Role of OH-stretch/torsion coupling and quantum yield effects in the first OH overtone spectrum of cis-cis HOONO

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

130.102.42.98 On: Sat, 22 Nov 2014 01:33:04

THE JOURNAL OF CHEMICAL PHYSICS 122, 104311 ͑2005͒

Role of OH-stretch/torsion coupling and quantum yield effects

in the ﬁrst OH overtone spectrum of cis-cis HOONO

a͒

Anne B. McCoy

Department of Chemistry, The Ohio State University, Columbus, Ohio 43210

Juliane L. Fry

Arthur Amos Noyes Laboratory of Chemical Physics, California Institute of Technology, Pasadena,

California 91125

Joseph S. Francisco

Department of Chemistry, Purdue University, West Lafayette, Indiana 47907

b͒

Andrew K. Mollner and Mitchio Okumura

Arthur Amos Noyes Laboratory of Chemical Physics, California Institute of Technology, Pasadena,

California 91125

͑

Received 24 November 2004; accepted 21 December 2004; published online 11 March 2005͒

A joint theoretical and experimental investigation is undertaken to study the effects of OH-stretch/

HOON torsion coupling and of quantum yield on the previously reported ﬁrst overtone action

spectrum of cis-cis HOONO ͑peroxynitrous acid͒. The minimum energy path along the HOON

dihedral angle is computed at the coupled cluster singles and doubles with perturbative triples level

with correlation consistent polarized quadruple ␨ basis set, at the structure optimized using the triple

␨

basis set ͑CCSD͑T͒/cc-pVQZ//CCSD͑T͒/cc-pVTZ͒. The two-dimensional ab initio potential

energy and dipole moment surfaces for cis-cis HOONO are calculated as functions of the HOON

torsion and OH bond length about the minimum energy path at the CCSD͑T͒/cc-pVTZ and QCISD/

AUG-cc-pVTZ ͑QCISD—quadratic conﬁguration interaction with single and double excitation and

AUG-augmented with diffuse functions͒ level of theory/basis, respectively. The OH-stretch

vibration depends strongly on the torsional angle, and the torsional potential possesses a broad shelf

at ϳ90°, the cis-perp conformation. The calculated electronic energies and dipoles are ﬁt to simple

functional forms and absorption spectra in the region of the OH fundamental and ﬁrst overtone are

calculated from these surfaces. While the experimental and calculated spectra of the OH

fundamental band are in good agreement, signiﬁcant differences in the intensity patterns are

observed between the calculated absorption spectrum and the measured action spectrum in the 2␯_O_H

region. These differences are attributed to the fact that several of the experimentally accessible states

do not have sufﬁcient energy to dissociate to OH+NO and therefore are not detectable in an action

2

spectrum. Scaling of the intensities of transitions to these states, assuming D =82.0 kJ/mol, is

0

shown to produce a spectrum that is in good agreement with the measured action spectrum. Based

on this agreement, we assign two of the features in the spectrum to ⌬n=0 transitions ͑where n is the

HOON torsion quantum number͒ that are blue shifted relative to the origin band, while the large

−

1

peak near 7000 cm is assigned to a series of ⌬n= +1 transitions, with predominant contributions

from torsionally excited states with substantial cis-perp character. The direct absorption spectrum of

−

1

cis-cis HOONO ͑6300–6850 cm ͒ is recorded by cavity ringdown spectroscopy in a discharge ﬂow

−

1

cell. A single band of HOONO is observed at 6370 cm and is assigned as the origin of the ﬁrst OH

overtone of cis-cis HOONO. These results imply that the origin band is suppressed by over an order

of magnitude in the action spectrum, due to a reduced quantum yield. The striking differences

Institute of Physics. ͓DOI: 10.1063/1.1859273͔

I. INTRODUCTION

OH + NO + M → HONO + M,

2

0

Peroxynitrous acid, HOONO, is a weakly bound isomer

of nitric acid that is formed as a secondary product of the

reaction:

⌬ H ͑0͒ = − 204 kJ/mol,

͑1a͒

͑1b͒

1

a

OH + NO + M → HOONO,

2

0

a͒_A_u_t_h_o_r_t_o_w_h_o_m_c_o_r_r_e_s_p_o_n_d_e_n_c_e_s_h_o_u_l_d_b_e_a_d_d_r_e_s_s_e_d_._E_l_e_c_t_r_o_n_i_c_m_a_i_l_:

⌬ H ͑0͒ = − 82.0 kJ/mol.

1

b

mccoy@chemistry.ohio-state.edu

Author to whom correspondence should be addressed. Electronic mail:

mo@caltech.edu

b͒

Reaction ͑1a͒ is a critical radical chain termination step in

the atmosphere, as OH and NO combine to form stable

2

0

021-9606/2005/122͑10͒/104311/14/$22.50

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

122, 104311-1

1

104311-2

McCoy et al.

J. Chem. Phys. 122, 104311 ͑2005͒

vibrational predissociation. The observed spectrum was un-

−

1

usual, with a series of broad peaks from 6300 to 7200 cm .

−

1

A peak at 6365 cm was identiﬁed as the 2␯_O_Horigin, but

the spectrum was dominated by a large, partially blended

−

1

peak at 6970 cm . Nizkorodov and Wennberg tentatively

assigned the peaks based on ab initio predictions to all three

conformers of HOONO, but the lack of observable structure

did not allow a deﬁnitive assignment.

More recent studies have shown that trans-perp

HOONO has a spectrum distinct from that shown in Fig.

6

,16

1

,

and isomerizes quickly to the more stable cis-cis con-

former ͑millisecond time scale at room temperature and

6

ϳ10 Torr͒. Fry et al. therefore attribute all of the bands

reported by Nizkorodov and Wennberg to the lowest energy

1

7

conformer, cis-cis HOONO. Matthews et al. conclude that

−

1

the strong 6970 cm band is due to trans-perp HOONO;

however, this is at odds with the experimental observations

of Fry et al.

