7-118.7 11.4 ( 0.5
2
2
2
2
2
2
3
3
59.8
95.0
95.0
95.2
95.2
96.0
73.7
83.9
8.9-59.1
16.4-74.9
11.6-101.1
10.6-53.3
11.7-85.4
10.8-69.3
9.0-58.8
10.5 ( 0.5
8.4 ( 0.7
7.9 ( 0.3
8.3 ( 0.5
8.1 ( 0.5
7.8 ( 0.5
3.5 ( 0.5
16.5-100.9 4.75 ( 0.17
1). Thus, comparing the rate coefficients in Table 1 with simple
OD(V ) 1) + DNO
3
(k
5
)
259.8
10.8-45.7
9.9-82.5
7.1-39.7
8.0-77.7
10.2-76.6
4.9-52.6
25.1 ( 1.1
24.6 ( 0.8
15.3 ( 1.0
18.0 ( 0.6
19.4 ( 0.7
12.2 ( 0.4
estimates of the rate coefficients for collisions between OH/
OD and HNO3/DNO3 on a simple hard-sphere, or hard-sphere
with Lennard-Jones attraction, basis leads to collisional prob-
abilities for loss of OH/OD(V ) 1) of between 0.1 and 0.01.
Second, we note that all of the rate coefficients show a quite
steep negatiVe dependence on temperature and that they vary
with temperature in a very similar fashion. Table 2 lists the
parameters for nonlinear least-squares fits (unweighted) of the
rate coefficients k2-k5 to two temperature-dependence formal-
isms, the Arrhenius expression and the expression
2
2
2
2
3
59.8
95.2
95.2
95.2
73.1
a
Concentration of reactant in excess. Units are 1014 cm-3. c Units
b
are 10- cm molecule
12
3
-1 -1 d
s
. Reported uncertainties are statistical
only.
-
n
k(T) ) k(298)(T/298)
(II)
The Arrhenius fits are also shown in Figure 4. Thus, expressed
as negative activation energies, the temperature dependences
yield values of Ea/R varying from -(520 ( 140) K for reaction
2
to -(750 ( 200) K for reaction 3. If, instead, the rate
coefficients are fit to expression II, then n varies from (1.8 (
.5) for reaction 2 to (2.6 ( 0.8) for reaction 3. The negative
0
activation energies and n are not physically meaningful, but
rather are empirical quantities to describe the negative temper-
ature dependence of these reactions, within the temperature
range of these experiments. Finally, any proposed mechanism
should explain the differences between the absolute magnitudes
of the rate coefficients, particularly the fact that those for the
fully hydrogenated and fully deuterated isotopic pairs (reactions
2
and 5) are greater than those for the mixed isotopic pairs
(reactions 3 and 4).
One possible explanation for the larger values of k2 and k5
compared to k3 and k4 is that at least some fraction of the OH/
OD(V ) 1) loss occurs via near-resonant intermolecular vibra-
tion-vibration (V-V) energy transfer. Such processes can be
written as
Figure 4. Arrhenius representation of k
open symbols, ref 7. Circles, k ; diamonds, k
. The solid lines are Arrhenius fits (see Table 2 for fit parameters),
and the dashed lines are fits to expression III (see text). The error bars
are statistical only (2σ), representing the uncertainty from the fit of k′
vs [HNO ] or [DNO ].
3 3
2
-k
5
. Filled symbols, this work;
2
3
; squares, k ; triangles,
4
k
5
OH(V ) 1) + HNO f OH(V ) 0) + HNO (ν ),
3
3
1
-1
(∆E/hc) ) -20 cm (10)
OD(V ) 1) + DNO f OD(V ) 0) + DNO (ν ),
3
3
1
-1
excited radicals by HNO3 and DNO3. We also note that the
arguments presented below do not rely upon any assumption
about whether OH/OD(V ) 1) is lost by reaction with HNO3/
DNO3 or is relaxed by HNO3/DNO3 to OH/OD(V ) 0).
