OH overtone spectra and intensities of pernitric acid

Article abstract of DOI:10.1016/S0009-2614(99)00858-1

The integrated absorption intensities of the 3ν_OH overtone transitions of the OH stretch of pernitric acid (HO₂NO₂) and nitric acid (HONO₂) were measured in concentrated sulfuric acid (H₂SO₄) solution. As well, oscillator strengths for the OH overtones of isolated HO₂NO₂ were calculated ab initio. Both results indicate that the OH overtones of HO₂NO₂ have larger oscillator strengths than the corresponding transitions in HONO₂ (1.3±0.5 times greater in H₂SO₄ solution; about 2 times greater according to the ab initio calculations). Overtone-induced photodissociation of HO₂NO₂ could play a role in HO_x production at high solar zenith angles.

Full text of DOI:10.1016/S0009-2614(99)00858-1

2
4 September 1999
Chemical Physics Letters 311 Ž1999. 131–138
www.elsevier.nlrlocatercplett
OH overtone spectra and intensities of pernitric acid
a
a,)
b
b
Lorien Fono , D.J. Donaldson , Robert J. Proos , Bryan R. Henry
a
Scarborough College and Department of Chemistry, UniÕersity of Toronto, Toronto, Ont. Canada M5S 1A1
b
Department of Chemistry and Biochemistry, UniÕersity of Guelph, Guelph, Ont. Canada N1G 2W1
Received 14 June 1999; in final form 19 July 1999
Abstract
The integrated absorption intensities of the 3nOH overtone transitions of the OH stretch of pernitric acid ŽHO NO . and
2
2
nitric acid ŽHONO . were measured in concentrated sulfuric acid ŽH SO . solution. As well, oscillator strengths for the OH
2
2
4
overtones of isolated HO NO were calculated ab initio. Both results indicate that the OH overtones of HO NO have
2
2
2
2
larger oscillator strengths than the corresponding transitions in HONO Ž1.3"0.5 times greater in H SO solution; about 2
2
2
4
times greater according to the ab initio calculations.. Overtone-induced photodissociation of HO NO could play a role in
2
2
HO production at high solar zenith angles. q 1999 Published by Elsevier Science B.V. All rights reserved.
x
1
. Introduction
To account for at least a portion of the ‘excess’
HO production, Donaldson et al. w4x recently pro-
x
Recently, measurements of HO Žthe sum of OH
posed a HOx production mechanism which relies
only on visible light, which is not strongly attenuated
at high SZAs. In this mechanism, dubbed ‘direct
overtone photodissociation’ ŽDOP., OH overtone
x
and HO . concentrations in the stratosphere have
shown a very rapid onset of HO production at very
high solar zenith angles ŽSZA., corresponding to the
sun’s being below the local horizon w1,2x. This is
surprising, and was not predicted by atmospheric
models, because in these models HO production is
predominantly due to the reaction of OŽ D. with
water. The OŽ D. species is produced predominantly
by UV photolysis of ozone at l(310 nm, although
there is a small quantum yield for its production at
longer wavelengths w3x. At high SZAs, ultraviolet
ŽUV. solar radiation is strongly attenuated by the
long atmospheric pathlength through the ozone layer,
so OŽ D. production is severely reduced, reducing
2
x
transitions of trace species X–OH present in the
stratosphere absorb visible radiation; if this light is
sufficiently energetic to cause the X–O bond to
cleave, then OH will be formed w5x. This ‘overtone-
x
1
1
mediated dissociation’ mechanism has been studied
for such molecules as HONO2 Žnitric acid. w6x,
HOOH Žhydrogen peroxide. w7x, HONO Žnitrous acid.
w8x and HOOCŽCH . Žt-butyl peroxide. w9x in the
3
3
context of measuring unimolecular reaction dynam-
ics.
Incorporation of the DOP mechanism into radia-
tive transfer models w4x suggested that of the known
1
HO production to a minimum.
x
species with low R–OH bond strengths present in
the stratosphere, H O , HNO , HONO and HO NO
2
2
3
2
2
Žpernitric acid., OH production from DOP of HOOH
would be insignificant, HNO3 and HONO would
)
Corresponding author. E-mail: jdonalds@chem.utoronto.ca
0
009-2614r99r$ - see front matter q 1999 Published by Elsevier Science B.V. All rights reserved.
