NO2 quantum yield from the 248 nm photodissociation of peroxynitric acid (HO2NO2)

Article abstract of DOI:10.1021/jp001982x

Investigations of the atmospheric impacts of aircraft emissions, industrial, and agricultural emission, and changes in the radiative balance of the atmosphere invariably focus on the chemical mechanisms, which control the production and loss of O₃. The quantum yield of NO₂ production from peroxynitric acid (PNA) photodissociation at 248 nm was determined by comparison to HNO₃photolysis data taken under identical experimental conditions. Measurements made over a range of pressures, flows, and precursor concentrations resulted in an NO₂ yield of 0.56 ± 0.17. Calculations of potential energy curves for several low-lying singlet and triplet states of PNA showed that while the singlet excitations occurred via an n-π* transition on the NO₂ moiety, the dissociative channels forming OH + NO₃ and HO₂ + NO₂ occurred via predissociation on different surfaces. Excitation energies at the multireference internally contracted configuration interaction and CCSD(T) showed that excited states of PNA were not accessible at wavelengths longer than 407 nm.

Full text of DOI:10.1021/jp001982x

1592
J. Phys. Chem. A 2001, 105, 1592-1598
NO2 Quantum Yield from the 248 nm Photodissociation of Peroxynitric Acid (HO2NO2)†
Coleen M. Roehl,‡ Troy L. Mazely,§ Randall R. Friedl, Yumin Li, Joseph S. Francisco, and
Stanley P. Sander*
Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak GroVe DriVe,
Pasadena, California 91109
ReceiVed: May 31, 2000; In Final Form: October 31, 2000
Peroxynitric acid (PNA) was photolyzed at 248 nm, and the NO2 photoproduct was detected by laser-induced
fluorescence (LIF). The quantum yield for the production of NO2 was determined by comparison with HNO3
photolysis data taken under identical experimental conditions. Measurements made over a range of pressures,
flows, and precursor concentrations resulted in an NO2 quantum yield of 0.56 ( 0.17, where the statistical
uncertainty is 2 standard deviations. Calculations of potential energy curves for several low-lying singlet and
triplet states of PNA are presented. The calculations show that while the singlet excitations occur via an
n-π* transition on the NO2 moiety, the dissociative channels forming OH + NO3 and HO2 + NO2 likely
occur via predissocation on different surfaces. Excitation energies at the MRCI and CCSD(T) level of theory
show that excited states of PNA are not accessible at wavelengths longer than 407 nm (∼3.0 eV).
Introduction
thermodynamic data5):
Investigations of the atmospheric impacts of aircraft emis-
sions,1,2 industrial and agricultural emissions,3 and changes in
the radiative balance of the atmosphere4 invariably focus on
the chemical mechanisms which control the production and loss
of ozone. These mechanisms are strongly influenced by species
which couple the members of different chemical families. An
example is peroxynitric acid, HO2NO2, which couples odd
hydrogen (HOx) and odd nitrogen (NOx) in the upper troposphere
and lower stratosphere. PNA is formed by the recombination
reaction
HO2NO2 + hV f HO2 + NO2 λT ) 1230 ( 120 nm (1a)
f OH + NO3 λT ) 725 ( 60 nm
(1b)
where the threshold wavelengths are derived from thermody-
namic data at 298 K.5 The atmospheric modeling community
uses values of 0.67 and 0.33 for the quantum yields for channels
1a and 1b, respectively, over the entire ultraviolet spectrum of
PNA. Other photodissociation channels and their thresholds are
HO2NO2 + hV f O(3P) + HNO3 λT ) 718 nm
f H + NO2 + O2 λT ) 394 nm
(1c)
(1d)
HO2 + NO2 + M f HO2NO2 + M
and removed mainly by reaction with OH, photolysis, and
thermal decomposition. PNA plays a key role in the destruction
of HOx radicals by the catalytic cycle
f OH + NO2 + O(3P) λT ) 320 nm (1e)
f HO2 + NO + O(3P) λT ) 296 nm (1f)
HO2 + NO2 + M f HO2NO2 + M
OH + HO2NO2 f NO2 + H2O + O2
Although there have been a number of measurements of PNA
cross sections in the 200-350 nm spectral region,6-10 relatively
little work has been carried out on photodissociation quantum
yields. The only study is that of MacLeod et al., who measured
the quantum yield for OH formation from PNA photolysis at
248 nm.11 In the latter study, the OH product yield from PNA
photolysis was measured relative to that from H2O2 in a very
low pressure photolysis reactor/mass spectrometer system. A
quantum yield of 34 ( 16% was determined for the production
of OH radicals, implying a quantum yield of roughly 66% for
the NO2 + HO2 product channel. Difficulty in measuring
absolute, unambiguous PNA concentrations by mass spectrom-
etry complicated their results.