FIG. 1. Smoothed action spectrum of the ﬁrst overtone of the OH stretch

2␯_O_H͒ band in cis-cis HOONO, recorded at 10 Torr, 300 K ͑Refs. 1 and 6͒.

͑

The appearance of the action spectrum of HOONO is

unusual. A typical overtone spectrum comprises a strong ori-

gin band and weaker satellite features; the weak bands are

either combination bands ͑including Fermi Resonances͒ or

hot bands ͑including sequence bands͒. The majority of the

oscillator strength is typically carried in the local 2␯_O_Hmode.

Any assignment of the cis-cis HOONO overtone spectrum

must explain why this origin band is relatively weak.

One factor that may be important in understanding the

anomalous intensity pattern of the overtone action spectrum

is the wavelength dependence of the quantum yield. The

quantum yield is unity if the photon energy is greater than

the dissociation energy D . Konen et al. now estimate a D

nitric acid. Channel ͑1b͒ reduces the efﬁcacy of reaction ͑1a͒

as a sink of HO and NO radicals, because HOONO rapidly

x

redissociates under atmospheric conditions. HOONO has

been the focus of many studies in the past two decades, but

only recently has there been direct evidence for its existence

1

–4

and formation in the gas phase under thermal conditions.

Theoretical studies establish the stability of HOONO

and have identiﬁed several minima and transition states. The

lowest energy conformer is cis-cis HOONO, a planar ring

structure in which the OH forms a weak intramolecular hy-

drogen bond with the terminal O atom. The trans-perp con-

former, with the heavy atoms in a plane in a trans conﬁgu-

ration and the OH bond perpendicular to that plane, has been

0

−

1

18

for cis-cis HOONO of 82.0 kJ/mol ͑6860 cm ͒. All of the

bands in the action spectrum that are below the dissociation

−

1

−

1

predicted to lie 15.3 kJ/mol ͑1280 cm ͒ higher in energy at

threshold of 6860 cm

must be hot bands. Rice–

5

,6

the CCSD͑T͒/cc-pVTZ level of theory.

Ramsperger–Kassel–Marcus ͑RRKM͒ calculations by Fry et

al. indicate that the quantum yield falls off rapidly below this

A third postulated minimum corresponds to the cis-perp

geometry in which the heavy atoms are in a plane in a cis

conﬁguration and the OH bond is perpendicular to that plane.

This conformer was ﬁrst investigated by McGrath and Row-

land at the MP2/6-311+G͑3df ,2p͒ ͑MP2–Møller-Plesset

perturbation theory truncated at the second order͒ level of

6

limit. Thus, the intensity of bands at frequencies below D₀

would be suppressed.

A second feature of cis-cis HOONO that may be respon-

sible for the anomalous intensity patterns is the presence of

an intramolecular hydrogen bond, which couples the OH

stretch with low-frequency, large amplitude modes near the

equilibrium conﬁguration. The internal hydrogen bond

causes a redshift in the OH-stretch frequency in cis-cis

HOONO, with an observed fundamental frequency of

7

theory, and has since been studied by several groups at vari-

4

,8–12

ous levels of theory.

The calculations of Golden et al. at

5

the QCISD͑T͒/cc-pVDZ level ͑QCISD—quadratic conﬁgu-

ration interaction with single and double excitation͒ and

Bean et al. at the CCSD͑T͒/cc-pVTZ level demonstrate that

higher levels of theory are required for a good description of

2

1

2

−

3306 cm ; based on the expected anharmonic correction,

the origin of the cis-cis HOONO ﬁrst overtone is the

−

1

HOONO. Attempts to locate a cis-perp HOONO minimum at

6365 cm band. Torsional motion about the OO bond ͑the

HOON dihedral angle, here referred to as ␶͒ will break the

hydrogen bond and lead to a blueshift of the OH bond fre-

5

,6,13

the CCSD͑T͒ level have been unsuccessful,

calling into

question earlier predictions of a cis-perp minimum. In the

ﬁrst spectroscopic detection of HOONO, Lee and co-workers

reported matrix-isolation spectra of two of the HOONO iso-

mers, cis-cis and trans-perp HOONO, but did not observe

−

1

quency back to Ϸ3530 cm . The intensity of the OH stretch

is predicted to be weaker in the hydrogen-bonded form, by a

factor of 1.5 in the double harmonic approximation ͑QCISD/

AUG-cc-pVDZ͒. Thus, torsionally excited molecules, in

which the hydrogen bond is effectively broken, may have

absorption bands that are blueshifted and possess larger cross

sections than the cis-cis HOONO origin.

1

4,15

the cis-perp isomer.

Recently, Nizkorodov and Wennberg reported the ﬁrs₁t

direct spectroscopic detection of HOONO in the gas phase,

reproduced in Fig. 1. In this experiment, excitation of

HOONO in the near infrared at the ﬁrst OH stretch overtone

The above considerations lead to the following hypoth-

esis. The overtone action spectrum observed by Nizkorodov

͑

2␯_O_H͒ region was detected by monitoring OH products from

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

130.102.42.98 On: Sat, 22 Nov 2014 01:33:04

104311-3

Spectrum of cis-cis HOONO

J. Chem. Phys. 122, 104311 ͑2005͒

and Wennberg ͑Fig. 1͒ is that of cis-cis HOONO, but the

relative band intensities deviate signiﬁcantly from the ex-

pected absorption spectrum due to quantum yield effects.

minimize the total energy. The potential was calculated using

the coupled-cluster method with single and double excita-

tions which incorporated a perturbative estimate of the ef-

1

9–21

The origin and combination bands below D will appear

fects of connected triple excitations ͓CCSD͑T͔͒.

The

0

weak, because they have a reduced quantum yield for disso-

Dunning correlation consistent cc-pVTZ basis set was

−

1

22–24

ciation. The strong bands at 6970 cm arise from hot, tor-

sionally excited molecules for which the internal hydrogen

bond has been broken, resulting in a vibrationally averaged

frequency of the OH bond close to that of a free OH stretch.