PII: S0009-2614Ž99.00858-1

1
32
L. Fono et al.rChemical Physics Letters 311 (1999) 131–138
play minor roles, but that HO NO , because of its
Absorption spectra of HONO and HO NO were
2 2 2
2
2
y1
low N–O bond energy Žabout 8760 cm . w10x,
measured between 400 and 1100 nm in a commercial
single-beam spectrometer, using cells of pathlength
5.00 or 10.00 cm. Background spectra were mea-
sured at the same time and subtracted from the
sample spectra as detailed below. No absorbing
species are present in these mixtures other than those
could be a significant source of HO at high SZAs.
x
These calculations required some knowledge of
the absorption cross-sections for the OH overtone
transitions. In the initial work, these were estimated,
based on the values suggested by Crim w11x. Subse-
quently, we have performed a thorough study of the
of interest and H O. At the time of the visible scan
2
OH overtones in HONO w12x, combining measured
of the HO NO solutions, a UV spectrum from 250
2
2
2
and ab initio values to obtain cross-sections we
estimate to be good to better than 25%. In the
present work, we extend this study to HO NO . Ab
to 400 nm was also measured in a 1.00 cm cell, with
the reaction mixture diluted by a factor of 10 in
order to get an absorbance between 0.2 and 0.8 in
the 310–330 nm region.
2
2
initio values for the oscillator strengths are obtained
and are found to be a factor of 2 larger than those for
Neither H O nor H SO absorbs in this region of
2
2
4
HONO . The integrated absorption cross-sections for
the spectrum. To determine the concentration of
HO NO , the total absorbance of the reaction mix-
2
the 3nOH overtones of HO NO and HONO were
2
2
2
2
2
measured in concentrated H SO4 solution. The
ture at each of 310, 320 and 330 nm was measured.
The total absorbance at each wavelength was as-
sumed to be the sum of that due to absorption by
HO NO and by HOOH,
2
HO NO result is a factor of 1.3 times larger than
2
2
that of HONO2.
2
2
2
. Experimental
A
totalrlscH 2 O2
´
H 2 O2 qcHNO4
´
HNO4
.
1
Ž .
Solutions of various concentrations of HONO2
Simultaneous solution of Eq. Ž1. at any two wave-
lengths yielded calculated concentrations of both
HO NO and the remaining, unreacted HOOH. The
and HO NO in concentrated Ž95–98 wt%. H SO
2
2
2
4
were prepared volumetrically from stock concen-
trated solutions and cooled to room temperature prior
to measurement. Commercial reagent-grade concen-
2
2
absorption coefficient of aqueous HOOH at the
wavelengths used was measured here. The absorp-
tion coefficient of aqueous HO NO reported by
trated HNO was used; the exact concentration of the
3
2
2
stock solution was determined to be 79.3% Žwrw.
by titration against standardized KOH. The synthesis
of HO NO followed a modification of that given
Jesson et al. w14x at these wavelengths was used.
Errors in concentration were estimated to be the
standard deviation of the three values arising from
this procedure.
2
2
by Appleman and Gosztola w13x. In this procedure,
2
.7 g NaNO were dissolved in 11.8 ml 20% wrw
Visible spectra of H O in cold Ž;08C. H SO
2
2
2
4
HOOH and cooled to 08C. Next, 7.8 ml of 20 wt.%
HOOH were added to 0.75 ml of concentrated
H SO , stirred, and cooled to y238C in a carbon
Žand HOOH in H SO , in the case of HO NO . are
2 4 2 2
unstructured in the 3nOH and 4nOH regions from 700
to 1100 nm. These background spectra were scaled
and subtracted from the visible spectrum of the
reaction mixture in the OH overtone regions. The
scaling was varied Žwithin reasonable stoichiometric
limits. until the overtone peak had a flat baseline in
2
4
tetrachloriderdry-ice bath. The NaNO –peroxide so-
2
lution was added to the acid–peroxide solution drop-
wise, using a syringe. After each of the first four
additions of 2 ml, 0.5 ml of H SO was added to the
2
4
reaction mixture. This procedure gave a stock solu-
tion of approximately 1.75 M HO NO . Repeating
the subtracted spectrum. The integrated absorbance
of each overtone band was determined at a number
of concentrations; a plot of Arl versus concentra-
tion yields a straight line with the integrated cross-
section as slope. The errors in the integrated ab-
sorbance arise primarily from the range of scaling
which could be applied to the background spectrum
2
2
this procedure with smaller amounts of NaNO2
yielded lower concentrations of HO NO . The spec-
2
2
trum of the solution was measured immediately after
preparation, to minimize decomposition of the
HO NO .
