Exploration of the Mechanism of the Activation of ClONO2 by HCl in Small Water Clusters Using Electronic Structure Methods

Article abstract of DOI:10.1021/jp9932261

High-level electronic structure calculations were used to study the mechanism of the reaction of ClONO₂ with HCl in neutral water clusters containing one to five solvating water molecules. For the reaction between molecular HCl and ClONO₂, the barrier decreases from 42 kcal mol^-1 (uncatalyzed) to essentially zero when catalyzed by only two water molecules, where the reaction products involve Cl₂ and HONO₂. The calculations thus predict that the gas-phase reaction may be important in the stratospheric reactivation of ClONO₂. The reaction between ClONO₂ and solvated H₃O⁺Cl^-, as on the polar stratospheric cloud (PSC) surface, was investigated with clusters involving up to seven water molecules. The ice-catalyzed reaction involves an ionic mechanism whereby charge transfer to ClONO₂ from the attacking nucleophile leads to significant ionization along the Cl-ONO₂ bond. The effect of the size of the first solvation shell of Cl^- is addressed by our calculations. In a cluster containing three waters and a five-water cluster structurally related to hexagonal ice, ClONO₂ reacts spontaneously with HCl to yield Cl₂/HONO₂ in the three-water reaction and Cl₂/H₃O⁺-NO₃^- in the five-water-catalyzed reaction. The calculations thus predict that the reaction of ClONO₂ with HCl on PSC ice aerosols can proceed spontaneously via an ionic pathway.

Full text of DOI:10.1021/jp9932261

4030
J. Phys. Chem. A 2000, 104, 4030-4044
Exploration of the Mechanism of the Activation of ClONO2 by HCl in Small Water
Clusters Using Electronic Structure Methods
Jonathan P. McNamara, Gary Tresadern, and Ian H. Hillier*
Department of Chemistry, UniVersity of Manchester, Manchester, M13 9PL, United Kingdom
ReceiVed: September 13, 1999; In Final Form: January 5, 2000
High-level electronic structure calculations were used to study the mechanism of the reaction of ClONO2
with HCl in neutral water clusters containing one to five solvating water molecules. For the reaction between
molecular HCl and ClONO2, the barrier decreases from 42 kcal mol-1 (uncatalyzed) to essentially zero when
catalyzed by only two water molecules, where the reaction products involve Cl2 and HONO2. The calculations
thus predict that the gas-phase reaction may be important in the stratospheric reactivation of ClONO2. The
reaction between ClONO2 and solvated H3O+Cl-, as on the polar stratospheric cloud (PSC) surface, was
investigated with clusters involving up to seven water molecules. The ice-catalyzed reaction involves an
ionic mechanism whereby charge transfer to ClONO2 from the attacking nucleophile leads to significant
ionization along the Cl-ONO2 bond. The effect of the size of the first solvation shell of Cl- is addressed by
our calculations. In a cluster containing three waters and a five-water cluster structurally related to hexagonal
ice, ClONO2 reacts spontaneously with HCl to yield Cl2/HONO2 in the three-water reaction and Cl2/H3O+-
-
NO3 in the five-water-catalyzed reaction. The calculations thus predict that the reaction of ClONO2 with
HCl on PSC ice aerosols can proceed spontaneously via an ionic pathway.
1. Introduction
have been extensively studied by the complementary techniques
of Fourier transform (FTIR) or reflection-absorption infrared
spectroscopy (RAIR) and mass spectrometric (MS) methods.13-17
FTIR and RAIR spectroscopy allow direct observation and
discrimination of different surface-adsorbed species, whereas
MS methods allow determination of products evolved from
surfaces. Methods to study the interaction of ClONO2 with ion-
containing water clusters31-33 are also well documented.
Calculations indicating the polarized nature of the chlorine
nitrate molecule34 suggest that chlorine is the most accessible
electrophilic site and is thus susceptible to nucleophilic attack
by surface-adsorbed species,6,13-22 with the reactivity of ClONO2
being controlled by the relative nucleophilic/electrophilic strength
of the species involved.
Polar stratospheric cloud (PSC) particles have been implicated
as important sites for the heterogeneous catalysis of a number
of atmospheric reactions1,2 responsible for the depletion of ozone
in the Antarctic stratosphere.1-12 Of particular importance are
those reactions involving the conversion of the reservoir species
chlorine nitrate (ClONO2) and HCl to photochemically labile
Cl2. Photolysis of Cl2 leads to chlorine radicals capable of
destroying ozone via efficient catalytic chain reactions. The PSC
ice aerosols are classified as type I (nitric acid trihydrates, NAT)
or type II (water ice). However, it is now widely accepted that
the activation of both ClONO2 and HCl can occur throughout
the atmosphere on other atmospheric aerosols originating from
volcanic eruptions, such as Mount Pinatubo (1991).2
Under acidic conditions the direct reaction in eq 2 is believed
to be dominant over hydrolysis (eq 1).6,13-22 Wofsy et al.35
propose an acid-catalyzed mechanism producing incipient Cl+
or NO2+ leaving groups that then undergo ion-ion recombina-
tion with the solvated chloride anion of ionized HCl. The
majority of experimental evidence suggests that at temperatures
>50 K, ionized HCl is present at the surface36-41 as well as in
the bulk. Given the dynamic nature of the ice surface,42 it is
likely that adsorbates will be introduced into the bulk. Bianco
and Hynes43 propose two possibilities regarding the nature of
the ionized HCl. Dissociation involving a proton transfer to an
adjacent water molecule may lead to the formation of the
H3O+Cl- contact ion pair; or alternatively, there is a high
probability that the proton will be transferred far from the
ClONO2 reaction site such that the direct reaction
The activation of ClONO2 has been suggested to occur via a
two-step mechanism (eqs 1a and 1b) or via a direct mechanism
(eq 2).6,13-22
ClONO2 + H2O f HOCl + HONO2
HOCl + HCl f Cl2 + H2O
(1a)
(1b)
ClONO2 + HCl f Cl2 + HONO2
(2)
Both the hydrolysis reaction 1a, leading to hypochlorous acid
(HOCl) and nitric acid (HONO2), and reaction 2, which we refer
to as the direct reaction, have been the subject of a number of
experimental studies.6,13-22
The availability of modern laboratory apparatus allows the
control of experimental parameters such as reactant partial
pressure and substrate temperature so that conditions close to
stratospheric ones can be simulated. Hanson,18,22-24 Zhang,25,26
and other workers19-21,27-30 have used flow tubes to measure
reaction probabilities or sticking coefficients (γ). The hydrolysis
reaction and to a somewhat lesser extent, the direct reaction,
ClONO2 + Cl- f Cl2 + NO3
(3)
-
would then proceed. This reaction has been shown to be
thermodynamically favorable in the gas phase.44 Wincel et al.30
have also shown reaction 3 to occur efficiently at temperatures
10.1021/jp9932261 CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/27/2000

Electronic Structure Calculations of Activation of ClONO2
J. Phys. Chem. A, Vol. 104, No. 17, 2000 4031
between 170 and 298 K, when the chloride ion is complexed
with up to three water molecules.
At present there are no available experimental data regarding
the rate of the direct reaction, yet Horn et al.17 reported an
instantaneous production of Cl2 at 160 K, which indicates a
facile reaction. However, the catalytic role of water in both
homogeneous and heterogeneous catalysis has yet to be resolved
and will be addressed in this work by focusing on the following
issues:6,13-22
(i) Whether the direct reaction between ClONO2 and HCl is
ionic or molecular and particularly the role of water in this
reaction.
(ii) The reasons for the low activation barrier inferred from
the observed rapid reaction on PSC ice aerosols.
(iii) The role of the PSC ice aerosol in both catalyzing the
reaction and whether the product nitric acid is ionic or molecular.
The structure and properties of ClONO2 itself have been the
subject of a number of theoretical investigations. The structure
and harmonic frequencies of free ClONO2 were calculated by
Lee45 and Ying and Zhao46 using CCSD(T) and density
functional (DFT) theories, respectively. The structure of the
chlorine nitrate anion has been studied using Møller-Plesset
perturbation theory (MP2) by Seeley et al.47 Early ab initio
investigations of the reactivity of ClONO2 focused on the
mechanism of hydrolysis. Lee et al.48 studied the acid catalyzed
mechanism proposed by Wofsy et al.35 at the CCSD(T) level,
whereas La Manna34 calculated the structure of six isomers of
ClNO3 using Hartree-Fock (HF) and MP2 methods and
proposed that the strongest interaction with water occurs along
the Cl-O intermolecular axis. More recently, Bianco and
Hynes49 and Xu and Zhao50,51 modeled the hydrolysis of
ClONO2 in small water clusters leading to solvated molecular
HOCl and HONO2 at the MP2//HF level. The latter authors
highlighted the role of the ice surface in both structure catalysis
and hydration of heterogeneous reactions, where the hydrolysis
is compatible with the proposed ion-catalyzed mechanism.6,13-22
However, Bianco and Hynes49 have also identified the ionic
products (HOCl and H3O+NO3-) along an MP2//HF/6-31+G-
(d,p), STO-3G microsolvated reaction path, where the barrier
is 3 kcal mol-1. In accord with this result, the hydrolysis reaction
leading to the ionic products has also been shown by DFT
calculations to occur essentially spontaneously in a cluster
containing six water molecules that is related to hexagonal
ice.52,53
The direct reactivation of ClONO2 has been of some interest
theoretically. Haas et al.44 found no barrier for the gas-phase
reaction with the chloride anion (eq 3), whereas Mebel et al.54
found that the barrier for the gas-phase reaction with molecular
HCl (∼45-60 kcal mol-1) became negative when catalyzed
by NO3-. Bianco and Hynes43 reported the results of ab initio
calculations for the direct reaction on a model ice lattice
containing nine water molecules. In agreement with experi-
ment,6,13-22 they calculated a barrier of 6.4 kcal mol-1 for the
reaction of ClONO2 coordinated to the ion pair (H3O+Cl-),
leading to the ionic products H3O+NO3- and Cl2. Moreover, in
an ab initio study, Xu and Zhao55 calculated the direct reaction
to occur with essentially zero barrier in a relatively small water
cluster. Other important atmospheric reactions have also been
studied by ab initio methods.56-72
Figure 1. Fragment of proton-ordered hexagonal ice, showing water
molecules (A, B) removed to accommodate chlorine nitrate.
phase reaction involving small water droplets where HCl is un-
ionized occurs readily at stratospheric temperatures. For the
reaction in larger clusters related to the ice surface, where HCl
is ionized to form either the H3O+Cl- contact ion pair or the
solvent-separated ion pair, a facile reaction is also predicted.
The calculations indicate that the structure of the local adsorption
site is important in effecting ionization of species involved in
the reaction. Furthermore, our cluster models confirm the
experimental observations that at stratospherically relevant
temperatures, the final reaction products contain the ionized form
of nitric acid H3O+NO3- 6,13-22 Finally, in light of our calcula-
.
tions, the implications for stratospheric chemistry are discussed.
2. Modeling Small Water Droplets and the Ice Surface
The construction and orientation of water molecules at the
ice surface are largely unknown. When liquid water freezes at
atmospheric pressure, the oxygen atoms comprise a wurzite
lattice,73 where the hydrogen atoms are distributed throughout
the lattice along the O-O bonds according to the Bernal and
Fowler ice rules.74 The disorder in the proton positions exists
down to 0 K and is responsible for the residual entropy at this
temperature.75
The nature of the ice surface with particular reference to its
role in low-temperature heterogeneous catalysis has been
highlighted by a number of experimental studies. The structure
of the external surface of an ice film, crystallized on a Pt(111)
surface at 90 K, was investigated by Materer et al.76 Using a
variety of techniques [low-energy electron diffraction (LEED),
molecular dynamics (MD), and ab initio calculations], they
found that the ice surface had full bilayer termination (Figure
1). Devlin and Buch used FTIR spectroscopy and MD/Monte
Carlo simulations to investigate small adsorbate interactions with
ice-like surfaces.77-81 They assigned surface molecules to one
of three categories; three coordinated molecules with either
dangling hydrogen or dangling oxygen coordination and four
coordinated molecules with distorted tetrahedral geometry. Buch
et al.81 also noted that both simulation and experiment indicate
that the surface of ice contains rings of water molecules large
enough to accommodate several adsorbate species.
The bulk properties of the condensed phase have been the
focus of a number of theoretical studies.73,75,82-85 In particular,
MD simulations have been successful in elucidating the ioniza-
tion of HCl at the ice surface.36 However the interaction of small
atmospherically relevant species, such as HOCl and HCl, with
the ideal ice surface have also been studied by electronic
structure methods.58,60 Geiger et al.58 used a four-water cluster
In this paper we present the results of ab initio (MP2) and
DFT calculations to understand both the homogeneous gas-phase
reaction and the heterogeneous PSC aerosol-catalyzed reaction
of ClONO2 + HCl. A number of important issues are addressed
by our cluster models. The calculations indicate that the gas-