The pure absorption spectrum will possess a strong origin

band, but will still have signiﬁcant hot bands and sequence

bands with large spectral shifts to the blue that arise from

strong coupling between the OH stretch and the low-

frequency torsional modes, particularly the HOON torsion.

The present work is a joint theoretical and experimental

effort to investigate this hypothesis. We have calculated the

absorption and action spectra of cis-cis HOONO in the fun-

damental and ﬁrst overtone region from ﬁrst principles, using

a two-dimensional model that explicitly treats the OH stretch

and HOON torsional degrees of freedom, and experimentally

measured the ﬁrst overtone spectrum of HOONO in direct

absorption.

used.

The Hessian from a quadratic conﬁguration inter-

2

5

action with single and double excitation ͑QCISD͒ method

optimization of the geometry was used for the geometry op-

timization at the CCSD͑T͒ level. The energies along the

minimum energy path were also calculated at the CCSD͑T͒/

cc-pVQZ level for the geometries previously optimized at

the CCSD͑T͒/cc-pVTZ level of theory and basis set. To dis-

tinguish between these two minimum energy paths, we de-

MEP

note them by V

͑␶͒ and V

͑␶͒, respectively. The

pVTZ

pVQZ

2

6

GAUSSIAN98 program was used in all the calculations.

B. Two-dimensional torsion-stretch potential energy

and dipole surfaces

To explore further the vibrational states of cis-cis

HOONO, two-dimensional potential and dipole moment sur-

faces were evaluated as functions of r and ␶, with the re-

maining seven coordinates constrained to their values on the

minimum energy path surface. The potential points were

evaluated at the CCSD͑T͒/cc-pVTZ level of theory/basis set,

while the dipole moment surface was evaluated at the

QCISD/AUG-cc-pVTZ level of theory/basis set. The poten-

tial and dipole surfaces were evaluated for 13 angles and 13

͑

i͒ An ab initio minimum energy path along the HOON

torsional coordinate ␶ was generated at the CCSD͑T͒/cc-

pVTZ level with the OONO dihedral angle frozen so that the

heavy atoms lie in a plane. CCSD͑T͒ energies were also

calculated with a cc-pVQZ basis set at the same nuclear ge-

ometries. This calculation provides the most accurate predic-

tion to date of the HOON torsional potential, including the

cis-perp HOONO conformer.

OH distances. The values of r ranged from r ͑␶͒−0.3 Å to

e

r ͑␶͒+0.3 Å, where r ͑␶͒ is the value of the OH bond length

e

͑

ii͒ Electronic energies and the dipole moment compo-

along the minimum energy surface at the speciﬁed value of

nents were calculated as functions of the OH stretch coordi-

nate r for ﬁxed HOON angles ␶ at the CCSD͑T͒/ccpVTZ and

QCISD/AUG-cc-pVTZ level of theory/basis, respectively.

␶

. The values of ␶ that were used to construct the two-

dimensional surface were 0°, 10°, 20°, 30°, 40°, 60°, 70°,

0°, 90°, 100°, 120°, 150°, and 180°. These points were cho-

sen to span the range of this angle ͓by symmetry V͑r,␶͒

8

͑

iii͒ Two-dimensional quantum mechanical calculations

were performed to solve for the torsion-vibration energy lev-

els and transition moments from the resulting surfaces.

=

V͑r,−␶͔͒ with a higher density of points near the potential

minimum and the cis-perp conformation.

͑

iv͒ Both absorption and action spectra at 300 K were

To correct for the small differences in the energies ob-

tained with the cc-pVQZ and cc-pVTZ basis sets along the

minimum energy path, for each value of ␶ the energies were

scaled by ␴͑␶͒, the ratio of the energy differences computed

by the two basis sets:

calculated. The action spectrum was predicted by convolut-

ing the calculated absorption spectrum with a simpliﬁed

quantum yield. The temperature dependence of this action

spectrum was investigated by calculating spectra at 233 and

3

00 K.

v͒ The absorption spectrum was recorded in the region

MEP

V

͑␶͒ − V

͑0°͒

͑

pVQZ

V͑r,␶͒ = V_p_V_T_Z͑r,␶͒

ͫ

ͬ

−

1

MEP

pVTZ

6

300–6800 cm by cavity ringdown spectroscopy and com-

V

pVTZ

pared to the experimental action spectrum and to the theoret-

ical predictions.

=

V_p_V_T_Z͑r,␶͒␴͑␶͒.

͑2͒

As ␴͑␶͒ is not deﬁned at ␶=0°, we extrapolated the scaling

from its values at 10°, 20°, and 30° and used ␴͑0°͒=0.943.

The one-dimensional slices along r through the unscaled

cc-pVTZ basis potential at each value of ␶ were ﬁt to a

Morse oscillator,

II. THEORETICAL AND COMPUTATIONAL METHODS

A. Torsional potential

In order to investigate the role of torsional motion in the

cis-cis HOONO spectrum in the region of the ﬁrst overtone,

2

U_p_V_T_Z͑r,␶͒ = D ͑␶͒†1 − exp„− ␣͑␶͓͒r − r ͑␶͔͒…‡ + U ͑␶͒,

e

we ﬁrst calculated a one-dimensional potential surface

͑

3͒

MEP

V

͑␶͒. This surface is a minimum energy path through the

HOONO surface, as a function of the HOON torsion angle ␶

evaluated at increments of 10°. The OONO dihedral angle

was ﬁxed to 0° ͑the cis conﬁguration͒ and the values of the

remaining seven internal coordinates were optimized to

in order to facilitate extrapolation to a larger range of OH

distances. In these ﬁts, all parameters were allowed to vary.

Evaluation of the full scaled potential U͑r,␶͒ was achieved

by multiplying D ͑␶͒ and U ͑␶͒ by the appropriate scaling

e

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-4

McCoy et al.