The weak HO2-NO2 bond in PNA results in a photodisso-
ciation threshold which lies in the near-infrared and can be
broken by a process mediated by overtones and combination
bands involving the O-H stretching mode.12 This mechanism
has been invoked to explain the observed dependence of OH
and HO2 at high solar zenith angles in the lower stratosphere.13
While cross sections of the second and third overtone bands
Net:
OH + HO2 f H2O + O2
PNA also acts as a temporary reservoir for HOx and NOx
radicals.
Atmospheric models currently consider two PNA photodis-
sociation pathways (thresholds are derived from tabulated
† Part of the special issue “Harold Johnston Festschrift”.
* Also with the Division of Geological and Planetary Sciences, and
Division of Engineering and Applied Science, California Institute of
Technology, Pasadena, CA 91125. To whom correspondence should be
addressed. E-mail: ssander@jpl.nasa.gov. FAX: (818) 393-5019. Ad-
dress: M/S 183-901, Jet Propulsion Laboratory, 4800 Oak Grove Dr.,
Pasadena, CA 91109.
‡ Present address: Division of Geological and Planetary Sciences,
California Institute of Technology, Pasadena, CA 91125.
§ Present address: Science Applications International Corporation, 21151
Western Ave, Torrance, CA 90501.
Department of Chemistry and Department of Earth and Atmospheric
Sciences, Purdue University, West Lafayette, IN 47907-1393.
10.1021/jp001982x CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/22/2000

Photodissociation of HO2NO2
J. Phys. Chem. A, Vol. 105, No. 9, 2001 1593
Figure 1. Diagram of the experimental apparatus. Details of the optical configuration of the detection axis are presented in the text.
(3ν1 and 4ν1) have been measured recently in solution14 and
gas phase,15 there are as yet no photodissociation quantum yield
measurements for these or other infrared bands.
photolysis cell was adjusted by varying the total flow rates
between 500 and 1400 sccm and the photolysis cell pressure,
which was maintained at 3.0, 7.0, or 10.0 Torr. These conditions
ensured that the NO2 photoproduct was removed from the
detector viewing zone between excimer pulses.
In this study, PNA was photolyzed at 248 nm and NO2
photoproduct was detected by laser-induced fluorescence (LIF).
PNA vapor concentrations were calculated from on-line ultra-
violet (UV) absorption measurements. The NO2 quantum yield
from PNA was determined by comparison to HNO3 photolysis
data taken under identical experimental conditions. The quantum
yield was investigated as a function of total flow rate and total
pressure over a range of precursor concentrations. Similar
photolysis experiments with HNO3 and PAN (peroxyacetyl
nitrate) have been performed in this lab.16-18 The experimental
results are complemented by calculations of the ground and
excited state potential surfaces which help to rationalize the
measured quantum yield.
An excimer laser operated at 248 nm with an energy density
of 75 mJ/cm2 and at a repetition rate of 30 Hz was used to
photolyze the PNA and HNO3 precursors. The weakly focused
excimer beam entered and exited the photolysis cell through
quartz windows located perpendicular to flow of gas and
attached to the body of the photolysis cell by O-ring joints. By
separating the windows from the cell body with O-ring joints
and adding blackened baffles in the extended sidearms of the
photolysis cell, the fluorescence background induced by the
excimer laser was reduced. A portion of pure carrier gas flow
was directed through the sidearms to keep the dead volume
flushed during measurements. Typically 16 000 laser shots were
averaged for one experiment (approximately 10 min).
Experimental Section
The 511 nm line of a 10 kHz pulsed copper vapor (probe)
laser with ∼2 W of power was utilized for LIF detection of
NO2. Light emitted from the copper vapor laser was passed
through two dichroic filters, to eliminate the 578 nm line of
the copper vapor laser, and telescoping optics, to reduce the
beam diameter to about 0.5 cm, before entering the photolysis
cell. The high repetition frequency of the probe laser allowed
for the acquisition of LIF data points every 100 µs.
Fluorescence from the excited NO2 was detected by a cooled
photomultiplier tube (PMT; Burle C31034-02) that was oriented
perpendicular to the laser beam axes. Collection optics served
to focus the light onto the cathode of the PMT. Blackened
baffles lining the housing tube leading to the PMT and a quartz
window (rather than Pyrex) were utilized to reduce undesirable
background fluorescence from the excimer laser and scattered
light. A 3.8 cm long quartz cell, filled with 0.36 M K2Cr2O7
and containing an aluminum honeycomb baffle, were used to
filter out scattered light below about 530 nm. Finally, a 578
nm notch filter selectively eliminated any residual Cu vapor
emission at that wavelength. The LIF signal was amplified,
discriminated, and counted with a multichannel scalar card
configured with 10 µs bin widths, and stored on a PC.