4032 J. Phys. Chem. A, Vol. 104, No. 17, 2000
McNamara et al.
(O2N3O5)
TABLE 1: Geometric Parameters of ClONO2
geometric parametera
method/basis set
r(Cl1-O2)
r(O2-N3)
r(N3-O4)
HF
r(N3-O5)
(Cl1O2N3)
(O2N3O4)
6-311++G(d,p)
6-311++G(3df,3pd)
1.663
1.642
1.372
1.368
1.165
1.164
1.166
1.163
116.4
116.0
111.0
111.1
118.4
118.3
MP2
6-311++G(d,p)
6-311++G(3df,3pd)
1.700
1.668
1.557
1.547
1.192
1.188
1.190
1.185
112.0
111.2
107.8
108.0
116.9
116.3
B3LYP
1.189
1.186
6-311++G(d,p)
6-311++G(3df,3pd)
1.711
1.679
1.516
1.513
1.188
1.184
113.9
113.4
108.5
108.8
117.6
117.2
other
1.195
1.196
CCSD(T)/TZ2Pb
experimentc
1.707
1.673
1.511
1.499
1.197
1.196
111.9
113.0
108.7
108.8
117.8
118.6
a Bond lengths (Å), angles (deg). Refer to Figure 2 for atom labeling. All levels of theory calculate ClONO2 to be planar. b Lee45. c Vibrationally
averaged gas-phase ClONO2 structure.45
excised from the ideal surface of hexagonal ice. Similarly Brown
and Doren60 studied the interaction of HOCl with both (H2O)4
and (H2O)26 cluster models excised from the hexagonal ice
surface. The latter models were constrained to retain the
structural features of the ideal surface. In view of the proposed
dynamic nature of the ice surface,42 the mechanisms of
atmospheric reactions were also explored in small water clusters
by high-level ab initio techniques. Vincent et al.64 elucidated
the mechanism of oxidation of SO2 by H2O2 in water droplets,
whereas Smith et al.65 examined the process of acid dissociation
in water clusters. Our recent calculations concerning the
hydrolysis of ClONO2,52,53 in addition to the work of Bianco
and Hynes,49 have shown these prototypical clusters to suitably
reproduce the observed reactivity of the PSC surface and small
water droplets. Models that represent the local adsorption site
by fragments excised from an ideal ice lattice58,60 may not
account for the solvating effect of the ice surface beyond the
adsorption site. As the foreign molecule approaches the ice
surface, there is a high probability that it will be at least partially
solvated by either surface-bound H2O42 or solvated within a
ring of water molecules.81 In view of Buch’s findings81 and the
dynamic nature of the ice surface,42 we chose to study the direct
reaction in water clusters, which are relevant to the study of
reactions in small water droplets and on the PSC aerosol surface.
Figure 2. Core reaction system.
as minima or transition structures on the potential energy surface
by calculation of harmonic vibrational frequencies. Intrinsic
reaction coordinate calculations were performed to confirm that
the transition structures connected the reactant and product
minima. Free energies were calculated within the perfect gas,
rigid rotor, harmonic oscillator approximation.
3. Computational Methods
The calculations presented herein were carried out using the
Gaussian 9486 and Gaussian 9887 suites of programs. Electron
correlation was included at the DFT (B3LYP)88-90 and MP291
levels. For the larger systems studied, DFT was chosen to
minimize computational expense, MP2 optimization being too
time-consuming for systems with ∼300 basis functions. The
B3LYP functional was chosen because previous calculations
by Ying and Zhao46 and our recent studies52,53 in which the
calculated structure of free ClONO2 at this level was shown to
be of comparable quality to that of the CCSD/TZ2P method45
(Table 1). All DFT optimizations used the flexible 6-311++G-
(d,p) gaussian basis set, because of recent electronic structure
calculations that show that both polarization and diffuse
functions are required to describe hydrogen bonded systems;
with smaller basis sets often leading to preferential stabilization
of ionic species.64 For the larger systems, single-point energy
calculations were carried out at the MP2/6-311++G(3df,3pd)//
B3LYP/6-311++G(d,p) level for comparison. Thus, the cal-
culations described are at a significantly higher level than those
previously reported.43 Stationary structures were characterized
4. Computational Results
In discussing the various structures we refer to the atom-
labeling scheme of the reactant pair (Figure 2), which contains
ClONO2 and the attacking nucleophile HCl. All of the structures
are denoted by the number of complete water molecules they
contain before reaction. Individual structural parameters are at
the B3LYP/6-311++G(d,p) level unless otherwise stated.
Relative energies and reaction barriers given in the proceeding
text are free energies at 180 K, a temperature appropriate to
the experimental conditions.6,13-22 The reaction energy is defined
as the difference in energy between the reactant and product
structures.
A. Gas-Phase ClONO2. First we consider the structure of
free ClONO2. Table 1 contains the optimized structure of
ClONO2 calculated using the HF, MP2, and DFT (B3LYP)
methods. Both the experimental structure45 and that from a
coupled cluster calculation by Lee45 are given for comparison.
The structure of ClONO2 at the HF/6-311++G(d,p) level shows