J. Chem. Phys. 122, 104311 ͑2005͒

factor ␴͑␶͒, and using a cubic spline interpolation to evaluate

two lowest frequency modes will not have enough energy to

dissociate. As such, the intensity of each transition in the OH

overtone absorption spectrum was multiplied by the fraction

of the population at temperature T that has at least 490 cm

of vibrational excitation in other modes. This fraction is the

these potential parameters at intermediate angles. The scaled

͑

s͒

versions of the parameters will be denoted as D ͑␶͒ and

e

͑

s͒

−1

U ͑␶͒ in the discussion that follows. For the spline, the grid

e

was extended to Ϫ360° through 360° by symmetry in order

to avoid unphysical edge effects.

estimated quantum yield for photodissociation, ⌽n ͑T͒,

Ј

A similar approach was employed to ﬁt the dipole mo-

ment surface. Here we ﬁrst transformed the dipole moment

operator into the principal-axis system for the four heavy

atoms. We then ﬁt each of the three Cartesian components of

which was multiplied by the absorption intensity to deter-

mine the action spectrum line intensities,

I

͑␯,T͒ = ⌽ ͑T͒ ϫ I

͑␯,T͒. ͑6͒

ABS;n v ,n v

Љ Љ Ј Ј

ACTION;n ,v ,n v

Љ Љ Ј Ј

n

Ј

the dipole moment to a cubic polynomial in ͓r−r ͑␶͔͒. The

e

form of the surface and the resulting intensities will depend

on the choice of the embedding of the body-ﬁxed axis sys-

tem. A principal-axis system was selected in which the three

principal axes were determined at each of the 13 geometries

along the minimum energy path and the axes are labeled to

ensure that the rotational constants obey AϾBϾC. Since the

transformation is based on the four heavy atoms, which do

not move signiﬁcantly over the full range of ␶, we ﬁnd that

the rotational constants vary by no more than 5%. This is

much smaller than the differences between the rotational

constants. Evaluation of the vector components of the dipole

moment surface was achieved by spline interpolation of the

ﬁt parameters.

III. EXPERIMENT

Cavity ringdown spectroscopy ͑CRDS͒ was employed to

experimentally obtain the direct absorption spectrum of cis-

cis HOONO in the overtone region. HOONO was generated

in a ﬂow tube reactor at 20–40 Torr, with a gas mixture of

1

00 SCCM 1% H /He mix ͑SCCM denotes cubic centimeter

2

per minute at STP͒ and 100 SCCM Ar ﬂowing through a

microwave discharge, combining with 200 SCCM 4%

NO /N₂͑similar to the ﬂow conditions for generating

2

6

HOONO used in Bean et al. and Fry et al. ͒. A purge ﬂow

of 200 SCCM dry N per mirror was used to prevent destruc-

2

tion of the mirror surfaces. These ﬂows result in a total com-

position in the cell of 0.125% H and 1.0% NO in bath N ,

2

Ar, and He.

C. Vibrational energy levels, transition intensities,

and spectra

At these concentrations, the primary reactions were

H + NO → OH + NO,

͑7͒

͑1a͒

͑1b͒

Having obtained a potential and dipole moment surface,

the energies and intensities were determined using a two-

dimensional discrete variable representation method,

where both the radial and angular dependences are treated on

evenly spaced grids of points with 30 and 75 points, respec-

tively. The value of ␶ was allowed to vary from Ϫ180° to

2

7,28

OH + NO + M → HONO + M,

2

OH + NO + M → HOONO + M,

2

but a key side reaction was

180°, while r ranged from 0.500 to 2.225 Å. With these

OH + NO + M → HONO + M.

Reference spectra of the two main interfering species,

͑8͒

energies and wave functions, the transition strengths of the

lines were evaluated using

2

HONO and HONO , were recorded for spectral subtraction.

I

͑␯͒ = ␯͉͗vЈnЈ͉␮͑r,␶͉͒vЉnЉ͉͘ ,

͑4͒

2

vЉ,nЉ;vЈ,nЈ

Both species were generated by bubbling N ﬂow through

2

where ␯ is the transition wave number in cm⁻¹and ␮ is the

dipole moment in D, and v and n are the quantum numbers

for OH stretch and torsion, respectively. The intensities of

lines in the absorption spectrum can then be determined by

multiplying the transition strength by a Boltzmann factor

liquid samples and ﬂowing through the CRDS cell at 40 Torr.

The HONO sample was generated by reaction of H SO with

2

4

NaNO salt in aqueous solution, and the HONO sample was

2

70% in H O, used as purchased from Aldrich.

2

Spectra were recorded by ﬂowing the relevant gas mix-

ture ͑discharge products or reference gas͒ through the center

50 cm long section of a cavity ringdown cell 2 cm in diam-

eter. The optical cavity comprised two highly reﬂective mir-

rors ͑Nova Wave, R=99.995% centered at 1.5 ␮m͒ posi-

tioned on either end of the cell, spaced 82 cm apart. The

mirrors were separated from the central ﬂow region by 16 cm

long purge volumes. Resulting empty-cell ringdown times

were on the order of 10 ␮s.

f

͑T͒ accounting for the population of the lower state in

vЉ,nЉ

the transition,

I

͑␯,T͒ = C ϫ f_v_Љ_,_n_Љ͑T͒ ϫ I

͑␯͒,

v ,n ;v ,n

Љ Љ Ј Ј

͑5͒

ABSv ,n ;v ,n

Љ Љ Ј Ј

where the zero in energy is taken to be the zero-point energy

of HOONO, and C is chosen so that the maximum value of

I

͑␯,T͒ at 300 K is 100.

ABSv ,n ;v ,n

Љ Љ Ј Ј

The above expression gives the absorption intensity for a

speciﬁc ͑vЈ,nЈ͒� ͑vЉ,nЉ͒ transition. Since the photodisso-

ciation product OH is detected in the action spectroscopy

experiments, the detection efﬁciency depends on the quan-

tum yield for dissociation. The dissociation threshold for cis-

cis HOONO lies 490 cm above the 2␯_O_Hovertone band.