The apparatus used in this study has been thoroughly
described in previous publications;16-18 hence only a brief
description along with the slight modifications and particular
experimental conditions will be discussed here. A diagram of
the apparatus is shown in Figure 1. Gaseous PNA or HNO3
was introduced into a flowing system by passing a calibrated
Ar flow through either a PNA or HNO3 sample, which was
contained in a quartz reservoir placed in an ice bath. Ar was
chosen as the carrier gas because of its relatively slow diffusivity
and low fluorescence quenching efficiency. The pressure in the
reservoirs (30-100 Torr) was adjusted using a Teflon needle
valve located between the reservoirs and a quartz photolysis
cell and was monitored with a 100 Torr MKS Baratron
capacitance manometer. Typical Ar flows through the reservoirs
were 30-100 standard cm3 min-1 (sccm). Also located upstream
of the needle value were two bypass cells (5 and 30 ( 0.1 cm
in length)used during UV absorption measurements. The precur-
sor gases were further diluted with Ar in a mixing volume
downstream of the needle valve and before entering the
photolysis cell. The partial pressure of the precursors was
maintained low enough so that the fluorescence quenching was
dominated by the Ar carrier gas. The residence time in the

1594 J. Phys. Chem. A, Vol. 105, No. 9, 2001
Roehl et al.
Nitric acid was prepared by collecting the vacuum distillate
of a 50:50 by volume mixture of 95% H2SO4 with NaNO3. The
HNO3 was maintained at 0 °C by placing the sample in an ice
bath. Peroxynitric acid was synthesized by the reaction of H2O2
(>90%) with NO2BF4 using the procedure described by Kenley
et al.19 90% H2O2 was prepared by a vacuum distillation of
70% Semiconductor Grade H2O2 (FMC Corporation), and gravi-
metric techniques were used to verify the percent composition.
of sample age, and new PNA samples were produced when the
measured PNA vapor pressure dropped by 50%. Since H2O2
photolysis at 248 nm has no effect on the NO2 signal, the
primary effect of H2O2 impurity is to introduce a small error in
the determination of [PNA].
After further dilution, the concentration ranges of photolyte
in the PNA and HNO3 photolysis experiments were (1-43) ×
1013 cm3 and (2-23) × 1014 cm3, respectively.
The PNA synthesis method of Kenley et al. produces several
impurities that must be removed and/or quantified by careful
analysis before use in photodissociation experiments. The spectra
of a number of different PNA samples were obtained off-line
from the photodissociation experiment using mid-infrared
Fourier transform infrared (FTIR) and near-infrared long-path
absorption/diode array spectrometer techniques. PNA funda-
mental vibrational bands (V2-V7) as well as bands from HNO3,
SiF4, H2O2, H2O, and NO2 impurities were identified in the mid-
IR spectra, and HNO3, H2O2, H2O overtone bands, and NO2
electronic bands were identified in the near-IR spectra.15
Compositions of the samples were determined using literature
values of cross sections and band strengths for both spectral
regions.15,20,21 The dependence of the impurity levels as a
function of the PNA sample age has been discussed by Li et
al.22 and Zhang et al.,15 who found that both NO2 and HNO3
decreased with time due to fractional distillation. In the present
work it was found that the ratio [PNA]/[HNO3] was ∼10-33
while the ratio [PNA]/[H2O2] ranged from ∼0.5 to 5. The NO2
band centered near 1600 cm-1 was observed in spectra of freshly
made samples, but decreased rapidly after several minutes of
bubbling. No attempt was made to quantify this impurity. The
SiF4 and H2O impurities were not quantified since they do not
absorb in the 210-250 nm spectral region and hence could not
interfere with the on-line UV absorption measurements (dis-
cussed below), nor do they interfere with the quantum yield
determinations.
Computational Methods
The ground state of HO2NO2 was optimized at the CCSD(T)
level of theory24 with the cc-pVDZ basis set.25 At the optimized
geometry of the ground state, the vertical excitation energies
for the lowest two singlet excited states, 21A and 31A, and the
lowest two triplet excited states, 13A and 23A, are calculated at
the complete active space self-consistent field (CASSCF)26,27
and multireference internally contracted configuration interaction
(MRCI)28,29 level of theory with the cc-pVDZ basis set. The
potential energy curves for the 11A, 21A, and 31A states are
calcualated at the CASSCF level of theory with the cc-pVDZ
basis set along the RO-O′′, RO′′-N, and RN-O′ coordinates,
separately, while fixing other parameters at their equilibrium
values of the ground state (see Figure 4 for coordinate
nomenclature). Some calculations were also carried out for the
potentials along the RO-O′′ coordinate using the aug-cc-pVDZ
basis set which inclues diffuse functions. Inclusion of the diffuse
functions did not change the resulting potential surfaces
significantly.