Electronic Structure Calculations of Activation of ClONO2
J. Phys. Chem. A, Vol. 104, No. 17, 2000 4033
TABLE 2: Mulliken Charges (e) B3LYP/6-311++G(d,p) of Reactant Pair (Figure 2)
atomic chargea
total fragment charge
reaction/structure
ClONO2
Cl1
O2
N3
O4
O5
Cl6
Hb
ClONO2
0.00
HCl
Cl2
H2Oc
HONO2
0.08
0.07
-0.26
0.03
0.08
ClONO2 + HCld
reactants
transition state
products
0.09
0.08
0.05
0.09
0.14
0.00
-0.28
-0.41
-0.32
0.04
0.08
0.01
0.07
0.14
0.06
-0.16
-0.25
-0.01
0.15 (7)
0.22 (7)
0.21 (7)
0.01
0.03
-0.01
0.04
0.01
-0.04
0.02
ClONO2 + HCl‚(H2O)e
reactants
transition state
products
0.08
0.01
-0.02
0.11
0.12
-0.02
-0.30
-0.34
-0.30
0.04
0.02
-0.01
0.02
-0.10
-0.08
-0.19
-0.25
0.03
0.20 (7)
0.17 (7)
0.43 (9)
-0.05
-0.29
0.01
-0.01
-0.03
0.04
0.37
ClONO2 + HCl‚(H2O)2 (isomer 1)f
reactants
transition state
Products
0.09
0.07
-0.06
0.11
0.12
-0.07
-0.33
-0.35
-0.35
0.03
0.02
0.01
0.00
-0.04
-0.03
-0.25
-0.31
0.02
0.24 (7)
0.25 (7)
0.46 (12)
-0.10
-0.18
0.09
0.23
-0.04
-0.01
0.02
ClONO2 + HCl‚(H2O)2 (isomer 2)g
reactants
transition state
products
0.08
0.05
-0.04
0.10
0.12
-0.03
-0.31
-0.33
-0.31
0.03
0.02
0.01
0.03
-0.03
-0.09
-0.24
-0.33
0.03
0.21(7)
0.22 (7)
0.46 (9)
-0.07
-0.17
0.08
0.26
0.04
a Refer to figures for atom labeling. b (In parentheses) atom label. c Atoms O8, H9 and H10. d Figure 4. e Figure 5. f Figure 6. g Figure 7.
important inadequacies, particularly with respect to the O2-N3
bond length. This length differs by ∼0.13 Å from the experi-
mental value and thereby overestimates its ionicity; however,
at this level, the Cl1-O2 bond distance (1.66 Å) is quite close
to the experimental value (1.67 Å). The addition of further
polarization functions [HF/6-311++G(3df,3pd)] decreases both
the Cl1-O2 and O2-N3 distances, thus increasing the discrep-
ancy between calculation and experiment. In contrast, the MP2/
6-311++G(d,p) geometry has Cl1-O2 and O2-N3 bond
lengths of 1.70 and 1.56 Å, respectively, which are somewhat
longer than the corresponding experimental values of 1.67 and
1.50 Å, respectively. In the B3LYP/6-311++G(d,p) structure,
the O2-N3 bond length (1.52 Å) is very close to the experi-
mental value (1.50 Å) and is in accord with the MP2 result,
with the Cl1-O2 bond of length 1.71 Å being overestimated.
Expansion of the basis [B3LYP/6-311++G(3df,3pd)] results
in a structure that is in excellent agreement with experiment,
with Cl1-O2 and O2-N3 bond lengths of 1.68 and 1.51 Å,
respectively. The Mulliken charges [B3LYP/6-311++G(d,p)]
for ClONO2 (Table 2) indicate a polarization of charge, with a
large negative charge of -0.26 on the nitrogen and a small
positive charge of 0.08 on the chlorine. Having considered the
structure of free ClONO2, we now briefly consider the reactivity
of HCl solvated in small water clusters relevant to the study of
Figure 3. HCl solvated by (a) three water molecules and four water
molecules, (b) isomer 1, (c) isomer 2, and (d) isomer 3. In this and
subsequent figures, distances are in angstroms (Å) and correspond to
the optimized B3LYP/6-311++G(d,p) geometries unless otherwise
stated.
the reaction of ClONO2 and HCl in such clusters.
B. HCl‚(H2O)3 and HCl‚(H2O)4. Recent ab initio and Monte
Carlo investigations have revealed that in smaller water clusters
with up to three solvating waters, un-ionized HCl is favored,59
whereas in clusters containing at least four water molecules,
ionized HCl is favored.65-67 Solvation of HCl in a three-water
ring (Figure 3a) leads to partial ionization of HCl (H-Cl 1.36
Å, isolated molecule 1.29 Å) and the addition of a further ring
water (isomer 1, Figure 3b) enhances this effect, which is in
agreement with previous investigations.59 We identified two
other minimum energy structures containing HCl solvated by
four water molecules. Isomer 2 (Figure 3c) is essentially the
HCl‚(H2O)3 cluster with an extra solvating water (A) and
contains un-ionized HCl. In contrast, isomer 3 (Figure 3d),
which is also based on the HCl‚(H2O)3 cluster, contains the ion
pair H3O+Cl-, with the molecular form of the acid (HCl) being
unstable in this cluster. Evidently, for isomer 3, the extra
bridging water (A) is able to stabilize both H3O+ and Cl-, each
of these species being stabilized by three hydrogen bonds. We
also note the cooperative nature of the ionization process where
in parallel with the lengthening H-Cl bond upon solvation, there
is a notable shortening of the ring hydrogen bonds. At 180 K,
isomers 2 (Figure 3c) and 3 (Figure 3d) containing the ionized
and un-ionized forms of HCl, respectively, are essentially
isoenergetic, whereas the single-ring structure (Figure 3b) is
∼2-3 kcal mol-1 lower in energy. Thus, in agreement with
previous studies,59,65-67 we have shown that HCl can exist in
both ionized and un-ionized forms depending not only on the
number of solvating waters but on their structural arrangement.
We now consider the reaction of ClONO2 and HCl in the
absence of water.
C. ClONO2 + HCl. A number of recent works49-51,55 have
modeled the reactions of ClONO2 using relatively low levels

4034 J. Phys. Chem. A, Vol. 104, No. 17, 2000
McNamara et al.
kcal mol-1 [MP2/6-31+G(d)] for the pathway we have studied.
Our calculations show the effect of different levels of theory
with barriers of 65.7 (HF), 55.9 (MP2), and 41.8 kcal mol-1
(B3LYP). There is also a variation in the calculated reaction
energies; HF predicts the product structure (Cl2/HONO2) to be
lower in energy than reactants (24.8 kcal mol-1), whereas MP2
calculates a much smaller value (9.8 kcal mol-1), and the
reaction energy is -13.8 kcal mol-1 at the DFT level. Notably,
the size of the basis used with a particular method results in
only small changes in the calculated barriers and reaction
energies. The inclusion of thermodynamic corrections (180 K)
yields barriers and reaction-free energies that do not radically
differ from the corresponding internal energy values (0 K; Table
3). We find that the reaction follows essentially the same
pathway at all levels of calculation; that is, in the reactant
complex, HCl is hydrogen bonded to O5 of ClONO2, and the
reaction occurs via a high-energy transition state containing
species akin to Cl- and protonated ClONO2. This transition state
leads to the weakly bound molecular Cl2/HONO2 product
complex. The structure of ClONO2 in the reactant complex
(Figure 4a) is closely related to that of the free molecule (Table
1), with the charge distribution in the complexed ClONO2 being
essentially the same as that in the isolated molecule (Table 2)
and with HCl being somewhat polarized with small charges of
-0.16 and 0.15 on Cl6 and H7, respectively. The transition
structure (Figure 4b) is characterized by an extension of the
Cl1-O2 bond and a transfer of H7 from Cl6 to O5 of ClONO2.
The structural parameters of the transition structure are close
to those at the MP2 level; namely, the increased Cl1-O2 distance
of 1.99 Å (1.71 Å reactant complex) and the compressed O2-
N3 bond (from 1.50 to 1.32 Å) of the forming HONO2.
However, the length of the forming Cl2 bond (2.62 Å) is
somewhat different to the MP2 value of 2.46 Å. Thus, Cl2
formation is more advanced in the MP2 transition structure. The
breaking Cl6-H7 bond (1.71 Å) is notably longer than the
forming O5-H7 bond (1.16 Å), which is consistent with the
ionization of HCl to form Cl- and protonated ClONO2. Evidence
supporting charge separation in the transition structure is
afforded by a Cl6 charge of -0.25 and a net charge of 0.25 on
the protonated ClONO2 entity (Table 2). The product complex
(Figure 4c) involves weakly bound Cl2 and molecular nitric acid
(HONO2) as evidenced by the nonbonded Cl1-O2 and Cl6-H7
distances of 4.02 and 2.57 Å, respectively. For the MP2 surface,
the Cl1-O2 nonbonded distance is somewhat shorter (3.45 Å),
indicating a stronger interaction.
Figure 4. Stationary structures for reaction of ClONO2 with HCl: (a)
reactants, (b) transition state, and (c) products (Cl2/HONO2).
of calculation [e.g., HF/6-31G(d)]. Thus, we calculated structures
at the HF level for comparison and to extend our study to
clusters involving explicit water molecules that may involve
the formation of ion pairs; a large basis including diffuse
functions was employed [6-311++G(d,p)] to correctly describe
anionic species.65 Furthermore, calculations on free ClONO2
(Table 1) indicate that to correctly describe the Cl1-O2 bond,
both f and d polarization functions must be included [6-311++G-
(3df,3pd)]. Therefore, we calculated stationary structures with
both these large basis sets to quantify the energetic and structural
effects of additional polarization and diffuse functions. The
stationary structures calculated at the B3LYP/6-311++G(d,p)
level are shown in Figure 4 and are representative of the
corresponding structures calculated at other levels. We now
discuss the structures and barriers obtained at the B3LYP/6-
311++G(d,p) level and compare them with the corresponding
barriers and structures calculated at the HF and MP2 levels with
this basis (Tables 3 and 4).
Finally, at all levels used (HF, MP2, DFT), the extension of
the basis to include more polarization functions to better describe
ClONO2 [6-311++G(3df,3pd)] results in small changes in the
structures. For example, in the transition structure (DFT), the
Cl1-O2 and Cl1-Cl6 distances both decrease to 1.93 and 2.57
Å, respectively, compared with 1.99 and 2.62 Å, respectively,
with the smaller basis [6-311++G(d,p)]. Having considered the
effect of model of electron correlation and basis sets on the
energetics and structures of the prototype reaction in the absence
of water, we now consider the catalytic effect of solvation by
a single water molecule.
D. ClONO2 + HCl‚(H2O). In view of the high-energy barrier
calculated for the reaction of ClONO2 with HCl (Table 3), we
studied the catalytic effect of the addition of a single water
molecule to our model reaction system (Figure 4a) to yield a
structure (Figure 5a) similar to that found previously in our study
of the hydrolysis of ClONO2.52 As for the uncatalyzed reaction,
we examined the effect of different models of electron correla-
tion and different basis sets on both the structures and energetics
Our study of the direct reaction in the gas phase follows that
reported by Mebel and Morokuma54 (also Xu and Zhao55) who
found barriers of ∼45 kcal mol-1 (B3LYP/6-31+G(d)) and ∼64