Therefore, based on harmonic frequencies, the overtone and

combination bands involving one quantum of either of the

The near-infrared ͑NIR͒ radiation, tunable through the

−

1

range 6000–7000 cm , was generated as the idler beam of a

optical parametric oscillator ͑OPO͒ pumped by a 100 Hz

tripled Nd:YAG ͑355 nm͒ ͑YAG—yttrium aluminium gar-

net͒ laser. Typical NIR pulse energies were 20 mJ with a

−

1

−

1

laser linewidth of 4 cm . Ringdown traces were accumu-

lated with a GaGe Compuscope board ͑14-bit digitizer͒.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-5

Spectrum of cis-cis HOONO

J. Chem. Phys. 122, 104311 ͑2005͒

FIG. 2. Calculated torsional minimum energy potential ͑MEP͒, as a function

of ␶ with the OONO dihedral angle constrained to 0°. Both potentials were

calculated at the CCSD͑T͒ level of theory using the cc-pVTZ ͑ᮀ͒ and cc-

pVQZ ͑b͒ basis sets.

Traces from 40 laser shots were averaged before ﬁtting. The

averaged traces were ﬁt to exponential functions using the

Levenberg–Marquardt algorithm and averaged at each wave-

length as the OPO was scanned through the range at a step

FIG. 3. ͑a͒ Harmonic frequency, ␻ ͑␶͒, and ͑b͒ anharmonicity, ␻ x ͑␶͒ for

the OH stretch in cis-cis HOONO, derived from the parameters in the Morse

ﬁts of U͑r,t͒ in Eq. ͑3͒, plotted as a function of torsional angle ␶.

e

e e

−

1

size of 2 or 4 cm . Scans of the full range were repeated in

both scanning directions and at various step sizes to ensure

that the relative intensities of the peaks were not inﬂuenced

by slow changes in product concentrations due to ﬂuctua-

tions in the microwave discharge. Absorption spectra were

obtained by subtracting the empty-cell ringdown spectra to

account for background losses due to the ringdown mirrors.

calculation, and it is likely that it depends on the OONO

torsion angle, which was ﬁxed at 0° in these calculations. As

such, we do not ascribe too much signiﬁcance to the pre-

dicted depth ͑or lack͒ of the well, although, as we will see,

the plateau has important implications for the spectroscopy.

B. Two-dimensional potential energy surface U„r,␶…

and dipole moment surfaces ␮„r,␶…

IV. RESULTS

The two-dimensional potential energy surface U͑r,␶͒

was evaluated over the range r ͑␶͒−0.3 Åഛrഛr ͑␶͒

A. Ab initio one-dimensional torsional potential

e

The CCSD͑T͒/cc-pVQZ and CCSD͑T͒/cc-pVTZ tor-

+0.3 Å and 0°ഛ␶ഛ180° by ﬁtting the calculated CCSD͑T͒/

cc-pVTZ points, and introducing scaling factors ␴͑␶͒ to cor-

rect for differences between the energies calculated using the

cc-pVTZ and cc-pVQZ basis sets. This energy surface as a

function of ␶ and r is reported in Table II of the EPAPS

MEP

sional minimum energy potential ͑MEP͒ functions, V

͑␶͒

pVTZ

MEP

pVQZ

and V

͑␶͒, are plotted in Fig. 2. The energies and opti-

mized structures along ␶ are reported in Table I of the EPAPS

2

9

supplemental material for this paper. The potentials that

were obtained at these two levels of theory are nearly iden-

tical, indicating that the calculation is well converged with

respect to basis set. The present results represent the highest

level of theory applied to this system to date. Using the cc-

pVQZ basis, the barrier for hindered rotation is 14.1 kJ/mol

2

9

supplemental material for this paper. The quality of the ﬁt

was tested by comparing the values of re͑␶͒ and Ue͑␶͒ to the

values obtained from electronic structure theory. In all cases,

−

1

MEP

Ue͑␶͒ is within 12 cm of V

͑␶͒, or better than 1%. The

pVTZ

values of re͑␶͒ agree to better than 0.05% or 0.0005 Å.

The harmonic frequencies ␻ ͑␶͒ and anharmonicities

␻ x ͑␶͒ for the one-dimensional OH vibration determined

͑

s͒

−

1

͑

1180 cm ͒, and occurs at ␶=180°, the cis-trans conﬁgura-

e

͑

s͒

tion of HOONO. Near the equilibrium geometry, the poten-

tial appears to be fairly harmonic, but there is a plateau in the

potential that extends from 50° to 110° that is roughly 8.4

e

from the scaled Morse ﬁts are plotted as a function of ␶ in

Fig. 3 ͑Morse parameters are reported in Table III of the

−

1

29

kJ/mol ͑700 cm ͒ above the global minimum. On this sur-

face, there is a slight local minimum at 90°. The depth of this

well is sensitive to both the level of theory and basis set. An

extensive search for a true local minimum, varying all geo-

metric parameters, was conducted at the CCSD͑T͒/cc-pVTZ

level but failed to ﬁnd a minimum.

EPAPS supplemental material for this paper. ͒ Here we use

the superscript ͑s͒ to indicate that they were calculated using

the scaled surface. These quantities depend strongly on the

HOON torsional angle, and change most rapidly at small

angles. As ␶ is increased from 0° to 50°, the harmonic OH

−

1

stretch frequency increases by 300 cm , while the anharmo-

nicity decreases sharply. This is consistent with an intramo-

lecular hydrogen-bonding interaction between the OH group

and the terminal O atom in the planar cis-cis conﬁguration.

Such a hydrogen bond softens the OH bond and makes it

This region of the torsion potential near ␶=90° will be

referred to as the cis-perp conﬁguration of HOONO. The

shape of the torsional potential near this geometry also de-

pends on the OH distance r, which was optimized in this

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-6

McCoy et al.