The active space used for the CASSCF calculation in this
work is (16e,11mo), which includes eight doubly occupied
molecular orbitals and three virtual molecular orbitals. The size
of the CAS in this active space is 9075 CSFs (configuration
state functions) for the singlet states and 13 068 CSFs for the
triplet states.
Concentrations of PNA and HNO3 in the photolysis cell were
determined immediately before and/or after every photolysis
experiment by passing the gas flows through parallel 5 and 30
cm long UV absorption cells upstream of the photolysis cell.
UV absorption measurements were made at two wavelengths,
214 nm (Zn lamp, 5 cm cell) and 254 nm (Hg Pen Ray lamp,
30 cm cell), and then were used along with calibrated flow rates
and the Beer-Lambert law to calculate vapor concentrations
in the photolysis cell. Vapor concentrations obtained at the two
wavelengths were compared and averaged. In the HNO3
photolysis experiments, vapor pressures inferred from the
photometric measurements were always within several percent
of each other and typically in the range of 13.5-14.5 Torr, in
excellent agreement with known vapor pressures for pure HNO3
at 0 °C.23 In the PNA photolysis experiments, the photometric
measurements provided a way to measure the fractional impurity
of HNO3. This is possible because there is a significant
Results
Temporal fluorescence data following the photolysis of PNA
and HNO3 were collected over a range of concentrations,
pressures, and flow rates in 10 µs bins. Background noise from
cell and filter fluorescence induced by the excimer laser, which
decays more rapidly than the LIF signal, was easily deduced
from data collected between probe pulses and was subtracted
from these raw data. The residual LIF signal consisted of (1)
NO2 generated from PNA or HNO3 photolysis, (2) NO2 from
PNA or HNO3 decomposition, and (3) probe laser scatter
through the optical and chemical filters. These temporal profiles
were collected until the NO2 photoproduct completely left the
detection viewing zone, at which time the LIF signal was
constant and solely resulted from processes 2 and 3 above. The
fluorescence signal from the PNA photofragment was calculated
by subtracting this constant background signal. Typical curves
generated from the photolysis of PNA for 3.0, 7.0, and 10.0
Torr total pressure are illustrated in parts a, b, and c, respectively,
of Figure 2. Similar plots were obtained for HNO3 photolysis
in experiments performed immediately before or after the PNA
measurements.
difference in the ratio σλPNA/σHλ NO at λ ) 254 and 214 nm (the
3
values are 18 and 4.1, respectively). It was found that the vapor
pressure of the PNA samples varied over the range 0.3-1.0
Torr depending on the age of the sample. Differences in the
PNA concentrations obtained at the two wavelengths were
usually less than 10%, and samples with concentrations differing
by more than 20% (implying a significant HNO3 impurity) were
discarded. Unlike HNO3, the H2O2 impurity in the PNA sample
could not be checked in the on-line UV absorption cells because
The temporal fluorescence signals such as those shown in
Figure 2 are proportional to the NO2 concentrations present at
a particular time after the photolysis pulse and can be related
to the NO2 quantum yield (from PNA or HNO3 photolysis) by
the following equation:
the ratio σλPNA/σHλ O does not differ significantly between 254
2
and 214 nm (5.42 and 5.8, respectively). The infrared spectra
[NO2]i ) ΛφNi O (λphoto)[i]σi(λphoto
)
(2)
2
provided a means for estimating the H2O2 impurity as a function

Photodissociation of HO2NO2
J. Phys. Chem. A, Vol. 105, No. 9, 2001 1595
Figure 3. Plot of the ratios of NO2 fluorescence from back-to-back
248 nm photolysis experiments using HNO3 and HO2NO2 as a function
of the corresponding relative concentrations of each precursor. The
circles, squares, and triangles indicate data obtained at 3.0, 7.0, and
10.0 Torr total pressure, respectively. The r2 value for the least-squares
fit is 0.85.