(∆E/hc) ) -11 cm (11)
where HNO3(ν1) and DNO3(ν1) denote molecules with one
quantum of vibrational excitation in the OH/OD stretching mode

Removal of OH and OD by HNO3 and DNO3
J. Phys. Chem. A, Vol. 107, No. 39, 2003 7767
of the molecule. In view of the close resonance between the
vibrational transition energies in the collision partners in these
processes, intermolecular V-V energy transfer might contribute
to the loss rate coefficients in the cases of OH(V ) 1) + HNO3
and OD(V ) 1) + DNO3. However, in the mixed isotopic pairs,
such V-V exchange processes
of OH(V ) 1) by HNO3 will be given by
) k [kIVR/(kIVR + k-a)]
k
(III)
X
a
We can recognize two limiting cases: (i) when kIVR . k-a and
kX ) ka and (ii) when kIVR , k-a and kX ) ka[kIVR/k-a]. When
strong covalent bonds are made to form a complex, as, for
6
OD(V ) 1) + HNO f OD(V ) 0) + HNO (ν ),
example, in OH(V ) 1) + NO2, the first limit is almost certainly
3
3
1
reached. In this case, association to form the complex is the
rate-determining step. The rate coefficient for loss of the
vibrationally excited radical then provides a good estimate of
the rate coefficient for association of the two species (usually
free radicals) in the limit of high pressure, and we would expect
little or no temperature dependence of the rate coefficient for
the loss of the vibrationally excited radical. For example, the
-
1
(∆E/hc) ) +908 cm (12)
OH(V ) 1) + DNO f OH(V ) 0) + DNO (ν ),
3
3
1
-
1
(∆E/hc) ) -940 cm (13)
are nonresonant, and reaction 12 is also strongly endothermic.
(Acceptance of the vibrational energy by other modes of HNO3
19
rate coefficients for the removal of CH(V ) 1) by H2 or D2
would not be endothermic, but it would be nonresonant.) Clearly,
intermolecular V-V energy transfer will not be efficient for
reactions 3 and 4. Yet, the rate coefficients k3 and k4 are still
quite large: for example, the room-temperature rate coefficients
for the removal of OH(V ) 1) by CH4 and CO2 are about 10
and 20 times smaller, respectively.18
20
and for the removal of OH(V ) 1) by SO2 show small negative
temperature dependences: when they are fit to expression II,
n, the absolute value of the exponent, is at most 0.27. If this
limit is reached in the present systems, we would also expect
the rate coefficients for the loss of OH/OD(V ) 1) to be similar
for all isotopic pairs that we have studied, as observed for the
These observations can be explained if the loss of OH/OD(V
1) in the presence of HNO3/DNO3 occurs via formation of
19
loss of CH(V ) 1) by H2 and D2.
)
In the second limit, kX will be much less than ka. Moreover,
the expression for kX is similar to that for recombination in the
limit of low pressure, with kIVR replacing the rate coefficient
for collisions that remove energy from the addition complex.
Generally, such rate coefficients show a steep, negatiVe tem-
perature dependence that can be viewed in one of two ways:
either one can show, via statistical mechanics, that the ratio ka/
k-a decreases with increasing temperature, or equivalently, one
can show that k-a increases more steeply with temperature than
either ka or the frequency of deactivating collisions. In the
present cases, we do not believe that either limit (i) (kIVR .
k-a) or limit (ii) (kIVR , k-a) applies. The temperature
dependence of the rate coefficients, which gets somewhat steeper
as the rate coefficients decrease, depends largely on the increase
in k-a as the temperature is increased. We note that, even for
OH(V ) 1) + HNO3, which has the largest value of kX, the rate
coefficients have an appreciable negative dependence on tem-
perature, strongly suggesting that, in all cases we have studied,
kX < ka and IVR is not very fast compared to the dissociation
of OH(V ) 1)‚HNO3*. Given the structure of the OH‚HNO3
the hydrogen-bonded complex that has been invoked to explain
the unusual kinetic behavior of the chemical reaction between
OH radicals and HNO3. This mechanism can be discussed in
terms of the following scheme of elementary processes, here
written for OH(V ) 1) + HNO3
2
Here, OH(V ) 1)‚HNO3* signifies a hydrogen-bonded complex
formed in collisions of OH(V ) 1) with HNO3 in which the
energy originally in the vibration of the OH radical remains
localized in this vibration in the complex. The rate coefficients
ka and k-a are those associated with the bimolecular formation
of this complex and its unimolecular dissociation, respectively.