J. Chem. Phys. 122, 104311 ͑2005͒

FIG. 4. One-dimensional slices of the dipole moment surface ␮͑r,␶͒ evalu-

ated at equilibrium OH bond lengths, r ͑␶͒, projected onto the three prin-

e

ciple axes, deﬁned in the text. ␮_A͑᭿͒, ␮_B͑b͒, and ␮_C͑ᮀ͒. Surface com-

puted at QCISD/AUG-cc-pVTZ level.

more anharmonic, but it is very directional and its effect on

the OH vibration weakens quickly as ␶ is changed.

The two-dimensional surfaces for the three components

of the dipole moment were determined at the QCISD/AUG-

cc-pVTZ level, at the same values of r and ␶ that were used

to obtain the two-dimensional potential surface. In order to

estimate the effects of the size of the basis set on the com-

puted value of the dipole moment, the dipole moment was

reevaluated at a subset of the conﬁgurations ͑three values of

r at ␶=0°, 20°, and 60°͒ at the QCISD/AUG-cc-pVDZ level.

For these points, the dipole moment values computed with

the two basis sets differed by less than 3%.

FIG. 5. Vibrationally adiabatic potential energy surfaces U_v_O_Hfor v_O_H=0, 1,

and 2, plotted as a function of ␶. The thin lines on each of the adiabatic

potential curves represent the energies E͑v_O_H,n͒ of the torsional states that

are associated with the speciﬁc v_O_Hquantum level.

The values of the dipole components along the minimum

energy path ͑r=r ͑␶͒͒ are plotted as a function of r in Fig. 4,

e

and the values of all of the calculated points on the three

components of the dipole surface are reported in Tables

by the quantum numbers n for the HOON torsion mode and

v for the OH stretch mode. The effect of torsion-stretch cou-

pling is most clearly seen by plotting the vibrationally adia-

batic torsion potentials Uv͑␶͒, obtained by averaging the two-

dimensional surface over a speciﬁc v state in the OH stretch.

This is achieved by evaluating

2

9

IV–VI of the EPAPS supplemental material for this paper.

The two in-plane components ͑A and B͒ are symmetric with

respect to ␶ and are relatively ﬂat over the range −50°Ͻ␶

Ͻ50°. However, the C ͑out-of-plane͒ component has a

strong dependence on ␶, and is of odd symmetry. The ﬁrst

derivative of this component of the dipole is 0.0258 D/deg at

1

2

1

2

͑

s͒

− ␻_ex^͑_e^s^͒

͑s͒

U ͑␶͒ = ␻ ͑␶͒

ͩ

v +

ͪ

͑

␶

͒

ͩ

v +

ͪ

+ U ͑␶͒, ͑9͒

v

e

0°. The differences in the properties of the three components

of the dipole moment are consistent with ⌬n= ±1 transitions

carrying signiﬁcant intensity and having C-type band con-

tours, whereas the ⌬n=0 transitions will have A- and/or

B-type band contours. The slices through the dipole moment

surface as a function of r for a given ␶ are ﬁt to a cubic

where the harmonic frequency and anharmonicity are ob-

tained from the scaled potential U͑r,␶͒, using the standard

relationships. These potentials are shown in Fig. 5. In con-

MEP

trast to V ͑␶͒, which is the bare electronic energy com-

puted along the minimum energy path, U ͑␶͒ is the one-

v

polynomial, centered at r ͑␶͒. The raw data and resulting

e

dimensional effective potential which includes the

parameters are reported in Tables IV–VII of the EPAPS

vibrational energy of the vth OH stretch state.

2

9

supplemental material for this paper. Fits to lower and

higher order polynomials were also performed and it was

found that quadratic through quartic expansions of the dipole

moment surface lead to spectra with intensities in near exact

agreement.

As illustrated in Fig. 5, the shape and depth of the adia-

batic torsional potentials depend on the amount of OH-

͑

s͒

͑s͒

stretch excitation. The ␶ dependence of ␻ and ␻ x , plot-

e

ted in Fig. 3, illustrates that the OH vibrational energy in Eq.

9͒ increases sharply with ␶. Consequently, the vibrationally

͑

MEP

pVQZ

adiabatic torsional potentials are deeper than V

͑␶͒, plot-

C. Calculated vibrational energy levels, wave

functions, and vibrationally adiabatic potentials

ted in Fig. 2, and they become deeper with increasing v.

The ﬁrst-order effect of the OH-stretch torsion coupling

is to increase the barrier height of the hindered rotation, i.e.,

the energy of cis-trans HOONO. The barrier on the scaled

The vibrational energy levels E͑v,n͒ were obtained by

solving the two-dimensional Hamiltonian, and are identiﬁed

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-7

Spectrum of cis-cis HOONO

J. Chem. Phys. 122, 104311 ͑2005͒

surface increases from 1342 to 1715 cm⁻¹and 2107 cm⁻¹in

In Fig. 6 we plot the projections of the probability den-

the v=0, 1, and 2 potentials, respectively. This can be com-

sities onto ␶ for the six lowest energy states ͑v=0͒ depicted

−

1

MEP

pared to the barrier of 1175 cm in V

vibrational dependence of U ͑␶͒ is seen at small angles, and

͑␶͒. Much of the

by horizontal lines in U ͑␶͒ of Fig. 5. The ﬁrst four of these

pVQZ

0

states can be best described as a progression of states in the

cis-cis well with n=0–3. There is a sharp increase in the

range of the potential sampled by these states and the states

with nϾ3. In fact, the n=4 state has most of its probability

amplitude in the region 45°Ͻ͉␶͉ Ͻ120° and can be best de-

scribed as having an anomalously large probability in the

cis-perp region of the potential. This apparent localization

results from the fact that as the potential well opens up near

␶=90°, the zero-order description of this motion, e.g., small

vibrations about the cis-cis HOONO minimum, is no longer

valid. As a result of this rapid increase in the range of motion

along ␶ we ﬁnd that there are two states that are close in

energy, one of which is the next state in the progression of

vibrational levels from the cis-cis HOONO minimum and

another which has greater probability amplitude near the cis-

perp shelf. This is seen in spite of the fact that the well at the

cis-perp HOONO conﬁguration is quite shallow. While we

have not plotted the states for other values of v, the general

trends are the same, and the wave functions that correspond

to the horizontal lines in Fig. 5 retain their clean nodal struc-

ture. The states that have energies lower than the cis-perp

shelf are localized in regions where ͉␶͉Ͻ60°, while the states

just above the shelf have anomalously large probability am-

plitude in the cis-perp region of the potential.