The recommended cross sections,5 σHNO (248 nm) ) 2.00 ×
3
10-20 cm2 molecule-1 and σPNA(248 nm) ) 4.58 × 10-19 cm2
molecule-1, were used to obtain φPNNOA2 . To determine the
relative production of NO2 from each precursor, the time-
dependent fluorescence signals from both HNO3 and PNA were
ratioed and averaged at set delays between the photolysis and
probe pulses. This ratio of independent values gives the relative
Figure 2. NO2 LIF signals (arbitrary units) from the photolysis of
PNA. Fluorescence from background NO2, excimer laser-induced cell
fluorescence, and probe laser scatter have been subtracted from the
temporal profiles. For the three experiments shown, the concentrations
of PNA and the Ar buffer gas pressures were (a) 3.4 × 1013 molecules
cm-3 and 3.0 Torr, (b) 9.5 × 1013 molecules cm-3 and 7.0 Torr, and
(c) 1.9 × 1014 molecules cm-3 and 10.0 Torr. The signal rise observed
in (a) and (b) is due to collisional energy transfer from high vibrational
levels of the X(2A1) state of NO2.16
yield of NO2, SPNA(τ)/SHNO (τ), for eq 4. This procedure was
3
repeated for all sets of data. The results of the measurements
made at total flows g700 sccm are plotted versus the ratio of
the precursor concentrations in Figure 3. Results for flows <700
sccm were not used as discussed in the next section. Each data
point represents a temporally averaged ratio of LIF signals. The
vertical error bars were determined by the statistical error in
the signal ratio, while the horizontal error bars were dominated
by the uncertainty in the PNA concentration measurement. The
line represents the linear regression of these data points and
where Λ is a detector response function, φNi O (λphoto) is the
2
NO2 quantum yield of the precursor, i, which has an initial
concentration, [i], and an optical cross section, σi(λphoto) at the
photolysis wavelength, λphoto. To avoid the difficulty in having
to directly determine Λ, HNO3 is used as a calibrant. By
measuring the NO2 LIF signals from back-to-back experiments
with PNA and HNO3 precursors, the detector response function
is eliminated, i.e.
has a slope ) SPNA(τ)/SHNO (τ) × [HNO3]/[PNA] ) 12.3 ( 1.0.
3
Substituting this value and the PNA and HNO3 concentrations
calculated from the UV absorption measurements into eq 4
results in a quantum yield of 0.56 for the production of NO2
from PNA photolysis at 248 nm. The results of all data sets are
tabulated in Table 1.
S
(PNA)(τ)
[NO2]PNA
[NO2]HNO
)
)
S(HNO )(τ)
3
3
φNPNOA2 (λphoto
)
σPNA(λphoto)
[PNA]
Discussion
(3)
NO2
[HNO ]
σ
HNO (λphoto)
3
φ
(λphoto
)
3
HNO3
As stated above, the quantum yield of NO2 production from
PNA photolysis was determined by solving eq 4 using LIF data
collected over a variety of PNA and HNO3 concentrations at
total flow rates greater than 700 sccm and at total pressures of
3.0, 7.0, and 10.0 Torr. The data shown in Figure 3 are
independent of total pressure and linear over the concentration
ranges employed. This is in contrast to LIF data collected at
similar pressures and flows of 500 sccm, which exhibited a
marked pressure dependence (see Table 1). This dependence is
attributed to decomposition of PNA on the walls as results of
the longer residence times.
Si(τ) is the fluorescence signal from each of the precursors
at time, τ, after photolysis. Since HNO3 photodissociates at 248
nm to produce NO2 product with a quantum yield of unity,17
eq 3 can be rewritten as
σ
HNO (λphoto)
S(PNA)(τ)
[HNO3]
[PNA]
3
φNPNOA2 (λphoto) ) φNO2 (λphoto
)
HNO3
S(HNO )(τ)
σ
PNA(λphoto
)
3
(4)

1596 J. Phys. Chem. A, Vol. 105, No. 9, 2001
Roehl et al.
TABLE 1: Measurement Conditions and Results for 248 nm PNA Photolysis Experiments
pressure, Torr
3.0
flow rate, sccm
no. of measurements
slope
[PNA], 1013 molecules cm-3
[HNO3], 1014 molecules cm-3
500
700-900
500
700-1400
500
700-900
13
8
17
10
5
14.4
11.4
8.3
16.6
5.3
2.6-31
1.4-12
2.3-41
7.7-29
4.6-14
14-43
4.9-36
1.8-23
7.7-81
4.2-23
6.9-21
5.9-17
7.0
10.0
7
9.2
Evidence of diffusion, fluorescence quenching, and collisional
deactivation of the internally excited nascent NO2 photoproduct
is seen in the LIF profiles (Figure 2) at short time scales. Similar
profiles were obtained in the earlier studies with HNO3 and
PAN.16-18 At total pressures of 3.0 Torr, the vibrational
relaxation of NO2 was not completed before the first LIF data
was recorded and hence there is an observable initial growth in
the LIF signal. This growth is much less pronounced in the 7.0
Torr data and is not detectable by 10.0 Torr. Fluorescence
quenching is more efficient at 10.0 Torr total pressure, however,
so the maximum signal for comparable PNA or HNO3 concen-
trations is less than that at lower pressures. For measurements
conducted at constant flow rates, but varying total pressures
(such as those in Figure 2), the effects of diffusion out of the
viewing zone can be detected. The fluorescence decay at 3.0
Torr is more rapid than at 10.0 Torr due to the faster diffusion
and decreased residence time in the photolysis cell.
Figure 4. Optimized structure of HO2NO2. Detailed geometrical
parameters are given in Table 4.