Electronic Structure Calculations of Activation of ClONO2
TABLE 3: Reaction Energies and Barriers (kcal mol-1) and Transition State Imaginary Frequencies (i cm-1
internal energy (0 K) free energy (180 K)
barrier reaction energy
J. Phys. Chem. A, Vol. 104, No. 17, 2000 4035
)
method/basis set
imaginary frequency
barrier
reaction energy
ClONO2 + HCl
HF
6-311++G(d,p)
6-311++G(3df,3pd)
MP2
376
326
64.0
62.0
-24.8
-25.7
65.7
64.0
-24.8
-25.1
6-311++G(d,p)
1674
54.8
-12.5
55.9
-9.8
B3LYP
6-311++G(d,p)a
6-311++G(3df,3pd)
1159
1123
40.9
41.7
-16.1
-17.7
41.8
43.0
-13.8
-14.2
ClONO2 + HCl‚(H2O)
HF/6-311++G(d,p)
MP2/6-31++G(d,p)
B3LYP/6-311++G(d,p)b
298
188
179
25.3
16.7
6.1
-28.3
-13.4
-18.1
34.7
21.0
5.1
-22.7
-11.0
-20.6
ClONO2 + HCl‚(H2O)2 (isomer 1)c
B3LYP/6-311++G(d,p)
MP2/6-311++G(3df,3pd)f
194
245
52
0.5
3.3
-18.8
-20.0
1.1
-17.5
3.9g
-18.7g
ClONO2 + HCl‚(H2O)2 (isomer 2)d
B3LYP/6-311++G(d,p)
MP2/6-311++G(3df,3pd)f
1.1
3.7
-18.8
-18.6
2.1
-16.7
4.8g
-16.5g
ClONO2 + HCl‚(H2O)3 (cluster 4)e
B3LYP/6-311++G(d,p)
MP2/6-311++G(3df,3pd)f
0.4
0.0
-7.7
-14.8
-0.1
-17.4
-0.3g
-24.4g
a Figure 4. b Figure 5. c Figure 6. d Figure 7. e Figure 10. f Single-point energy evaluations using B3LYP/6-311++G(d,p) structures. g Includes
thermodynamic correction at B3LYP/6-311++G(d,p) level.
TABLE 4: Geometric Parameters for ClONO2 + HCl Reaction
geometric parametera
b
method/basis set
Cl1-O2
O2-N3
N3-O4
HF
N3-O5
O5-H7
Cl6-H7
Cl1-Cl6
reactants
6-311++G(d,p)
6-311++G(3df,3pd)
transition state
6-311++G(d,p)
6-311++G(3df,3pd)
products
1.664
1.643
1.367
1.364
1.162
1.161
1.171
1.167
2.377
2.468
1.271
1.266
4.261
4.212
1.794
1.753
1.288
1.289
1.138
1.137
1.283
1.276
0.998
0.999
2.205
2.209
2.671
2.641
6-311++G(d,p)
6-311++G(3df,3pd)
3.515
3.409
1.184
1.182
1.165
1.163
1.332
1.327
0.950
0.949
4.030 (2.823)
3.923 (2.694)
2.001
1.975
MP2
reactants
6-311++G(d,p)c
transition state
6-311++G(d,p)
products
1.702
1.991
3.453
1.542
1.321
1.215
1.191
1.208
1.204
1.193
1.292
1.406
2.280
1.197
0.973
1.275
1.618
2.462
4.006
2.457
2.024
6-311++G(d,p)c
B3LYP
reactants
6-311++G(d,p)d
6-311++G(3df,3pd)
transition state
1.712
1.680
1.503
1.503
1.187
1.184
1.193
1.188
2.287
2.351
1.289
1.283
4.054
4.016
6-311++G(d,p)e
6-311++G(3df,3pd)
products
1.992
1.928
1.322
1.325
1.189
1.185
1.300
1.295
1.158
1.154
1.708
1.727
2.622
2.567
6-311++G(d,p)f
6-311++G(3df,3pd)
4.020
3.844
1.211
1.208
1.195
1.193
1.412
1.403
0.975
0.975
2.567
2.498
2.053
2.011
a Distances (Å). All structures essentially planar. Refer to Figure 4 for atom labeling. b (In parentheses) Cl1-H7. c Nonplanar. d Figure 4a. e Figure
4b. f Figure 4c.
of the reaction. The stationary structures calculated at the
B3LYP/6-311++G(d,p) level are shown in Figure 5 and are
representative of the corresponding structures calculated at other
levels. With this basis, the calculated barriers (Table 3) differ
widely; they are 34.7 (HF), 21.0 (MP2) and 5.1 kcal mol-1
(B3LYP), and are reduced considerably to 31.0 (HF), 34.9
(MP2), and 36.7 kcal mol-1 (B3LYP) compared with those
found in the absence of H2O. The reaction energies (Table 3)
are essentially unchanged at the HF and MP2 levels; they are,
respectively, -22.7 and -11.0 kcal mol-1, (-24.8 and -9.8
kcal mol-1 uncatalyzed). The B3LYP product structure is
somewhat more stabilized relative to reactants (20.6 kcal mol-1
)
compared with the uncatalyzed reaction (13.8 kcal mol-1). We
note that the inclusion of thermodynamic corrections increases
somewhat the barriers at both HF and MP2 levels, whereas the
barrier decreases at the DFT level (Table 3).

4036 J. Phys. Chem. A, Vol. 104, No. 17, 2000
McNamara et al.
frequency of the transition state (179i cm-1, Table 3) corre-
sponds to transfer of H7 to the adjacent water molecule (O8)
and of Cl1 to Cl6, forming Cl2. In the transition structure the
length of the breaking and forming Cl1-O2 and O2-N3 bonds
of ClONO2 and HONO2 (1.93 and 1.35 Å) are close to the MP2
values with the smaller basis (1.95 and 1.34 Å, Table 5). The
transfer of the acidic proton (H7) to the adjacent water molecule
(O8) is well advanced, with a Cl6-H7 distance of 1.70 Å,
indicating incipient Cl- formation. The product structure (Figure
5c), which involves the molecular acid and Cl2, is characterized
by a strong hydrogen bond between HONO2 and H2O (1.69
Å). However, Cl2 is only weakly bound, with nonbonded Cl1-
O2 and Cl1-H7 distances of 2.91 and 2.59 Å, respectively. A
similar product structure is obtained at the MP2 level (Table
5).
In view of our finding that the catalysis of the ClONO2 +
HCl reaction by one water molecule results in a decrease in the
barrier height from 41.8 to 5.1 kcal mol-1 [B3LYP/6-311++G-
(d,p)], we now consider the catalytic effect of solvation by two
water molecules.
E. ClONO2 + HCl‚(H2O)2. Because the DFT model yields
structures for the uncatalyzed and water-assisted reactions of
ClONO2 and HCl, close to the MP2 structures, we obtained
stationary structures for the ClONO2 + HCl‚(H2O)2 reaction at
the B3LYP/6-311++G(d,p) level. However, in view of the wide
variation in energetics calculated using different models, MP2
single-point energy calculations were carried out on all DFT
structures with the larger 6-311++G(3df,3pd) basis [shown to
more accurately describe ClONO2 (Table 1)]. The free energies
at the MP2 level include the thermodynamic correction from
the appropriate B3LYP calculation at 180 K. We examined two
possible structures in which ClONO2 is solvated with two water
molecules. The first, isomer 1 (Figure 6a) contains ClONO2
and HCl solvated in a single ring structure, where the arrange-
ment of the water molecules is related to a fragment excised
from the surface of an ideal ice crystal (similar to that used by
Bianco and Hynes43,49 to study the hydrolysis of ClONO2).
Isomer 2 (Figure 7a) contains essentially the reactant structure
from the reaction catalyzed by one water molecule (Figure 5a)
with an additional water hydrogen bonded to O8.
Isomer 1. The inclusion of an extra water into the ring
structure (Figure 6a) results in a lowering of the barrier to 1.1
kcal mol-1 (B3LYP), with a facile reaction also being predicted
at the MP2 level. The reaction energy (-17.5 kcal mol-1
,
Figure 5. Stationary structures for reaction of ClONO2 with HCl
solvated by one water molecule: (a) reactants, (b) transition state, and
(c) products (Cl2/HONO2).
B3LYP) is a little less than for the one-water reaction (Table
3). The reactant structure (Figure 6a) can be considered as a
fragment excised from the ideal ice surface49-52 where molecular
HCl is adsorbed and there is insufficient solvation to stabilize
the (potential) ionized acid (H+Cl-). Both ClONO2 and HCl
are somewhat dissociated within this cluster (Table 1). Partial
ionization of ClONO2, induced by charge transfer, is supported
by a formal charge of -0.10 associated with this species (Table
2). This result is consistent with an ab initio study of the chlorine
nitrate anion by Seeley et al.,47 where the addition of an electron
to neutral ClONO2 resulted in lengthening and shortening of
the Cl1-O2 and O2-N3 bonds, respectively. The central feature
of the predicted transition state (Figure 6b, imaginary frequency
194i cm-1) is that it corresponds to a simple proton transfer
from HCl to an adjacent water to generate the strong nucleophile
Cl-. The extent of ionization of ClONO2 in the transition state
is less than that in the one-water-assisted reaction, which is
evident from the Cl1-O2 and O2-N3 bond breaking and forming
distances of 1.84 and 1.39 Å, respectively. Furthermore, a formal
charge of -0.18 associated with ClONO2 indicates increased
At the DFT level, the reactant complex contains molecular
ClONO2, HCl, and H2O forming a single ring structure (Figure
5a). The reaction proceeds via a proton transfer mechanism to
an almost planar transition structure (Figure 5b) containing
partially ionized ClONO2 and essentially fully dissociated HCl.
Collapse of the transition state leads to the solvated molecular
product complex Cl2/HONO2‚H2O (Figure 5c). The structures
of both ClONO2 and HCl in the reactant complex (Figure 5a)
are perturbed only a little from those of the isolated molecules
(Table 1). The transition state (Figure 5b) highlights important
differences between the catalyzed and uncatalyzed reactions.
There is significant charge separation in the transition structure,
with a large negative charge of -0.29 associated with the
ionizing ClONO2 (-0.05 reactant complex). There is also a large
positive charge (0.54) on the protonated water entity (H2O +
H7), in contrast to the uncatalyzed reaction where the positive
charge (0.25) is on the protonated ClONO2 entity. The imaginary