2
2

L. Fono et al.rChemical Physics Letters 311 (1999) 131–138
133
and still result in a physically reasonable subtracted
spectrum.
In these calculations, the dipole moment function
is expressed as a seventh-order Taylor series expan-
sion in the internal O–H bond length co-ordinate, q,
expanded about the equilibrium position:
3
. Calculations
i
m
Ž
q
.
sÝ m q ,
5
Ž .
i
The ab initio determination of the overtone oscil-
where the coefficients m are given by:
i
lator strengths was performed in the same manner as
1
Eml
previously reported for HONO w12x. This method
2
m s
< .
6
Ž .
i
i
e
has proved very successful in reproducing experi-
mental intensities for a number of molecules, over a
range of six orders of magnitude in intensity w12,15–
i! ER
The equilibrium O–H bond distance was deter-
mined by performing a geometry optimization at the
2
1x. Intensities for C–H overtones are reproduced to
Ž
.
QCISDr6-31qG d,p level. The coefficients given
better than a factor of two; calculated O–H overtone
intensities are within 25% of measured values. The
Ž .
in Eq. 6 were obtained by changing the H–O bond
˚
˚
length by "0.4 A in increments of 0.1 A and
calculating the dipole moment at each position using
ab-initio quantum calculations. Numerical analysis
was used to give the coefficients w6x which were used
to construct the dipole moment function w5x. This, the
Morse wavefunctions and the transition energies
given by w3x, were used in Eq. Ž2. to calculate the
oscillator strengths.
oscillator strengths, f
transitions from < g: to <e: are given by w22x:
, of vibrational overtone
g ™ e
s4.70165=10y7
cm Dy2
2
f
Ž
.
n < m < ,
2
Ž .
g ™ e
e g
e g
where ne g gives the transition frequency in cmy1
and me g represents the transition dipole moment
matrix element, ²e< m< g: in debye. The oscillator
strength can be shown to be related to the integrated
absorption cross-section, s Žn ., by w22x:
4
. Results and discussion
1
2
fs1.1296=10
Hs Žn
.
dn .
4
.1. Experimental
In order to calculate the oscillator strength of an
overtone transition, the transition energy and the
dipole moment function are required. The wavefunc-
tions describing the OH stretch are approximated as
the eigenfunctions of a Morse oscillator whose tran-
sition energies to level Õ from the ground vibrational
Overtone spectra are observed in the 3nOH region
for both HNO and HO NO and in the 4n re-
gion for HNO . The spectra of HONO and HO NO
2
in the 3nOH region are displayed in Figs. 1 and 2,
respectively. In both molecules, at low concentra-
tions a double peak is observed in the 3nOH spectra,
with maxima at 980 and 1015 nm in HONO and at
9
trum, the bands lie very close to the measured gas-
3
2
2
OH
3
2
2
state, n
, are given by:
Õ§ 0
2
n
sÕvy
Ž
Õ qÕ
.
vx ,
3
Ž .
Õ§ 0
2
70 and 995 nm in HO NO . In the HONO spec-
where v and vx represent the vibrational frequency
and anharmonicity constants Žin cm ., respectively.
These parameters are not known experimentally for
HO NO , so were calculated here. For a Morse
2 2 2
y1
phase absorption w12x. No overtone spectra have been
reported previously for HO NO . The somewhat
2
2
2
2
broader red component of each double peak becomes
relatively less intense as the concentration is in-
creased. This effect is shown in Fig. 2 for HO NO .
We interpret the double peaks as being due to a
combination of solvation and hydrogen bondingr
complex formation effects. Since the higher-energy
component of the double peak becomes more pro-
nounced with increasing concentration, we assign it
to the complexed species; the lower-energy compo-
oscillator, v and vx are related to the bond dissoci-
ation energy by:
2
2
2
D sv r4vx .