kIVR is the rate coefficient for intramolecular vibrational
redistribution (IVR) of the energy from the OH radical vibration
in OH(V ) 1)‚HNO3* to the other modes of the complex,
yielding OH(V ) 0)‚HNO3**, in which the radical OH stretch
is unexcited, but which contains more energy in its other
vibrational modes. Given the far higher density of states
associated with the low mode frequencies, it is fair to assume
that the reverse transfer is unimportant within the lifetime of
the OH(V ) 0)‚HNO3** complex. This complex will dissociate
rapidly (i.e., k-b > k-a) on account of its high content of internal
energy relative to ground-state OH(V ) 0) + HNO3. Of course,
the complexes OH(V ) 1)‚HNO3* and OH(V ) 0)‚HNO3**
might also dissociate to form the products of reaction 1, H2O
and NO3 (not shown in above scheme). In fact, the rate
coefficient for dissociation to H2O and NO3 of OH(V ) 0)‚
HNO3** will likely be significantly larger than the rate
coefficient for dissociation to H2O and NO3 of OH(V ) 0)‚
HNO3*, the complex formed by the association of OH(V ) 0)
and HNO3, because of the additional energy of the OH(V )
4,5
complex presented in ab initio studies (see Figure 1), it seems
plausible that IVR is relatively slow from the radical OH stretch
into the other modes of the complex, as the OH is only loosely
bound (<8 kcal) at some distance (∼2 Å) from the HNO3
molecule in the complex.
2
1
Herbert et al. observed that the removal of CH(V ) 1) by
-
11
3
-1 -1
N2 is rapid [k(294 K) ) 3.0 × 10
cm molecule s ] and
that the rate coefficient for this process shows a moderate
negative temperature dependence between 86 and 584 K: when
fit to expression II, n was ∼1.2. They attribute the fast removal
of CH(V ) 1) by N2 to the formation of the weakly-bound CHN2
species: the strength of the dative bond between CH and N2 is
-
1 22
approximately 30 kcal mol . Herbert et al. suggest that the
CHN2 complex might dissociate before the CH vibrational
excitation energy is redistributed to the other modes of the
complex. Thus, the rate coefficient shows a negative temperature
dependence. Furthermore, high-pressure measurements23 of the
rate coefficients for the reaction of ground-state CH with N2
from 200 to 500 K show that the high-pressure limit for this
rate coefficient is nearly temperature-independent (n ) 0.15)
and it is slightly higher than the rate coefficients for loss of
CH(V ) 1) by N2. In the case of OH(V ) 1) + HNO3, the bond
0
)‚HNO3** complex. This additional energy is greater than the
4
calculated height of the barrier between OH(V ) 0)‚HNO3*
and H2O + NO3.
According to this mechanism, if the concentration of OH(V
1)‚HNO3* is in steady state, the rate coefficient for the loss
)

7
768 J. Phys. Chem. A, Vol. 107, No. 39, 2003
McCabe et al.
5
to
formed between the two species is weaker still, and the rate
coefficients show a stronger negative temperature dependence.
Finally, we address the question of why the magnitude of
the rate coefficients for loss of OH/OD(V ) 1) in the different
isotopic systems differ, with those for the like-isotope combina-
tions (k2 and k5) being larger than those for the unlike-isotope
combinations (k3 and k4). We believe that the most likely
explanation for this lies in the different rates of intramolecular
vibrational relaxation in the different systems. The fully
hydrogenated and fully deuterated systems offer a route for IVR
that is not available in the mixed systems, namely, V-V
transfer, within the complex, from the vibration associated with
the radical to the OH/OD stretch in HNO3/DNO3. This pathway
is nonresonant in the isotopically mixed complexes. Subsequent
transfer into the low-frequency modes, i.e., those within the
HNO3 or DNO3 moiety or the intermolecular modes of the
complex, is likely to be more efficient from these modes that
are strongly coupled to the low-frequency modes than directly
from the vibration of the OH/OD radical.
Intermolecular V-V energy transfer is probably not a major
contributor to any of the rate coefficients k2-k5. For the unlike-
isotope processes, reactions 3 and 4, the lack of resonance
between the collision partners means that intermolecular energy
transfer will be inefficient: rate coefficients similar to those
for the quenching of OH(V ) 1) by CH4 and CO218 would be
expected if the removal of OH/OD(V ) 1) occurred via
intermolecular energy transfer. The like-isotope processes are
faster, but the negative temperature dependences are quite
similar for the like- and unlike-isotope processes. Therefore,
the process that occurs in the like-isotope systems but not in
the unlike-isotope systems also has a strong negative temperature
dependence. Intermolecular V-V energy transfer should be
weakly temperature-dependent, and thus, it is unlikely to account
for the differences between the rate coefficients for the like-
and unlike-isotope processes. Therefore, the larger rate coef-
ficients for the like-isotope processes are probably due to faster
V-V transfer within the hydrogen-bonded complex.