v

the curvature of the torsion potential at the cis-cis HOONO

minimum is found to increase with increasing OH stretch

excitation. This leads to an increase in the difference be-

tween the energies of the states with n=0 to n=1 from

80 cm⁻when v=0, to 462 cm and 537 cm when v=1

1

−1

3

and 2, respectively. Finally, the cis-perp HOONO minimum

in the one-dimensional surface becomes shallower with in-

creasing OH excitation. The value of the potential in the

cis-perp conﬁguration ͑␶=90°͒, measured relative to the cis-

cis HOONO minimum, e.g., U ͑90°͒−U ͑0°͒, increases

v

−

1

−1

from 848 cm

1

for v=0 to 1190 cm

for v=1 and

545 cm⁻for v=2. This can be compared to the value of the

1

potential along the minimum energy path at the cis-perp ge-

MEP

pVQZ

−1

ometry, where V

͑90°͒=690 cm . The dramatic changes

in the torsional potential and frequencies with OH-stretch

excitation reﬂect the large decrease in the OH-stretch fre-

quency in the hydrogen-bonded cis-cis conﬁguration.

On each of the one-dimensional adiabatic surfaces,

U ͑␶͒, in Fig. 5, we plotted the energy of the associated

v

two-dimensional vibrational states. Since we are solving the

Schrödinger equation for a two-dimensional system for v

Ͼ0, there will be states with multiple values of v in a given

energy range. As such, for a given v, we only showed the

energies of, at most, the 20 lowest energy states for which

D. Calculated absorption and action spectra

2

␲

m

2

P =

͵

d␶͉͗␺ ͑r͉͒⌿ ͑r,␶͉͒͘ Ͼ 0.75.

͑10͒

Our calculated spectra are shown in Fig. 7. The spectra

were generated using the two-dimensional model described

above, with a temperature of 300 K. In each of the plots we

present a stick spectrum as well as the results a convolution

of the stick spectrum by a Gaussian with a full width, half

v

m

0

Here ␺ ͑r͒ represents an eigenstate of the one-dimensional

stretch problem, obtained from the one-dimensional slice

through our two-dimensional surface at ␶=0, while ⌿ ͑r,␶͒

is a solution to the two-dimensional stretch-torsion Hamil-

tonian, described above. For these states, we replace the gen-

eral index m by the combined index v and n, where the value

of n is determined by the number of zeros in the integrand.

For v=0, the states with n=0, 1 and 2 lie below the

barrier that separates the cis-cis and cis-perp HOONO

v

m

−1

maximum of 70 cm , which is chosen to reproduce the ob-

served experimental broadening. In Table I we list some of

the most important ͑nЈ,vЈ=2͒� ͑nЉ,vЉ=0͒ overtone transi-

tions, along with their respective transition frequencies, tran-

sition strengths, absorption intensities, and action intensities,

as deﬁned in Eqs. ͑4͒–͑6͒.

minima and are conﬁned to the cis-cis well on the U ͑␶͒

0

−

1

1. Absorption spectra

adiabatic surface. The highest of these states lies 848 cm

above the minimum of the U ͑␶͒ adiabatic surface and less

The calculated absorption spectra for the fundamental

and overtone at 300 K are shown in Figs. 7͑a͒ and 7͑b͒. In

both spectral regions, the dominant peak is the origin, and

only one transition, the n=0� 0 band, contributes to its in-

tensity. All of the other bands with signiﬁcant intensity—

combination bands, sequence bands, and hot bands—are

shifted to the blue of the origin by 100–600 cm . In the

convoluted spectra, there is one very broad satellite band in

the fundamental region and four narrower satellite bands for

the ﬁrst overtone.

The overtone absorption spectrum is most relevant to the

current study. The ﬁrst two smaller bands adjacent to the

2␯_O_Horigin, B and C, are approximately 10% and 2% of the

intensity of the origin. Each has a single dominant contribu-

tion, from n=1� 1 and n=2� 2 transitions, respectively.

Thus, these levels are “Franck–Condon-like” ⌬n=0 se-

0

than 1 cm⁻below the cis-perp minimum on this surface.

There are six additional states that have energies that are at

or below the cis-trans barrier to free rotation of the OH about

the OO bond. In addition, we note that the density of states

increases sharply just above the cis-perp shelf, because the

well opens up rather abruptly at 50°. At energies above the

cis-trans barrier, the level spacings begin to increase and the

states become doubly degenerate as they approach the n²

dependence of a one-dimensional free rotor. When the OH

bond is excited to the v=1 or 2 states, the cis-cis well deep-

ens and narrows, while the plateau about the cis-perp con-

former and the cis-trans barrier height increase. As a result,

for v=2, four states have energies below the cis-perp shelf,

and these states have larger level spacings than those with

v=0.

1

−

1

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-8

McCoy et al.

J. Chem. Phys. 122, 104311 ͑2005͒

FIG. 6. Probability densities for the

six lowest energy states ͑v_O_H=0, n

=

0–5͒, projected onto ␶.

quence bands. Such large shifts in the sequence band fre-

quencies are a direct result of the stretch-torsion coupling,

which leads to the increased energy level spacing of the tor-

sion states in U ͑␶͒ versus U ͑␶͒, discussed above. These

bands are comparable in transition strength to the origin, but

appear weak due to their low populations.