4
Ž .
e
The dissociation energy was calculated ab initio,
using the GAUSSIAN98 suite of programs w23x. In
conjunction with the reported value of the fundamen-
y1
tal frequency Ž3540 cm . w24x, values for v and
vx could then be determined.

1
34
L. Fono et al.rChemical Physics Letters 311 (1999) 131–138
Plots of the integrated absorbance divided by the
cell pathlength of the 3nOH bands versus concentra-
tion, as given by Eq. Ž1., are shown in Fig. 3 for
HNO and in Fig. 4 for HO NO . These plots are
3
2
2
linear, with no indication of breaks or discontinu-
ities, suggesting that taking the sum of the integrated
absorbances of the two peaks which appear at lower
concentrations is a valid approach. The slopes of
these plots yield the integrated absorption cross-sec-
tions, which are presented in Table 1. The integrated
absorption cross-sections obtained for the 3nOH and
4
nOH transitions of HONO are factors of 3.6 and
2
6
.1 less than the corresponding values in the gas
y20
y21
2
phase, 2.63=10
and 2.37=10
cm mole-
cule cmy1, respectively w12x.
y1
Clearly, because of solvent effects on the over-
tone intensities, the solution-phase results should not
be directly compared with gas-phase values. How-
ever, the qualitatively similar nature of the overtone
spectra Ždouble peaks at lower concentrations; single
peaks at higher concentrations. suggests that solva-
Fig. 1. Overtone absorption spectrum in the 3 nOH region of 3.9
M HONO2 in concentrated H SO solution.
2
4
nent is assigned to the uncomplexed Žbut solvated.
species. We take the sum of the total integrated
intensity of both peaks for comparison to the inte-
grated absorption cross-section in the gas phase.
To ensure that the feature in the 3nOH spectral
region in the HO NO sample is really due to
2
2
HO NO , we performed several checks. A Raman
2
2
spectrum of a concentrated sample was measured
and showed peaks corresponding to those given by
Appleman and Gosztola w13x, in which the spectrum
was measured in perchloric acid solvent. As well, to
ensure that the overtone is not due to HONO, a
sample was prepared as above, but without addition
of H O . HONO may be synthesized from the reac-
2
2
tion of NaNO in Ždilute. H SO w25x. A very weak
2
2
4
peak is observed at 950 nm under these conditions;
no such feature is measured in either the HNO or
3
HO NO spectra. As well, the spectra of H O and
2
2
2
2
H O were each measured in dilute Ž2.6 M. H SO ;
2
2
4
these showed absorption features due to their 3nOH
bands at 1020 and 980 nm, respectively. We con-
clude that the absorption features shown in Fig. 2 are
assignable to the 3nOH overtone of HO NO .
Fig. 2. Overtone absorption spectrum in the 3 nOH region of
HO NO in concentrated H SO solution at two different con-
2
2
2
4
centrations. Upper panel: 0.5 M; lower panel: 1.2 M.
2
2

L. Fono et al.rChemical Physics Letters 311 (1999) 131–138
135
in conjunction with Eqs. Ž3. and Ž4. to predict the
vibrational frequency and anharmonicity constants.
The predicted and experimental values of these pa-
rameters are given in Table 3. Clearly, the best
agreement with experiment is shown by the
QCISDr6-31qGŽd,p. results, which lie within 0.2%
of experiment for the vibrational frequency constant
and 4.5% for the anharmonicity constant.
At this level of calculation, the calculated D for
e
HONO2 is 4.03% higher than that obtained using
Eq. Ž4. and the experimental values of v and vx.