TABLE 2: Electronic Configurations of HO2NO2 Singlet
and Triplet States
Finally, we report a quantum yield for the production of NO2
from PNA photolysis of 0.56 ( 0.17 (2σ). The stated uncertainty
includes the uncertainty in the slope of Figure 3 (16%), the
uncertainty in the relatiVe cross sections for HNO3 and PNA at
the photolysis and analysis wavelengths (10%), and the uncer-
tainties in the on-line UV absorbance measurements for HNO3
and PNA (10% and 20%, respectively). The derived quantum
yield is somewhat lower than the value inferred from the results
of MacLeod et al.,11 but is within the stated errors and can be
considered to be in agreement considering the complexity of
the reaction mechanism invoked to interpret the measurements
of MacLeod et al.11 The most difficult part of the earlier study
was associated with characterizing the purity and amounts of
PNA using mass spectrometric techniques. This problem was
circumvented, in part, in this study by measuring the vapor
concentrations in situ via UV absorption and by incorporating
the results from off-line FTIR and the near-IR experiment.
Another consideration in this regard is that the current work
and the MacLeod et al. study may be influenced by dissociation
of hot photofragments. In this study, NO2 can be formed from
the decomposition of excited NO3 (channel 1e) and removed
by decomposition of excited NO2 (channel 1f). The measurement
of OH in the MacLeod et al. work may be influenced by the
decomposition of excited HO2 (channel 1e). These effects may
be responsible for the failure of the observed OH and NO2
quantum yields to sum to unity at 248 nm. Both experiments,
however, measure the quantities that are important for the
purposes of atmospheric modeling.
state
electronic configuration
(core)10(1-17)a218a219a220a221a022a023a0
20a f 21a
19a f 21a
18a f 21a
20a f 21a
11A
21A
31A
13A
23A
TABLE 3: Vertical Excitation Energies (VEEs), in eV
CASSCF
MRCI
state
VEE(cc-pVDZ)
VEE(cc-pVDZ)
11A
21A
31A
13A
23A
0.0
0.0
4.78
5.41
4.03
4.57
5.05
5.52
4.26
4.76
of the five heavy atoms. 18a is a nonbonding molecular orbital
with contributions from the 2p orbitals of the two oxygen atoms
of NO2 and lies approximately perpendicular to the molecular
plane. The 19a and 20a molecular orbitals are the nonbonding
molecular orbitals consisting mainly of the 2p orbitals of the
two oxygen atoms in the NO2 group. These orbitals lie in the
heavy atom plane. The 21a orbital is the π-type antibonding
MO formed from the 2p orbitals of the two oxygen atoms and
the nitrogen atom of the NO2 group, and is perpendicular to
the heavy atom plane. The excited states, 21A, 31A, 13A, and
23A, are derived from the single electronic transitions 20a f
21a, 19a f 21a, 18a f 21a, and 20a f 21a, separately. The
excited states studied in this work can be described as n f π*
transitions on the NO2 group.
As shown in Table 3, the vertical excitation energy for the
21A state is 5.05 eV (246 nm). This is somewhat smaller than
the value of 5.823 eV (5.610 eV with electron correlation)
obtained by Saxon and Liu using the CI singlet method with
the DZP basis set.30 Recently, Chen and Hamilton calculated
the vertical excitation energies of the lowest singlet state to be
5.041 and 4.903 eV with CAS and CASPT2 wave functions,
respectively, using the B3LYP/6-31G* geometry.31 Our results
are consistent with both the CAS and CASPT2 results of Chen
and Hamilton. The present results are also consistent with the
Calculations of Potential Energy Surfaces. The optimized
geometry of the HO2NO2 ground state is shown in Figure 4.
The optimized results of calculations at the CCSD(T)/cc-pVDZ
level of theory indicate that the ground state of HO2NO2 has
no symmetry and that all the heavy atoms lie approximately in
one plane. The electronic configurations for the singlet and
triplet states in this study are summarized in Table 2. The
configuration of the ground state is
(core)10(1-17)a218a219a220a221a022a023a0
The core consists of molecular orbitals formed by the 1s orbitals

Photodissociation of HO2NO2
J. Phys. Chem. A, Vol. 105, No. 9, 2001 1597
TABLE 4: Optimized Geometries for the Ground and
Lowest Triplet State of HO2NO2 (Bond Lengths in Å, Angles
in deg, Energy in Hartrees)
CCSD(T)/cc-pVDZ
parameters
ground state
lowest triplet state
NO
NO′
1.200
1.200
1.292
1.294
NO′′
OO′′
1.518
1.418
1.515
1.427
OH
0.975
133.6
109.9
108.1
101.6
-12.2
84.2
0.976
108.7
108.9
104.4
100.7
-53.7
82.2
ONO
O′′NO
OON
HOO
OO′′NO
HOON
energy
-355.19609
-355.08444
RN-O′ than with RO-O′′ and RN-O′′. This is because the relevant
electronic transition, n f π*, is located primarily on the NO2
moiety. Also, the crossings between the 21A and 31A states and
the repulsive states occur at larger values of internuclear
separation on the N-O′ coordinate than on the O-O′′ and
N-O′′ coordinates. As noted above, inclusion of diffuse
functions in the basis set did not significantly affect the
potentials along the O-O′′ coordinate.