TABLE 3: Values of kIVR from the Fit of k
2
-k
Expression IIIa
reaction
(kIVR ( 2σ)/1010 s-1
OH(V ) 1) + HNO
OH(V ) 1) + DNO
OD(V ) 1) + HNO
OD(V ) 1) + DNO
3
3
3
3
9.2 ( 4.4
1.7 ( 0.7
2.3 ( 0.9
6.4 ( 2.8
a
Error bars represent the uncertainty of the fits.
are given in Table 3. Because these values of kIVR depend on
the estimated value of A, their absolute magnitude is not as
significant as their values relative to one another. However, these
values correspond to reasonable lifetimes with respect to IVR
10-60 ps). For example, Wheeler et al. reported that the OH-
V ) 1) complex with CH4 has a lifetime of 38 ( 5 ps. The
fact that we are able to fit k2-k5 to expression III using
reasonable values for the elementary rate coefficients demon-
strates that our data are consistent with the mechanism we
present for the removal of OH(V ) 1) and OD(V ) 1) by HNO3
and DNO3. The mechanism not only accounts for the different
magnitude of the rate coefficients k2-k5 and their temperature
dependence, but also accounts for the slight increase in the
strength of the negative temperature dependence in the slower
reactions.
24
(
(
Furthermore, the rate coefficients we report here are consistent
with the observed pressure- and temperature-dependent kinetics
for the reaction of OH(V ) 0) with HNO3. Previous work from
2
our laboratory has simulated the pressure- and temperature-
dependent rate coefficient for this reaction, k1, using the
-
11
mechanism shown in Figure 1. ka was estimated to be 1 × 10
3
-1 -1
cm molecule s . This estimate was based on the previous
measurements of k2 and k5 by Smith and Williams, with the
7
assumption that some loss of OH(V ) 1) occurs by resonant
V-V transfer without formation of the OH‚HNO3 complex.
Here, we argue that resonant V-V transfer without complex
formation is unimportant and that ka is probably larger than the
rate coefficients presented here and previously for reactions
To test the validity of expression III, we fitted the rate
coefficients k2-k5 to this expression. Figure 4 shows, as dashed
lines, the results of the global fit. ka and k-a were assumed to
be the same for reactions 2-5; only kIVR was allowed to vary
between reactions. Furthermore, ka and kIVR were assumed to
be temperature-independent, whereas k-a was assumed to have
an Arrhenius temperature dependence
-
11
2-5: the fit of our data to expression III gives ka ) 7 × 10
3
-1 -1
cm molecule s . Our fit to expression III also provides the
temperature dependence of k . Using this updated information
about the elementary rate coefficients for the mechanism shown
in Figure 1, k can be simulated as described previously, with
only minor adjustments to the previously presented rate coef-
-a
1
ficients for the various steps of the reaction.
k-a ) A exp(-E /RT)
(IV)
a
We conclude that the loss of vibrationally excited OH(V )
) and OD(V ) 1) in the systems that we have studied occurs
1
-
11
3
-1 -1
The fitted value of ka is (7 ( 5) ×10
cm molecule s ,
via relaxation by formation of hydrogen-bonded complexes.
However, in none of these cases does the rate coefficient for
loss of OH/OD(V ) 1) reach that for bimolecular formation of
the collision complex (ka), because the rate coefficients for
redissociation of the initially formed complexes (k-a) and for
intramolecular vibrational redistribution of the OH/OD excitation
energy into the other modes of the complex (kIVR) are similar.
We attribute the negative temperature dependence of the values
of k2-k5 to the increasing rate of redissociation of the complexes
as the temperature is increased. Finally, we propose that different
rates of intramolecular vibrational energy redistribution might
be responsible for the different loss rate coefficients in the four
systems that we have studied.
indicating that the OH‚HNO3* complex is formed efficiently
from the free reactants. (Quoted errors represent the uncertainty
of the nonlinear least-squares fit, at the 2σ level.) The “activation
energy” for dissociation of the complex, Ea/R, is 840 ( 280 K.
The remaining parameters, kIVR and A, the preexponential factor
of k-a, are not independent variables in expression III, so they
cannot be assigned by the fit. However, k-a, like ka, should be
unaffected by the vibrational excitation of the OH radical: it
should be the same for OH(V ) 1)‚HNO3* and OH(V ) 0)‚
HNO3*, the complex formed by the association of OH(V ) 0)
and HNO3, because the extra energy in the OH(V ) 1)‚HNO3*
complex is localized in the radical OH stretch and does not
2
affect other modes of the complex. As described previously,
the ratio ka/k-a is roughly constrained by the thermodynamics
of the OH‚HNO3 complex, which allows us to assign an
approximate value for A of 3 × 10 s . The resulting values
of kIVR for the four different reactions from the constrained fit
Acknowledgment. This work was funded in part by the
NASA Upper Atmospheric Research Program. D.C.M. acknowl-
edges an NSF graduate research fellowship.