Similarly, the ab initio values of D calculated for
e
H O and HOOH are overestimated by 2.04% and
2
3
.19%, respectively. Taking an average ‘overcalcula-
tion’ of D of 3.09%, we may correct the ab initio
e
values and use these corrected values to predict
vibrational parameters. These are displayed in Table
4
for all four molecules. Using the corrected parame-
ters for HNO3 yields predicted Žexperimental. w12x
Fig. 3. Beer’s law plot of the integrated absorbance divided by the
cell pathlength of the HONO2 3 nOH band as a function of
concentration in H SO . The units have been converted from the
2
4
conventional solution-phase units in order to obtain integrated
2
y1
y1
cross-sections in cm molecule
cm , for comparison to the
gas-phase values. The slope of this plot is the integrated absorp-
tion cross-section. It is Ž7.2"1.1.=10y21 cm molecule
2
y1
.
tion affects the overtone intensities of HONO and
2
HO NO similarly. If this is the case, the ratio of the
2
2
observed integrated absorption cross-sections of
HO NO to HNO may be used to scale the mea-
2
2
3
sured gas-phase intensity of HNO3 to obtain the
gas-phase intensity of HO NO . This procedure
2
2
yields a scaling of 1.3"0.5 for the intensity ratio of
the 3nOH band of HO NO acid relative to the same
2
2
band of HONO2.
.2. Calculations
The O–H dissociation energies of HO NO ,
4
2
2
HONO , H O and HOOH were calculated using two
2
2
basis sets Ž6-31GŽd. and 6-31qGŽd,p.. at the QCISD
Fig. 4. Beer’s law plot of the integrated absorbance of the
and QCISDŽT. levels. The results are shown in Table
HO NO 3 nOH band as a function of concentration in H SO .
2
2
2
4
2
. These values and the known fundamental frequen-
The slope of the best-fit line through the data points is Ž9.7"2.5.
y21
2
y1
cies for the OH stretch in these molecules were used
=10
cm molecule
.

1
36
L. Fono et al.rChemical Physics Letters 311 (1999) 131–138
Table 1
a
Integrated absorption cross-sections
Species
3nOH
4nOH
gasb
solution
gasb
solution
Ž2.63"0.26.=10y20
Ž7.5"1.1.=10y21
Ž9.7"2.5.=10
Ž2.37"0.24.=10y21
Ž3.9"1.1.=10y22
HNO3
y21
HO NO
2
2
a
y1
Integrated intensities in cm2 molecule cmy1
From Ref. w12x.
.
b
Table 2
Calculated H–O bond energiesa
Theory
HO NO
HONO2
HOH
HOOH
2
2
QCISDr6-31GŽd.
37622
40920
28806
54125
45241
46055
42872
46550
35977
38266
41346
28214
QCISDr6-31qGŽd,p.
QCISDŽT.r6-31GŽd.
a
y1
Energies in cm
.
transition energies in Žcmy1. to Õ s 3–6 of
rameters and transition energies as calculated above.
We also performed two other types of scaling to
obtain the vibrational parameters. In the first, the
ratio of the experimental values of the dimensionless
anharmonicity C Žsvr2vx . of HOOH ŽCs21.15.
1
1
0180 Ž10173.; 13261 Ž13245.; 16185 Ž16165.;
8952 Ž18940. respectively.
Table 5 presents the transition frequencies and
oscillator strengths calculated for HO NO . The first
2
2
column shows the results using the vibrational pa-
to that of H O ŽCs23.58. was used to scale that for
2
Table 3
Morse parameters for OH stretches
Theory
HO NO
HONO2
HOH
3881
87.8
3866
80.3
3919
107
3869.6
HOOH
2
2
QCISDr6-31GŽd.
v
3724
92.2
3708
84.0
3789
125
3675
62.4
3701
75.7
3698
74.2
3799
94.3
3783
86.5
3876
133
vx
v
QCISDr6-31qGŽd,p.
QCISDŽT.r6-31GŽd.
Experiment
vx
v
vx
v
b
c
d
3707
3789.4
89.6
vx
79
82.06
a
b
c
d
y1
Energies in cm
.
From Ref. w12x.
From Ref. w19x.
From Ref. w27x.
Table 4
Corrected morse parameters
HO NO
HONO2
HOH
HOOH
2
2
y1
D Žcm
v Žcm
.
39693
3714
43885
3706
78.3
45155
3871
40107
3789
e
y1
.
y1
vx Žcm
.
86.9
83.0
89.5

L. Fono et al.rChemical Physics Letters 311 (1999) 131–138
137
Table 5
Predicted peak positions and oscillator strengths for HO NO
2
2
V
Ab initio
C-scaling
D -scaling
e
n
Žcm
f
n
f
n
f
y1
Žcmy1.