The calculated potential energy curves in Figure 5 provide a
way to rationalize the observed PNA photolysis quantum yields.
At 248 nm, HO2NO2 dissociation through the NO2 moiety is
unfavorable because there is no curve crossing in the Franck-
Condon region along the RN-O′ coordinate for the 21A excited
state. There is curve crossing in the Franck-Condon region for
the weakly bound 21A state along the RO-O′′ coordinate, and
any leakage could predissociate along the RO-O′′ coordinate to
produce OH + NO3. There is curve crossing in the Franck-
Condon region for the 21A state along the RN-O′′ coordinate
which leads to the formation of HO2 + NO2 by predissociation.
Comparing Figure 5a and Figure 5b, we can see that photodis-
sociation yields along the RN-O′′ coordinate should be greater
than along the RO-O′′ coordinate because the N-O′′ surface is
shallower and curve crossing occurs at smaller internuclear
separations. This is consistent with the observed quantum yields
of 0.34 for OH formation11 and 0.56 for NO2 formation at
248 nm.
Figure 5. Potential energy curves of HO2NO2 in the Franck-Condon
region along the O-O′′ (a), N-O′′ (b), and N-O′ (c) coordinates.
Energies are in hartrees and coordinate distances are in angstroms.
The vertical excited state energies of the triplet states are
summarized in Table 3. The two lowest triplet states lie below
the lowest excited singlet state. The energy of the lowest triplet
state is estimated to be 4.26 eV. The next triplet state lies about
0.5 eV above the 13A state. The triplet excited states all have
predominantly valence character. Both the triplet states of HO2-
NO2 involve promotion into an orbital with antibonding
character in the NO2 chromophore, as is the case for the
corresponding second and third singlet excited states. To obtain
a better understanding of the band origin of the lowest triplet
state, the geometry of this state was optimized at the CCSD-
(T)/cc-pVDZ level of theory. The results for the optimized
ground state and lowest triplet state are presented in Table 4.
These results show that there is significant internal rotation that
occurs around the NO bond (HOO′′-NO2). There is also a
significant lengthening of the NO bonds on the NO2 group.
These changes are consistent with the reduction in bond order
associated with the n f π* transition for this state. A recent ab
initio study of HO2NO2 calculated at the MP2(Fc)/6-311++G**
presence of a small feature in the absorption spectrum at 250
( 10 nm (4.97 eV).9
The potential energy curves calculated for the three singlet
states along the RO-O′′, RN-O′′, and RN-O′ coordinates are shown
in Figure 5a-c. As can be seen in Figure 5a, the lowest singlet
excited states, 21A and 31A, are weakly bound along the RO-O′′
coordinate in the Franck-Condon region (defined as the region
within 0.25 Å of the equilibrium point in the ground state). Their
potential energy curves indicate, however, that there are
crossings which could lead to predissociation arising from
intersections with other unbound states. Along the RN-O′′
coordinate, the potential energy curve for the 21A state is
shallower than the curve for the RO-O′′ coordinate in the
Franck-Condon region and intersects with other unbound states
leading to predissociation. The 31A state is bound and also
intersects with other unbound states around the Franck-Condon
region. Referring to the 21A and 31A potentials, the states are
more strongly bound in the Franck-Condon region along the
RN-O′ coordinate than along the RO-O′′ and RN-O′′ coordinates.
The slopes of the potential curves also vary more rapidly with
32
level by Jitariu and Hirst shows similar trends in structural
changes in the NO2 group upon excitation into the triplet state.
The adiabatic singlet-triplet (S0-T1) splitting calculated in this

1598 J. Phys. Chem. A, Vol. 105, No. 9, 2001
Roehl et al.
(11) MacLeod, H.; Smith, G. P.; Golden, D. M. J. Geophys. Res. 1988,
93, 3813-3823.
study is 3.04 eV at the CCSD(T)/cc-pVDZ level of theory with
zero-point vibrational energy differences included. This is
identical to the value of 3.04 eV calculated by Jitariu and Hirst
using the G2MP2 method.32 These results imply that the origin
of the singlet-triplet absorption starts around 407 nm.
(12) Donaldson, D. J.; Frost, G. J.; Rosenlof, K. H.; Tuck, A. F.; Vaida,
V. Geophys. Res. Lett. 1997, 24, 2651-2654.
(13) Wennberg, P. O.; Salawitch, R. J.; Donaldson, D. J.; Hanisco, T.