12
-1

Removal of OH and OD by HNO3 and DNO3
J. Phys. Chem. A, Vol. 107, No. 39, 2003 7769
References and Notes
(14) Schiffman, A.; Nelson, D. D.; Nesbitt, D. J. J. Chem. Phys. 1993,
8, 6935.
9
9
(1) Brown, S. S.; Talukdar, R. K.; Ravishankara, A. R. J. Phys. Chem.
(
15) Sinha, A.; Vander Wal, R. L.; Crim, F. F. J. Chem. Phys. 1989,
1, 2929.
16) Sander, S. P.; Friedl, R. R.; Golden, D. M.; Kurylo, M. J.; Huie,
A 1999, 103, 3031.
(2) Brown, S. S.; Burkholder, J. B.; Talukdar, R. K.; Ravishankara,
(
A. R. J. Phys. Chem. A 2001, 105, 1605.
R. E.; Orkin, V. L.; Moortgat, G. K.; Ravishankara, A. R.; Kolb, C. E.;
Molina, M. J.; Finlayson-Pitts, B. J. Chemical Kinetics and Photochemical
Data for Use in Atmospheric Studies, EValuation Number 14; JPL
Publication No. 02-25; Jet Propulsion Laboratory: Pasadena, CA, 2003.
(
(
(
(
(
3) Margitan, J. J.; Watson, R. T. J. Phys. Chem. 1982, 86, 3819.
4) Xia, W. S.; Lin, M. C. J. Chem. Phys. 2001, 114, 4522.
5) Aloisio, S.; Francisco, J. S. J. Phys. Chem. A 1999, 103, 6049.
6) Smith, I. W. M. J. Chem. Soc., Faraday Trans. 1997, 93, 3741.
7) Smith, I. W. M.; Williams, M. D. J. Chem. Soc., Faraday Trans.
(17) Suto, M.; Lee, L. C. J. Chem. Phys. 1984, 81, 1294.
(18) Raiche, G. A.; Jeffries, J. B.; Rensberger, K. J.; Crosley, D. R. J.
2
6
1
1985, 81, 1849.
8) Jaffer, D. H.; Smith, I. W. M. Faraday Discuss. Chem. Soc. 1979,
7, 212.
Chem. Phys. 1990, 92, 7258.
19) Brownsword, R. A.; Canosa, A.; Rowe, B. R.; Sims, I. R.; Smith,
I. W. M.; Stewart, D. W. A.; Symonds, A. C.; Travers, D. J. Chem. Phys.
(
(
(9) Vaghjiani, G. L.; Ravishankara, A. R. J. Phys. Chem. 1989, 93,
1
997, 106, 7662.
20) Blitz, M. A.; Hughes, K. J.; Pilling, M. J. J. Phys. Chem. A 2003,
07, 1971.
21) Herbert, L. B.; Sims, I. R.; Smith, I. W. M.; Stewart, D. W. A.;
948.
(
(
10) Dieke, G. H.; Crosswhite, H. M. J. Quant. Spectrosc. Radiat.
Transfer 1962, 2, 97.
11) Clyne, M. A. A.; Coxon, J. A.; Woon Fat, A. R. J. Mol. Spectrosc.
973, 46, 146.
12) MacLeod, H.; Smith, G. P.; Golden, D. M. J. Geophys. Res. 1988,
3, 3813.
13) Luque, J.; Crosley, D. R. LIFBASE: Database and Simulation
1
(
(
Symonds, A. C.; Canosa, A.; Rowe, B. R. J. Phys. Chem. 1996, 100, 14928.
(22) Moskaleva, L. V.; Lin, M. C. Z. Phys. Chem. 2001, 215, 1043 and
references therein.
(23) Fulle, D.; Hippler, H. J. Chem. Phys. 1996, 105, 5423.
(24) Wheeler, M. D.; Tsiouris, M.; Lester, M. I.; Lendvay, G. J. Chem.
Phys. 2000, 112, 6590.
1
(
9
(
Program, v. 1.6; SRI Report No. MP 99-009; SRI International: Menlo
Park, CA, 1999.