Žcmy1.
.
y8
y8
y8
3
4
5
6
7
10099
13118
15963
18635
21132
5.66=10
4.02=10
4.86=10
7.13=10
1.15=10
10090
13100
15934
18591
21071
5.81=10
4.21=10
5.14=10
7.66=10
1.26=10
10076
13072
15887
18521
20973
6.06=10
4.52=10
5.63=10
8.60=10
1.46=10
y9
y9
y9
y10
y11
y11
y10
y11
y11
y10
y11
y11
HONO ŽCs23.46. to yield a predicted value for
result is robust. Given the uncertainties surrounding
gas-solution changes in overtone absorption
strengths, and the good agreement of theory and
2
HO NO ŽCs21.04.. Using this parameter with the
2
2
experimental fundamental frequency gives predicted
values of v and vx; the corresponding transition
energies and oscillator strength are shown in the
second column of Table 5. The second alternative
scaling method was to scale the O–H dissociation
energy of H O by that of H O, then to use that
experiment in the case of HONO , we feel that the
2
ab initio values are likely to be closer to the gas-phase
oscillator strengths for HO NO than those obtained
2
2
from the H SO solution experiments.
2
4
2
2
2
factor to predict the D for the OH bond of HO NO
e
2
2
5
. Conclusions
from that of HONO . This predicted D is then used
2
e
with the experimental value for the vibrational fun-
damental to obtain v and vx. The final column of
Table 5 shows the results of that approach for the
transition energies and oscillator strengths.
Both the ab initio results and the spectra measured
in H SO indicate that the OH overtone transitions
2
4
in HO NO have larger oscillator strengths than the
2
2
corresponding transitions of HONO . We conclude
2
To compare these results with experiment, we
take the ratio of the calculated HO NO oscillator
that the oscillator strengths for OH overtones in
HO NO are probably at least 1.5 times larger than
2
2
2
2
strengths divided by the corresponding values calcu-
the corresponding values for HONO . This result
2
lated w12x for HONO . The ab initio values of the
2
suggests that DOP of HO NO at high SZAs could
2
2
oscillator strengths for the 3nOH and 4nOH transi-
be significant w26x.
tions in HONO , calculated as outlined above but
2
using experimental transition frequencies, are 2.78=
y8
y9
1
0
and 2.14=10 , respectively w12x. These val-
Acknowledgements
ues differ on average by only 15% from the experi-
mentally measured oscillator strengths. The ratios of
the calculated oscillator strengths for HO NO2 to
The financial support of NSERC is gratefully
acknowledged. We thank Dr. A.F. Tuck and Prof. V.
Vaida for many productive discussions.
2
HONO range from 2.04 to 2.18 for the 3n
transi-
2
OH
tion, and 1.88–2.11 for the 4nOH transition, depend-
ing on which approach is used to calculate the
HO NO vibrational parameters.
References
2
2
Both experiment and theory indicate that the ab-
sorption cross-sections to the OH overtones in
HO NO are larger than the corresponding cross-
w x
1
P.O. Wennberg, R.C. Cohen, R.M. Stimpfle, J.P. Koplow,
J.G. Anderson, R.J. Salawitch, D.W. Fahey, E.L. Wood-
bridge, E.R. Keim, R.S. Gao, C.R. Webster, R.D. May, D.W.
Toohey, L.M. Avallone, M.H. Proffitt, M. Loewenstein, J.R.
Podolske, K.R. Chan, S.C. Wofsy, Science 266 Ž1994. 398.
2
2
sections in HONO . Every method used to calculate
2
the oscillator strength yields this qualitative result;
the quantitative agreement among the different ap-
proaches is quite good as well, suggesting that this
w2x P.O. Wennberg, R.J. Salawitch, D.J. Donaldson, T.F.
Hanisco, E.J. Lanzendorf, K.K. Perkins, S.A. Lloyd, V.
Vaida, R.S. Gao, E.J. Hintsa, R.C. Cohen, W.H. Schartz,

1
38
L. Fono et al.rChemical Physics Letters 311 (1999) 131–138
T.L. Kusterer, D.E. Anderson, Geophys. Res. Lett. 26 Ž1999.