F.; Lanzendorf, E. J.; Perkins, K. K.; Lloyd, S. A.; Vaida, V.; Gao, R. S.;
Hintsa, E. J.; Cohen, R. C.; Swartz, W. H.; Kusterer, T. L.; Anderson, D.
E. Geophys. Res. Lett. 1999, 26, 1373-1376.
(14) Fono, L.; Donaldson, D. J.; Proos, R. J.; Henry, B. R. Chem. Phys.
Lett. 1999, 311, 131-138.
Acknowledgment. We acknowledge the technical support
of Mr. Dave Natzic. The supercomputer used in this investiga-
tion was provided by funding from the NASA Offices of Earth
Science, Aeronautics and Space Science. The research was
carried out by the Jet Propulsion Laboratory, California Institute
of Technology, under contract with the National Aeronautics
and Space Administration. Support from the NASA Upper
Atmosphere Research, Tropospheric Chemistry and Atmo-
spheric Effects of Aircraft Programs is acknowledged.
(15) Zhang, H.; Roehl, C. M.; Sander, S. P.; Wennberg, P. O. J. Geophys.
Res. 2000, 105, 14593-14598.
(16) Mazely, T. L.; Friedl, R. R.; Sander, S. P. J. Chem. Phys. 1994,
100, 8040-8046.
(17) Mazely, T. L.; Friedl, R. R.; Sander, S. P. J. Phys. Chem. 1995,
99, 8162-8169.
(18) Mazely, T. L.; Friedl, R. R.; Sander, S. P. J. Phys. Chem. 1997,
101, 7090-7097.
(19) Kenley, R. A.; Trevor, P. L.; Lan, B. Y. J. Am. Chem. Soc. 1981,
103, 2203.
(20) May, R. D.; Friedl, R. R. J. Quant. Spectrosc. Radiat. Transfer
1993, 50, 257-266.
References and Notes
(21) Hojer, S.; May, R. D.; Miller, C. E. J. Quant. Spectrosc. Radiat.
Transfer 1996, 55, 273-278.
(1) Johnston, H. S.; Kinnison, D. E.; Wuebbles, D. J. J. Geophys. Res.
1989, 94, 16351-16364.
(2) AViation and the Global Atmosphere; Penner, J. E., Lister, D. H.,
Eds.; Cambridge University Press: Cambridge, 1999.
(3) WMO Scientific Assessment of Ozone Depletion: 1998; World
Meteorological Organization: Geneva, 1999.
(4) IPCC Climate Change 1995; Cambridge University Press: Cam-
bridge, 1996.
(5) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling, EValuation Number 12; Jet Propulsion Laboratory: California
Institute of Technology, 1997.
(6) Cox, R. A.; Patrick, R. Int. J. Chem. Kinet. 1979, 11, 635.
(7) Morel, O.; Simonaitis, R.; Heicklen, J. Chem. Phys. Lett. 1980,
73, 38.
(8) Graham, R. A.; Winer, A. M.; Pitts, J. N., Jr. Geophys. Res. Lett.
1978, 5, 909.
(9) Molina, L. T.; Molina, M. J. J. Photochem. 1981, 15, 97.
(10) Singer, R. J.; Crowley, J. N.; Burrows, J. P.; Schneider, W.;
Moortgat, G. K. J. Photochem. Photobiol. 1989, 48, 17-32.
(22) Li, Z.; Friedl, R. R.; Moore, S. B.; Sander, S. P. J. Geophys. Res.
1996, 101, 6795-6802.
(23) Handbook of Chemistry and Physics, 71st ed.; Lide, D. R., Ed.;
CRC Press: Boca Raton, FL, 1990.
(24) Watts, J. D.; Gauss, J.; Bartlett, R. J. J. Chem. Phys. 1993, 98,
8718-8733.
(25) Dunning, T. H. J. Chem. Phys. 1989, 90, 1007-1023.
(26) Werner, H. J.; Knowles, P. J. J. Chem. Phys. 1985, 82, 5053-
5063.
(27) Knowles, P. J.; Werner, H. J. Chem. Phys. Lett. 1985, 115, 259-
267.
(28) Werner, H. J.; Knowles, P. J. J. Chem. Phys. 1988, 89, 5803-
5814.
(29) Knowles, P. J.; Werner, H. J. Chem. Phys. Lett. 1988, 145, 514-
522.
(30) Saxon, R. P.; Liu, B. J. Phys. Chem. 1985, 89, 1227-1232.
(31) Chen, Z.; Hamilton, T. P. J. Phys. Chem. 1996, 100, 15731-15734.
(32) Jitariu, L. C.; Hirst, D. M. J. Phys. Chem. A 1999, 103, 6673-
6677.