Electronic Structure Calculations of Activation of ClONO2
J. Phys. Chem. A, Vol. 104, No. 17, 2000 4037
Figure 7. Stationary structures for reaction of ClONO2 with HCl
solvated by two water molecules (isomer 2): (a) reactants, (b) transition
state, and (c) products (Cl2/HONO2).
Figure 6. Stationary structures for reaction of ClONO2 with HCl
solvated by two water molecules (isomer 1): (a) reactants, (b) transition
state, and (c) products (Cl2/HONO2).
(H2O)2 complex, with nonbonded distances of 2.75 and 3.22
Å, respectively.
charge transfer to ClONO2 in the transition state compared with
the reactant structure. Compared with the one-water reaction,
the formation of a species akin to H3O+ is less well defined,
with the transferring proton (H7) being still some distance from
the accepting water molecule (1.26 Å), although formal charges
of 0.23 and 0.25 are associated with H2O and H7, respectively.
A shortening of the ring hydrogen bonds is also noted
[decreasing from 2.03 and 1.74 Å (reactant complex) to 1.90
and 1.59 Å (transition state), respectively]. The products of the
reaction (Figure 6c) are again molecular nitric acid and Cl2,
where HONO2 is solvated by two water molecules forming a
hydrogen-bonded ring structure. In agreement with the one-
water-assisted reaction, Cl2 is only weakly bound to the HONO2‚
Isomer 2. The second isomer (Figure 7a) is related to the
reactant structure in the one-water-catalyzed reaction (Figure
5a). We added an extra water molecule that is not directly
involved in the solvating ring to examine the catalytic effect of
a second solvation shell. At 180 K, isomer 2 is a little higher in
energy than isomer 1 (1.3 kcal mol-1). The calculated barriers
are a little larger than for the reaction involving isomer 1 [i.e.,
2.1 (B3LYP) and 4.8 kcal mol-1 (MP2)]. We now consider the
stationary structures. A number of differences between the
isomer 2 (Figure 7a) and isomer 1 (Figure 6a) reactant structures
are evident. First, both ClONO2 and HCl are ionized to a lesser
extent in the isomer 2 reactant structure than in isomer 1 and

4038 J. Phys. Chem. A, Vol. 104, No. 17, 2000
McNamara et al.
TABLE 5: Geometric Parameters for ClONO2 + HCl‚(H2O) Reaction
geometric parametera
O5-H9 O8-H9
HF/6-311++G(d,p)
b
method/basis set
Cl1-O2
O2-N3
N3-O4
N3-O5
1.173
O8-H10
H7-O8
Cl6-H7
Cl1-Cl6
reactants
transition state
products
1.667
1.885
3.150
1.362
1.285
1.189
1.162
1.165
1.168
2.283
1.618
0.963
0.944
0.991
1.782
0.941
0.950
0.942
1.977
1.002
0.944
1.282
1.948
3.596
2.361
2.002
1.222
1.318
4.192 (2.998)
MP2/6-31++G(d,p)
reactantsc
1.712
1.949
2.898
1.508
1.336
1.236
1.210
1.228
1.221
1.212
2.196
1.527
0.998
0.966
0.965
0.971
0.965
1.849
1.044
0.968
1.289
1.832
3.685 (2.515)
3.264
2.364
2.024
transition state
1.265
1.382
1.031
1.701
productsc
B3LYP/6-311++G(d,p)
reactantsd
1.734
1.928
2.914
1.470
1.354
1.221
1.190
1.201
1.199
1.200
2.210
1.678
0.998
0.967
0.963
0.968
0.963
1.762
1.115
0.967
1.321
1.697
3.855 (2.589)
3.096
2.478
2.057
transition statee
productsc,f
1.240
1.380
1.005
1.687
a Distances (Å). All structures essentially planar. Refer to Figure 5 for atom labeling. b (In parentheses) Cl1-H7 distance. c Nonplanar. d Figure
5a. e Figure 5b. f Figure 5c.
also the Cl1-Cl6 intermolecular distance in isomer 2 (3.03 Å)
is notably longer than that in isomer 1 (2.87 Å). These
differences possibly explain the small difference in barrier
heights (Table 3). We located a transition state for the two-
water-assisted reaction involving isomer 2 (Figure 7b). In line
with the reaction for isomer 1, the transition structure essentially
corresponds to a simple proton transfer. The internal hydrogen
bonds have notably shortened to 1.70 and 2.06 Å (from 1.82
and 2.30, respectively, in the reactant complex), and a formal
charge of -0.17 associated with ClONO2 indicates charge
transfer to this molecule. Collapse of the transition state yields
Cl2 and molecular HONO2 in a product structure (Figure 7c)
similar to that for the reaction of isomer 1 (Figure 6c) but 2.1
kcal mol-1 higher in energy.
We thus conclude that waters in a second solvation shell have
an appreciable catalytic effect, although to a lesser extent than
those directly involved in the solvating ring. We have shown
that a facile reaction can occur in a relatively small model
cluster, indicating the gas phase, water-catalyzed reaction of
ClONO2 + HCl may be important in contributing to the
atmospheric activation of chlorine. However, in the reactions
considered thus far, the products of the gas phase reaction
involve molecular nitric acid (HONO2), whereas at stratospheri-
cally relevant conditions the products of the ice-catalyzed
reaction involve the ionized acid (H3O+NO3-).6,13-22 Therefore,
we now consider the reaction catalyzed by three and four water
molecules.
F. ClONO2 + HCl‚(H2O)3 and ClONO2 + HCl‚(H2O)4.
To investigate the reasons for the formation of the ionic products
(Cl2/H3O+NO3-) observed at stratospherically relevant temp-
eratures,6,13-22 we considered essentially two different models.
The first model contains a single ring of water molecules and
reactants and is related to a fragment excised from the ideal ice
surface49-52 (Figures 8 and 9). The interaction of ClONO2 with
complete rings containing both three or four water molecules
and ionized HCl (Figures 10 and 11) was also studied. This
second model has features evident on the ice surface studied
by Buch et al.,81 who report rings of water molecules large
enough to accommodate several adsorbate species that are
proposed as the sites for acid ionization.
Single Ring Models. Our study begins by considering a range
of three- and four-water clusters analogous to those used by
other authors in the study of ClONO2 hydrolysis.49-52 With the
two-water reactant cluster (Figure 6a), we examined the effect
of solvation by a single water molecule at A or B, at distances
typical of hydrogen bonds in hexagonal ice (∼1.65 Å).
Optimization of the initial structures solvated at A or B led
Figure 8. ClONO2 and HCl solvated by three water molecules (cluster
1).
directly to molecular product clusters (Cl2/HONO2) analogous
to those found in the one- and two-water-catalyzed reactions,
with no stable reactant structure. However the addition of a third
ring-water molecule (cluster 1, Figure 8) leads to a stable
reactant structure containing essentially ClONO2 and HCl, where
the arrangement of the water molecules is related to a fragment
excised from the ideal ice surface with HCl adsorbed atop on
oxygen. In view of the very low barrier found for the two-water-
assisted reaction (Table 3), a facile reaction is probable for this
cluster and no transition state search was carried out.
We examined the further solvation of our three-water reactant
structure (cluster 1, Figure 8) with extra-ring water molecules
placed at A, B, and C. As expected, solvation at C, remote from
the site of acid ionization, yields an analogous structure
containing un-ionized HCl. However, solvation at A or B, close
to the proton acceptor of the ionizing acid (HCl), leads to
structures containing ionized HCl, as indicated in cluster 2
(Figure 9a) containing a well-defined hydroxonium ion and
strongly polarized ClONO2. Within this reactant structure, a
strong interaction between the transferring Cl1 and both the
forming nitrate and attacking Cl- is evident from Cl1-O2 and
Cl1-Cl6 bond breaking and forming distances of 2.01 and 2.38
Å, respectively. A shortened O2-N3 bond of 1.33 Å (1.52 Å
free molecule) indicates incipient nitrate formation, and further

Electronic Structure Calculations of Activation of ClONO2
J. Phys. Chem. A, Vol. 104, No. 17, 2000 4039
Figure 9. ClONO2 and ionized HCl solvated by four water mol-
ecules: (a) reactant structure (cluster 2) and (b) product structure (cluster
3).
evidence of the formation of hydroxonium ion is afforded by a
charge of 0.61 associated with the H3O entity (Table 6). The
calculated charges (Table 6) support the idea of charge transfer
from the chloride to ClONO2 being -0.30 on Cl6 and -0.40
on ClONO2. However, a Cl1-Cl6 distance 0.33 Å greater than
in free Cl2 indicates the reaction is far from complete. In view
of the very low barrier predicted for similar structures involving
two water molecules (Table 3), no transition state search was
carried out because a low barrier is also probable for this system.
Given that the arrangement of the water molecules in the
reactant-like structure is able to support the ionized acid,
(H3O+Cl-) and strongly ionized ClONO2, we investigated the
possibility that a product-like structure containing the ionized
nitric acid (H3O+NO3-) may also be stable within a related
structure.
Figure 10. Stationary structures for reaction of ClONO2 with ionized
HCl solvated by three water molecules: (a) reactants (cluster 4), (b)
transition state, and (c) products (Cl2/HONO2).
concerned, the product-like structure (Figure 9b) is 5.5 kcal
mol-1 (9.1 kcal mol-1, MP2) more stable than the reactant-like
cluster 2 (Figure 9a). Within this structure, formal charges of
0.71 (H3O) and -0.57 (NO3) indicate well-defined nitrate and
hydroxonium ions that are separated by a relatively short
hydrogen bond of length 1.49 Å. A nonbonded Cl1-O2 distance
of 2.32 Å results in a notable distortion of the Cl1-Cl6 distance
(2.16 Å) from the free molecule case (2.05 Å). These findings
suggest that four water molecules are required to support the
ionization of HCl and the formation of the ionized product nitric
acid (H3O+NO3-) which is in agreement with previous calcula-
tions.66,67 However, to understand the reaction on the PSC
We identified a product-like structure containing the ionized
nitric acid contact ion pair (cluster 3, Figure 9b). This new ion-
pair structure was obtained from the first ion-pair structure by
a manual shift of a proton of the hydroxonium ion onto an
adjacent water molecule. As far as overall energetics are