373.
w18x H.G. Kjaergaard, B.R. Henry, A.W. Tarr, J. Chem. Phys. 94
1
Ž1991. 5844.
w3x A.R. Ravishankara, G. Hancock, M. Kawasaki, Y. Matsumi,
Science 280 Ž1998. 60, and references therein.
w4x D.J. Donaldson, G.J. Frost, K.H. Rosenlof, A.F. Tuck, V.
Vaida, Geophys. Res. Lett. 24 Ž1997. 2651.
w5x F.F. Crim, Annu. Rev. Phys. Chem. 44 Ž1993. 397.
w6x A. Sinha, R.L. Vander Wal, F.F. Crim, J. Chem. Phys. 92
Ž1990. 401.
w19x H.G. Kjaergaard, B.R. Henry, Mol. Phys. 83 Ž1994. 1099.
w20x H.G. Kjaergaard, R.J. Proos, D.M. Turnbull, B.R. Henry, J.
Phys. Chem. 100 Ž1996. 19273.
w21x H.G. Kjaergaard, C.D. Daub, B.R. Henry, Mol. Phys. 90
Ž1997. 201.
w22x P.W. Atkins, Molecular Quantum Mechanics, 2nd edn., Ox-
ford University Press, Oxford, 1983.
w7x L.J. Butler, T.M. Ticich, M.D. Likar, F.F. Crim, J. Chem.
w23x M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria,
Phys. 85 Ž1986. 2331.
M.A. Robb, J.R. Cheeseman, V.G. Zakrzewski, J.A. Mont-
gomery, Jr., R.E. Stratmann, J.C. Burant, S. Dapprich, J.M.
Millam, A.D. Daniels, K.N. Kudin, M.C. Strain, O. Farkas,
J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G.A. Petersson,
P.Y. Ayala, Q. Cui, K. Morokuma, D.K. Malick, A.D.
Rabuck, K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V.
Ortiz, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith,
M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez,
M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W.
Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon, E.S.
Replogle, J.A. Pople, GAUSSIAN98, Revision A.5, Gauss-
ian, Pittsburgh, PA, 1998.
w8x S.M. Holland, R.J. Strickland, M.N.R. Ashfold, D.A. Newn-
ham, I.M. Mills, J. Chem. Soc. Faraday Trans. 87 Ž1991.
3
461.
w9x M.D. Likar, J.E. Baggott, F.F. Crim, J. Chem. Phys. 90
Ž1989. 6266.
w10x 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
Ž1997. 521.
w11x F.F. Crim, Annu. Rev. Phys. Chem. 35 Ž1984. 657.
w12x D.J. Donaldson, J. Orlando, G. Tyndall, R. Proos, B. Henry,
V. Vaida, J. Phys. Chem. A 102 Ž1998. 5171.
w13x E.H. Appleman, D.J. Gosztola, Inorg. Chem. 34 Ž1995. 787.
w14x J.P. Jesson, L.C. Glasgow, D.L. Filkin, C. Miller, Geophys.
Res. Lett. 4 Ž1977. 513.
w24x H. Niki, P.D. Maker, C.M. Savage, L.P. Breitenbach, Chem.
w15x D.M. Turnbull, H.G. Kjaergaard, B.R. Henry, Chem. Phys.
Phys. Lett. 45 Ž1977. 564.
1
95 Ž1995. 129.
w25x A. Bongartz, J. Kames, F. Welter, U. Schurath, J. Phys.
Chem. 95 Ž1991. 1076.
w26x D.J. Donaldson, A.F. Tuck, V. Vaida, submitted to Physics
and Chemistry of the Earth, May 1999.
w27x H.G. Kjaergaard, J.D. Goddard, B.R. Henry, J. Chem. Phys.
95 Ž1991. 5556.
w16x H.G. Kjaergaard, D.M. Turnbull, B.R. Henry, J. Chem. Phys.
9
9 Ž1993. 9438.
w17x H.G. Kjaergaard, B.R. Henry, H. Wei, S. Lefebvre, T.
Carrington Jr., O.S. Mortensen, M.L. Sage, J. Chem. Phys.
1
00 Ž1994. 6228.