4040 J. Phys. Chem. A, Vol. 104, No. 17, 2000
McNamara et al.
TABLE 6: Mulliken Charges (e) B3LYP/6-311++G(d,p) of Reactant Pair (Figure 2) in Clusters
atomic chargea
total fragment charge
Cl2 H3O
cluster
Cl1
O2
N3
O4
O5
Cl6
H7
ClONO2
HCl
NO3
HONO2
single ring models
ClONO2 + HCl‚(H2O)3
0.02 -0.27 0.24
1b
0.11
0.00
0.11 -0.38
0.04
-0.10
-0.40
-0.03
ClONO2 + HCl‚(H2O)4
0.08 -0.37 -0.02 -0.09 -0.30
2c reactants
3d products
-0.30 0.61 -0.40
-0.17 0.71 -0.57
-0.04 -0.01 -0.27 -0.06 -0.23 -0.13
ice surface ring models
ClONO2 + HCl‚(H2O)3
0.08 -0.33 -0.02 -0.17 -0.29
0.04 -0.32 -0.02 -0.17 -0.21
-0.06 -0.08 -0.31 0.02 -0.07 0.02
ClONO2 + HCl‚(H2O)4
0.08 -0.38 -0.01 -0.10 -0.37
0.05 -0.32 -0.01 -0.14 -0.31
ClONO2 + HCl‚(H2O)5
-0.03 -0.01 -0.29 -0.04 -0.26 -0.14
4e reactants
-0.04
transition state -0.06
products
-0.48
-0.53
-0.33 0.78 -0.44
-0.27 0.76 -0.47
-0.04
0.01
5f
-0.02
-0.03
-0.43
-0.45
-0.39 0.77 -0.41
-0.34 0.76 -0.42
6g
7h
-0.17 0.65 -0.60
a Refer to Figures for atom labeling. b Figure 8. c Figure 9a. d Figure 9b. e Figure 10. f Figure 11a. g Figure 11b. h Figure 12b.
ionization to produce the contact ion pair, H3O+Cl-, or
alternatively there is a high probability that proton transfer will
occur yielding a solvent-separated ion pair. In both cases,
desolvation of Cl- is required for reaction. Thus, we investigated
both the structural and energetic effects of increasing the first
solvation shell around Cl- from one to two waters and of
separating the ion pair (H3O+Cl-) by up to two layers of water
molecules.
Our first model system contains ClONO2 solvated by three
water molecules (cluster 4, Figure 10a). Within this structure,
ClONO2 is coordinated to the ionized acid, where the ion pair
(H3O+Cl-) is separated by a single layer of water molecules,
and the nucleophile Cl6 is solvated by two waters. Energetically,
this structure is 5.0 kcal mol-1 (2.1 kcal mol-1, MP2) higher in
energy than in the three-water structure (cluster 1, Figure 8)
containing un-ionized HCl; the small difference in stability is
probably due to the strained hydrogen bonds in cluster 4.
Compared with the three-water structure (Figure 8) containing
un-ionized HCl, cluster 4 (Figure 10a) contains a significantly
more polarized ClONO2 molecule, with respectively lengthened
and shortened Cl1-O2 and O2-N3 bonds (2.09 and 1.31 Å)
and a forming Cl1-Cl6 bond (2.30 Å) that is a little more
product-like. Charge transfer from the attacking nucleophile
(Cl6) is well advanced with a formal charge (Table 7) of -0.48
associated with the ClONO2 entity. In line with this, the large
positive charge of 0.78 on H3O indicates a well-defined
hydroxonium ion. Calculation of the transition state (Figure 10b)
leading to the molecular products (Cl2/HONO2) yields an
essentially zero barrier at both the B3LYP and MP2 levels
(Table 3). The charge distribution within this transition structure
is essentially the same as that in the reactant structure [-0.47
(NO3) and 0.76 (H3O)], and the transferring Cl1 is approximately
midway between O2 of the departing nitrate and Cl6 of the
forming Cl2. The product structure (Figure 10c) is formed by
the transfer of a proton from the hydroxonium ion adjacent to
O5 of the nitrate to yield solvated HONO2 and Cl2, and is 17.4
kcal mol-1 lower in energy than the reactant cluster (24.4 kcal
mol-1, MP2). We note that in the transition structure identified,
the developing nitrate is poorly solvated, whereas on the PSC
surface, the presence of additional solvating waters around the
reactive site may stabilize the developing nitrate entity and
collapse to ionized nitric acid (H3O+NO3-).
Figure 11. ClONO2 and ionized HCl solvated by four water mol-
ecules: (a) reactant structure (cluster 5) and (b) reactant structure
(cluster 6).
surface, we must increase the size of the first solvation shell of
Cl-, which may be important in determining the barrier to
reaction.
Ice Surface Ring Models. We now consider a second type of
cluster related to the rings identified on the 140 K annealed ice
surface studied by Buch et al.81 In a recent study of the direct
reaction, Bianco and Hynes43 identified two possible scenarios
regarding the fate of the ionized acid HCl. The first involves
To investigate the structural effects of increasing the number

Electronic Structure Calculations of Activation of ClONO2
J. Phys. Chem. A, Vol. 104, No. 17, 2000 4041
TABLE 7: Internal and Free Energies (Hartrees) of Clusters
internal energy (0 K)
free energy (180 K)
cluster
B3LYP/6-311++G(d,p)
MP2/6-311++G(3df,3pd)a
B3LYP/6-311++G(d,p)
MP2/6-311++G(3df,3pd)b
single ring models
ClONO2 + HCl‚(H2O)3
-1428.842877
1c
-1430.785544
-1430.712002
-1428.769335
ClONO2 + HCl‚(H2O)4
-1505.180172
-1505.194843
2d reactants
3e products
-1507.266260
-1507.275071
-1507.165753
-1507.174447
-1505.079665
-1505.094219
ice surface ring models
ClONO2 + HCl‚(H2O)3
-1428.853780
4f reactants
transition state
products
-1430.801110
-1430.800428
-1430.813415
-1430.720034
-1430.719856
-1430.747735
-1428.772704
-1428.773155
-1428.811606
-1428.853727
-1428.877286
ClONO2 + HCl‚(H2O)4
-1505.193708
-1505.197735
5g
6h
-1507.277148
-1507.280516
-1507.172641
-1507.176147
-1505.089201
-1505.093366
ClONO2 + HCl.(H2O)5
-1581.541919
7i
-1583.757755
-1583.629823
-1581.413987
a Single-point energy evaluation using B3LYP/6-311++G(d,p) structure. b Includes thermodynamic correction at B3LYP/6-311++G(d,p) level.
c Figure 8. d Figure 9a. e Figure 9b. f Figure 10. g Figure 11a. h Figure 11b. i Figure 12b.
of solvation shells around the reacting species, we located two
more minimum energy structures containing four water mol-
ecules. The first, cluster 5 (Figure 11a), is related to the three-
water cluster (Figure 10a) but with a single water inserted
between the hydroxonium ion and the developing nitrate of the
ionizing ClONO2. The extra water essentially represents a
surface-bound water above the plane of the ring solvating the
ionized acid. Cluster 6 (Figure 11b) is also related to the three-
water cluster (Figure 10a), but with an extra water molecule
inserted into the ring solvating the acid such that the ion pair
(H3O+Cl-) is now separated by single and double layers of water
molecules. Both these clusters have features evident on the ice
surface studied by Buch et al.81
As far as overall energetics are concerned (Table 7) an extra
internal hydrogen bond results in clusters 5 and 6 being,
respectively, 4.3 and 6.5 kcal mol-1 lower in energy than the
single-ring cluster 2 (Figure 9a). Turning now to consider the
structures of these species, in cluster 6 (Figure 11b) for example,
ClONO2 is significantly ionized. A breaking Cl1-O2 bond of
length 2.09 Å and a net charge of -0.45 (Table 6) associated
with the ClONO2 entity support the concept of charge transfer
from Cl6 affecting ionization of ClONO2. A shortened O2-N3
bond of 1.31 Å indicates incipient nitrate formation, and a charge
of 0.76 on H3O is evidence for a well-defined hydroxonium
ion. Having considered the ClONO2 + HCl reaction catalyzed
by both three and four water molecules in two different types
of structure related to the PSC surface, we now examine the
reaction in water clusters with structural arrangements of water
molecules closely related to that of hexagonal ice.52
G. ClONO2 + HCl‚(H2O)5. We investigated the reactivity
of ClONO2 and H3O+Cl- in a five-water cluster where the
arrangement of the water molecules is structurally related to
ordinary hexagonal ice. Figure 1 depicts a proton-ordered phase
of hexagonal ice with the model cluster water positions shown.
Surface bilayer water molecules (A, B) were removed to
accommodate ClONO2. In view of the findings of Buch et al.,81
our model system could be present either as part of a ring
structure reported on the annealed nanocrystal surface or as a
surface defect found on the unannealed surface.
The initial reactant structure was constructed by taking our
structure containing ClONO2 hydrated by six water molecules
(Figure 12a), used previously in the study of the hydrolysis
Figure 12. (a) Initial hydrolysis structure, ClONO2 solvated by six
water molecules; (b) ClONO2 and ionized HCl solvated by fiVe water
molecules, product structure (cluster 7, Cl2/H3O+NO3-).
reaction,52,53 and replacing the nucleophilic water molecule by
Cl- and protonating the adjacent water molecule to form the
contact ion pair H3O+Cl-. Optimization of this structure led
directly to the product-like structure (cluster 7, Figure 12b)
containing Cl2 and ionized nitric acid (H3O+NO3-). A strong
interaction between Cl2 and both the nitrate and hydroxonium

4042 J. Phys. Chem. A, Vol. 104, No. 17, 2000
McNamara et al.
ions is evident from the extended Cl1-Cl6 bond length of 2.18
Å [free molecule, 2.05 Å, B3LYP/6-311++G(d,p)]. Further
evidence of the strong interaction with the ionized acid is
afforded by a charge of -0.17 on Cl2. Thus, ClONO2 reacts
spontaneously with HCl to form ionic products in a five-water
cluster structurally related to hexagonal ice. However, the initial
reactant-like structure contains Cl- solvated by only a single
water, whereas the first solvation shell of Cl- may contain up
to three waters on a PSC surface.39-41,43 In view of these results,
we investigated the reactivity of ClONO2 in a corresponding
seven-water cluster in which two additional waters were added
to complete the first solvation shell of Cl-, at distances typical
of H-Cl nonbonded distance found in the clusters discussed
so far (∼2.35 Å). Optimization of this cluster at the B3LYP/
6-311++G(d,p) level leads to some distortion of the initial
arrangement of water molecules, however, the calculations
indicate that the final product structure contains species akin to
ionized nitric acid (H3O+NO3-) and Cl2, which is very close to
those identified in the five-water cluster (Figure 12b). These
findings suggest the reaction on the PSC surface will proceed
essentially spontaneously, even when Cl- has a complete first
solvation shell.
structures (Figures 5a, 6a, and 7a), and supporting this is a small
transfer of charge from the substrate (HCl water cluster) to
ClONO2. This effect, although small, is evidence for a change
toward an ionic mechanism as the number of solvating waters
is increased. A central feature of the transition states is that
they are somewhat akin to simple proton transfers from HCl to
ClONO2 in the absence of water (Figure 4b) and to an adjacent
water molecule in the water-assisted reactions (Figure 5b, 6b,
and 7b).
The PSC-catalyzed reaction, differs from the gas-phase
reaction in that the reactant structures now contain H3O+Cl-
and strongly ionized ClONO2 (Figures 9a, 10a, 11a, and 11b).
We considered two different types of model system. The first
model involves single rings of water molecules and is related
to fragments excised from an ideal ice surface, as used in
previous studies of the hydrolysis reaction.49-52 The second
model is related to the rings identified on the 140 K annealed
ice surface studied by Buch et al.81 These minimum-energy
structures have revealed at a molecular level a number of
important details concerning the mechanism of the reaction and
are in agreement with previous theoretical findings.43
(i) Laboratory studies and MD simulations36-41 suggest that
at stratospheric conditions ionized HCl is present at the ice
surface. In agreement with this result, our reactant structures
(Figures 9a, 10a, 11a, and 11b), which are representative of the
local adsorption sites on a PSC surface, all contain H3O+Cl-.
The reactant clusters also account for the ionized acid forming
a contact-ion pair (Figure 9a) or the solvent-separated ion pair
(Figures 10a, 11a, and 11b). Comparison of the four-water
clusters 2 and 6 suggests that the separation of H3O+Cl- by up
to two waters leads to a stronger nucleophile and thus to
increased charge transfer, resulting in increased ionization along
the Cl-ONO2 bond. Bianco and Hynes43 suggest the separation
of the ion pair would lead to the reaction proceeding as reaction
3. Thus, we conclude that the solvent separation of the H3O+Cl-
pair is an important step for a facile reaction on a PSC surface.
5. Discussion
We begin by commenting on the choice of basis set and
model of electron correlation. For ClONO2 itself (Table 1), we
find that the DFT (B3LYP) method with the 6-311++G(d,p)
basis yields a structure of essentially the same quality as the
CCSD(T)/TZ2P method45 and is in good agreement with the
experimental vibrationally averaged gas-phase structure.45 As
far as the uncatalyzed (Table 4, Figure 4) and the one-water-
assisted (Table 5, Figure 5) reactions of ClONO2 with HCl are
concerned, the transition structures calculated at the DFT level
are very close to the corresponding MP2 structures. In agreement
with recent studies,52,53,65 the 6-311++G(d,p) basis was used
to more accurately model the potential energy surface, because
both polarization and diffuse functions are required to describe
hydrogen bonded systems, with smaller basis sets tending to
predict preferential stabilization of zwitterionic structures.
Finally, we note that the HF, MP2, and DFT models yield
somewhat different barriers (Table 3).
(ii) As far as the reaction on the PSC surface is concerned,
desolvation of Cl- may be an essential step for the reaction.
Bianco and Hynes studied the ClONO2 + HCl reaction
involving nine water molecules at the MP2/(SBK+*,6-31G+G*)//
HF(HW*,3-21G) level,43 where the nucleophile Cl- has a first
solvation shell. In our study, we attempted to quantify the effect
on reactivity of increasing the size of the Cl- solvation shell
from one (Figure 9a) to three (seven-water cluster, related to
Figure 12b) water molecules. For the four-water cluster (Figure
9a), the first solvation shell of Cl- contains a single water and
a low barrier is predicted for the reaction to the product-like
structure (Figure 9b). Increasing the first solvation shell to two
waters, as in clusters 4, 5, and 6 (Figures 10a, 11a, and 11b),
does not radically alter the reactivity because an essentially zero
barrier is calculated for the three-water-catalyzed reaction
(Figure 10, Table 3). For the reaction catalyzed by five water
molecules (Figure 12b) leading to the ionic products with zero
barrier, increasing the solvation shell of Cl- from one to three
waters has little effect. These findings suggest that complete
solvation of Cl-, may not be as important as the separation of
the ion pair H3O+Cl-.
The calculations herein reveal a number of important features
concerning the reaction of ClONO2 with HCl in small water
clusters. Importantly, we observed a change from a molecular
mechanism for the gas-phase reaction to an ionic mechanism
for the PSC-catalyzed reaction.
For the smaller systems studied, containing one to three
solvating waters, the reactions involve ClONO2 and un-ionized
HCl. These model reactions indicate that the gas-phase or
homogeneous reaction of ClONO2 with HCl may also be
important in increasing active chlorine reservoirs. The calculated
barriers [Table 3, B3LYP/6-311++G(d,p)] show that the
involvement of catalytic waters reduces the barrier from 41.8
(uncatalyzed) to 1.1 kcal mol-1 on the inclusion of only two
water molecules, and further increasing the size of the water
cluster to three water molecules leads to an essentially spontane-
ous reaction. These findings are consistent with an ab initio
study by Xu and Zhao55 in which a facile reaction is also
predicted. Although at present there are no available experi-
mental data concerning the rate of the reaction for comparison,
the prompt appearance of Cl2 (180 K) reported by Horn et al.17
suggests a ready reaction. The proposed increase in the
electrophilicity of the Cl atom of ClONO216,17,52 resulting from
ionization along the Cl1-O2 bond is evident from our reactant
(iii) Experimental findings indicate that for the heteroge-
neously catalyzed ClONO2 + HCl reaction, the final reaction
product involve the ionized acid H3O+NO3- 6,13-22 We identified
.
two such clusters, one involving four waters where the ionized
nitric acid forms a contact ion pair (Figure 9b) and the other,
in which the ion pair is effectively separated by two solvation
shells (Figure 12b). For a three-water cluster (Figure 10a), which

Electronic Structure Calculations of Activation of ClONO2
J. Phys. Chem. A, Vol. 104, No. 17, 2000 4043
initially contains H3O+Cl-, the product structure involves Cl2/
HONO2 (Figure 10c). Analysis of the transition structure (Figure
10b) for this reaction reveals insufficient solvation of the
forming nitrate, and thus additional surface-bound waters may
be required to participate in the stabilization of the developing
nitrate on the PSC surface.
(iv) Finally we comment on the energetics of the PSC-
catalyzed reaction. An ab initio investigation by Bianco and
Hynes43 involving a nine-water cluster yielded a barrier of 6.4
kcal mol-1. In agreement with this, our calculations indicate
the reaction on the PSC surface may actually proceed spontane-
ously, particularly if the reaction involves the solvent-separated
ion pair (H3O+Cl-) and proceeds as reaction 3.
In summary, our calculations have shown that the homoge-
neous or gas-phase reaction between ClONO2 and un-ionized
HCl, when catalyzed by only two waters, can occur readily at
stratospherically relevant temperatures. This reaction may be
an important contributor to the increase in active chlorine in
the stratosphere. In larger clusters related to PSC ice aerosols,
the reaction between ClONO2 and the ionized acid (H3O+Cl-)
can proceed essentially spontaneously, whereby solvent separa-
tion of acid, leads to the reaction proceeding as in eq 3. Here
the formation of an ionic product requires the participation of
more water molecules than are needed to promote the reaction.
However, our central finding of atmospheric importance is that
the reaction of ClONO2 with HCl, a key element in the proposed
formation of the ozone hole, proceeds essentially spontaneously
via an ionic pathway in water clusters of a relatively small
critical size.
(21) Tolbert, M. A.; Rossi, M. J.; Malhotra, R.; Golden, D. M. Science
1987, 238, 1258.
(22) Hanson, D. R.; Ravishankara, A. R. J. Phys. Chem. 1992, 96, 2682.
(23) Hanson, D. R.; Ravishankara, A. R. J. Phys. Chem. 1994, 98, 5728.
(24) Hanson, D. R. J. Phys. Chem. A 1998, 102, 4794.
(25) Zhang, R.; Jayne, J. T.; Molina, M. J. J. Phys. Chem. 1994, 98,
867.
(26) Zhang, R.; Leu, M.-T.; Keyser, L. J. Phys. Chem. 1994, 98, 13 563.
(27) Oppliger, R.; Allanic, A.; Rossi, M. J. J. Phys. Chem. A 1997, 101,
1903.
(28) Barone, S. B.; Zondlo, M. A.; Tolbert, M. A. J. Phys. Chem. A
1997, 101, 8643.
(29) Zondlo, M. A.; Barone, S. B.; Tolbert, M. A. J. Phys. Chem. A
1998, 102, 5735.
(30) Wincel, H.; Mereand, E.; Castleman, Jr., A. W. J. Phys. Chem. A
1997, 101, 8248.
(31) Nelson, C. M.; Okumura, M. J. Phys. Chem. 1992, 96, 6112.
(32) Van Doren, J. M.; Viggiano, A. A.; Morris, R. A. J. Am. Chem.
Soc. 1994, 116, 6957.
(33) Schindler, T.; Berg, C.; Niedner-Schatteburg, G.; Bondybey, V. E.
J. Chem. Phys. 1996, 104, 3998.
(34) La. Manna, G. J. Mol. Struct. (THEOCHEM) 1994, 309, 31.
(35) Wofsy, S. C.; Molina, M. J.; Salawitch, R. J.; Fox, L. E.; McElroy,
M. B. J. Geophys. Res. 1988, 93, 2442.
(36) Gertner, B. J.; Hynes, J. T. Faraday Discuss. 1998, 110, 301.
(37) Rieley, H.; Aslin, H. D. J. Chem. Soc., Faraday Trans. 1995, 91,
2349.
(38) Horn, A. B.; Chesters, M. A.; McCoustra, M. R. S., Sodeau, J. R.
J. Chem. Soc., Faraday Trans. 1992, 88, 1077.
(39) Delzeit, L.; Powell, K.; Uras, N.; Devlin, J. P. J. Phys. Chem. B
1997, 101, 2327.
(40) Graham, J. D.; Roberts, J. T. J. Phys. Chem. 1994, 98, 5974.
(41) Delzeit, L.; Rowland, B.; Devlin, J. P. J. Phys. Chem. 1993, 97,
10 312.
(42) George, S. M.; Livingston, F. E. Surf. ReV. Lett. 1994, 4, 771.
(43) Bianco, R.; Hynes, J. T. J. Phys. Chem. A 1999, 103, 3797.
(44) Haas, B.-M.; Crellin, K. C.; Kuwata, K. T.; Okumura, M. J. Phys.
Chem. 1994, 98, 6740.
(45) Lee, T. J. J. Phys. Chem. 1995, 99, 1943.
Acknowledgment. We thank EPSRC for support of this
research and Dr. J. C. Whitehead for helpful discussion.
(46) Ying, L. M.; Zhao, X. S. J. Phys. Chem A 1997, 101, 6807.
(47) Seeley, J. V.; Miller, T. M.; Viggiano, A. A. J. Chem. Phys. 1996,
105, 2127.
(48) Lee, T. J.; Rice, J. E. J. Phys. Chem. 1993, 97, 6637.
(49) Bianco, R.; Hynes, J. T. J. Phys. Chem. A 1998, 102, 309.
(50) Xu, S. C.; Zhao, X. S. J. Phys. Chem. A 1999, 103, 2100.
(51) Xu, S. C.; Zhao, X. S. Acta Phys.-Chim. Sinica 1998, 14, 988.
(52) McNamara, J. P.; Hillier, I. H. J. Phys. Chem. A 1999, 103, 7310.
(53) McNamara, J. P.; Tresadern, G.; Hillier, I. H. Chem. Phys. Lett.
1999, 310, 265.
(54) Mebel, A. M.; Morokuma, K. J. Phys. Chem. A 1996, 100, 2985.
(55) Xu, S. C.; Zhao, X. S. Acta Phys.-Chim. Sinica 1998, 14, 5.
(56) Akhmatskaya, E. V.; Apps, C. J.; Hillier, I. H.; Masters, A. J.;
Palmer, I. J.; Watt, N. E.; Vincent, M. A.; Whitehead, J. C. J. Chem. Soc.,
Faraday Trans. 1997, 93, 2775.
(57) Beichert, P.; Schrems, O. J. Phys. Chem. A 1998, 102, 10 540.
(58) Geiger, F. M.; Hicks, J. M.; de Dios, A. C. J. Phys. Chem. A 1998,
102, 1514.
(59) Packer, M. J.; Clary, D. C. J. Phys. Chem. 1995, 99, 14 323.
(60) Robinson Brown, A.; Doren, D. J. J. Phys. Chem. B 1997, 101,
6308.
(61) Koput, J.; Peterson, K. A. Chem. Phys. Lett. 1998, 283, 139.
(62) Hanway, D.; Tao, F.-M. Chem. Phys. Lett. 1998, 285, 459.
(63) Estrin, D. A.; Kohanoff, J.; Laria, D. H.; Weht, R. O. Chem. Phys.
Lett. 1997, 280, 280.
(64) Vincent, M. A.; Palmer, I. J.; Hillier, I. H.; Akhmatskaya, E. V. J.
Am. Chem. Soc. 1998, 120, 3431.
(65) Smith, A.; Vincent, M. A.; Hillier, I. H. J. Phys. Chem. A 1999,
103, 1132.
(66) Planas, M.; Lee, C.; Novoa, J. J. J. Phys. Chem. 1996, 100, 16495.
(67) Lee, C.; Sosa, C.; Planas, M.; Novoa, J. J. J. Chem. Phys. 1996,
104, 7081
References and Notes
(1) Molina, M. J.; Rowland, F. S. Nature 1974, 249, 810.
(2) Tolbert, M. A. Science 1994, 264, 527.
(3) Solomon, S.; Garcia, R. R.; Rowland, F. S.; Wuebbles, D. J. Nature
1986, 321, 755.
(4) Solomon, S. ReV. Geophys. 1988, 26, 131.
(5) Solomon, S. Nature 1990, 347, 347.
(6) Molina, M. J.; Tso, T.-L.; Molina, L. T.; Wang, F. C.-Y. Science
1987, 238, 1253.
(7) Cicerone, R. J. Science 1987, 237, 35.
(8) Turco, R. P.; Toon, O. B.; Hamill, P. J. Geophys. Res. 1989, 94,
16 493.
(9) McElroy, M. B.; Salawitch, R. J.; Wofsy, S. C. Geophys. Res. Lett.
1986, 13, 1296.
(10) Henderson, G. S.; Evans, W. F. J.; McConnell, J. C. J. Geophys.
Res. 1990, 95, 1899.
(11) Wennberg, P. O.; Cohen, R. C.; Stimpfle, R. M.; Koplow, J. P.;
Anderson, J. G.; Salawitch, R. J.; Fahey, D. W.; Woodbridge, E. L.; Keim,
E. R.; Gao, R. S.; Webster, C. R.; May, R. D.; Toohey, D. W.; Avallone,
L. M.; Proffitt, M. H.; Loewenstein, M.; Podolske, J. R.; Chan, K. R.; Wofsy,
S. C. Science 1994, 266, 398.
(12) Crutzen, P. J.; Arnold, F. Nature 1986, 324, 651.
(13) Banham, S. F.; Horn, A. B.; Koch, T. G.; Sodeau, J. R. Faraday
Discuss. 1995, 100, 321.
(14) Sodeau, J. R.; Horn, A. B.; Banham, S. F.; Koch, T. G. J. Phys.
Chem. 1995, 99, 6258.
(15) Koch, T. G.; Banham, S. F.; Sodeau, J. R.; Horn, A. B.; McCoustra,
M. R. S.; Chesters, M. A. J. Geophys. Res. 1997, 102, 1513.
(16) Horn, A. B.; Sodeau, J. R.; Roddis, T. B.; Williams, N. A. J. Chem.
Soc., Faraday Trans. 1998, 94, 1721.
(17) Horn, A. B.; Sodeau, J. R.; Roddis, T. B.; Williams, N. A. J. Phys.
Chem. A 1998, 102, 6107.
(18) Hanson, D. R.; Ravishankara, A. R. J. Geophys. Res. 1991, 96,
5081.
(19) Leu, M.-T. Geophys. Res. Lett. 1988, 15, 17.
(20) Chu, L. T.; Leu, M.-T.; Keyser, L. F. J. Phys. Chem. 1993, 97,
12 798.
(68) Robertson, S. H.; Clary, D. C. Faraday Discuss. 1995, 100, 309.
(69) Ying, L. M.; Zhao, X. S. J. Phys. Chem. A 1997, 101, 3569.
(70) Ando, K.; Hynes, J. T. J. Phys. Chem. B 1997, 101, 10 464.
(71) Xu, S. C.; Zhao, X. S. Acta Phys.-Chim. Sinica 1999, 15, 193.
(72) Xu, S. C. J. Chem. Phys. 1999, 111, 2242.
(73) Lee, C.; Vanderbilt, D.; Laasonen, K.; Car, R.; Parrinello, M. Phys.
ReV. B 1993, 47, 4863.
(74) Bernal, J. D.; Fowler, R. H. J. Chem. Phys. 1933, 8, 515.
(75) Davidson, E. R.; Morokuma, K. J. Chem. Phys. 1984, 81, 3741.
(76) Materer, N.; Starke, U.; Barbieri, A.; Van Hove, M. A.; Somorjai,
G. A.; Kroes, G.-J.; Minot, C. J. Phys. Chem. 1995, 99, 6267.

4044 J. Phys. Chem. A, Vol. 104, No. 17, 2000
McNamara et al.
(77) Silva, S. C.; Devlin, J. P. J. Phys. Chem. 1994, 98, 10 847.
(78) Devlin, J. P.; Buch, V. J. Phys. Chem. 1995, 99, 16 534.
(79) Devlin, J. P.; Buch, V. J. Phys. Chem. B 1997, 101, 6095.
(80) Rowland, B.; Kadagathur, N. S.; Devlin, J. P.; Buch, V.; Feldman,
T.; Wojcik, M. J. J. Phys. Chem. 1995, 99, 8328.
(81) Buch, V.; Delzeit, L.; Blackledge, C.; Devlin, J. P. J. Phys. Chem.
1996, 100, 3732.
(82) Casassa, S.; Ugliengo, P.; Pisani, C. J. Chem. Phys. 1997, 106,
8030.
(83) Morrison, I.; Li, J.-C.; Jenkins, S.; Xantheas, S. S.; Payne, M. C.
J. Phys. Chem. B 1997, 101, 6146.
(84) Moore Plummer, P. L. J. Phys. Chem. B 1997, 101, 6247.
(85) Moore Plummer, P. L. J. Phys. Chem. B 1997, 101, 6251.
(86) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Petersson,
G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez G.; Pople, J. A. Gaussian94, revision E.1; Gaussian,
Inc.; Pittsburgh, PA, 1995.
(87) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K.
N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, K.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J.
V.; Sefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.;
Replogle E. S.; Pople, J. A. Gaussian98, revision A.1; Gaussian, Inc.;
Pittsburgh, PA, 1998.
(88) Lee, C.; Yang. W.; Parr, R. G. Phys. ReV. B 1988, 37, 785
(89) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett.
1989, 157, 200.
(90) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(91) Møller, C.; Plesset, M. S. Phys. ReV. 1934, 46, 618.