Boltzmann factor. However, this decrease in population is

compensated by the factor of 50 increase of the transition

strength for the n=4� 3 transition, compared to 1� 0 tran-

sition. A fourth weak band, labeled E, is another pileup of

weak bands with a signiﬁcant contribution from ⌬n= +2

transitions, which appear as a single peak after convolution.

^T^h^e^␯OH fundamental region of the absorption spectrum

is similar to the overtone region, although the magnitudes of

the blueshifts are less, as expected. Only one satellite band,

2

0

The two higher energy bands in the overtone spectrum

are pileup bands of several transitions. The third band D is

ϳ10% the intensity of the origin. We assign this band to a

series of largely unresolved ⌬n= +1 C-type transitions. The

n=4� 3 transition is the strongest single line, although the

transitions with nЉϽ6 all carry intensities that are within a

factor of 4 of this transition. As such, the main contribution

to the overall intensity of the convoluted band comes from

the sum of the intensities of transitions with nЉ=0 through 6.

As we noted above, the states with nЉജ3 have energies

above the cis-perp plateau, and thus many of the transitions

in this band arise from states with signiﬁcant probability am-

plitude in the cis-perp HOONO geometry. As seen in Table I,

the individual ⌬n= +1 transitions have large transition

strengths for nЉജ2, a reﬂection of the fact that the intensity

in this band comes from the C component of the dipole vec-

tor ͑e.g., the component that is perpendicular to the OONO

plane͒. It is interesting to note that the states with nЉജ2 are

at least 628 cm⁻higher in energy than the vibrationless

level, so that they are expected to be weak based on the

−

1

located 333 cm to the blue of the origin, has appreciable

intensity ͑15% of the origin͒. It arises from the series of

⌬

=

n=1 transitions, described above. The nЉ=1� 1 and nЉ

2� 2 are seen in the stick spectrum with 10% and 2% of

the intensity of the most intense transition, the smaller blue-

shift of these bands leads them to be incorporated as a blue

shoulder in the envelope of the most intense transition.

The overtone absorption spectrum at 233 K is shown as

a dotted line in Fig. 7͑b͒. The two curves are normalized

such that the origin peaks 6310 cm have the same intensity.

At the lower temperature, the sequence bands and hot bands

are approximately three times weaker, due to the lower the

−

1

Boltzmann factor f

͑T͒ of the torsionally excited levels.

vЉ,nЉ

2

. Action spectrum

The primary difference between the absorption and ac-

tion spectra arises from the fact that the action spectrum

1

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-9

Spectrum of cis-cis HOONO

J. Chem. Phys. 122, 104311 ͑2005͒

1

8

HOONO͒ of Konen et al. On the other hand, at 300 K hot

bands are populated. Therefore, if we neglect off-diagonal

anharmonicities between ͑r,␶͒ and the seven remaining vi-

brational modes, we can approximate the action spectrum by

weighting the absorption intensity by the fraction of the vi-

brational partition function that reﬂects contributions to the

intensities of transitions to states with at least 490 cm⁻¹of

vibrational energy above the 2␯_O_Hovertone.

The procedure for achieving this is described above. We

use the harmonic frequencies obtained at the CCSD͑T͒/cc-

pVTZ level of theory. When we analyzed the vectors associ-

ated with the two torsional modes, which had frequencies of

−

1

3

83 and 523 cm , we found that they were highly mixed. In

fact, when we performed the harmonic analysis making ei-

ther the mass of hydrogen or the masses of the heavy atoms

very large, thereby setting one of the torsion frequencies to

zero, we found that the HOON torsion frequency became

−

1

4

39 cm , while the OONO torsion frequency became

−

1

62 cm . This indicated that there is large mixing of the two

torsional modes at the harmonic level. As such, we use

the average of the two torsion frequencies in place of the

calculated OONO frequency. Operationally the above proce-

dure requires us to scale transitions to the 2␯_O_Hstate by

⌽₀

͑300 K͒=0.104 and transitions to the combination

band with 2 quanta in the OH stretch and one in the

torsion by ⌽ ͑300 K͒=0.288. At 233 K, we ﬁnd that the

1

quantum yields have decreased to ⌽ ͑233 K͒=0.0352 and

0

⌽₁

͑233 K͒=0.1566.

The resulting stick and convoluted spectra I_A_C_T_I_O_N͑␯,T͒,

taking into account the nonunity quantum yield ⌽ ͑300 K͒,

n

Ј

is plotted in Fig. 7͑c͒. Action spectrum intensities of speciﬁc

transitions are listed in Table I.

Comparing the spectra in Figs. 7͑b͒ and 7͑c͒ reveals a

dramatic change in the relative band intensities in the over-

tone spectrum upon inclusion of the quantum yield correc-

tion in Eq. ͑6͒. The ﬁrst three bands, A, B, and C, are the

same as observed in the absorption spectrum ͑the n=0� 0,

n=1� 1, and n=2� 2 transitions͒, but the origin is almost

ten times weaker, only a factor of 2 to 3 stronger than the

next two bands. The most prominent feature in the calculated

action spectrum is now the large peak at 6900 cm⁻¹͑D͒. This

peak has roughly twice the peak intensity of the origin and is

substantially broader, indicating that it accounts for more of

the total band strength.

As in the absorption spectrum, the D band is composed

of many transitions, the majority of which are ⌬n= +1 lines.

This is a reﬂection of the fact that the intensity in this band

comes from the C component of the dipole vector ͑e.g., the

component that is perpendicular to the OONO plane͒. As is

seen in Fig. 4, its value increases from zero in the planar

conﬁguration to 1.5 D in the cis-perp HOONO conﬁguration.

The most intense of these transitions originates from the nЉ

FIG. 7. Spectra, calculated at 300 K ͑solid͒ and 233 K ͑dashed͒ for the

fundamental ͑top͒, ﬁrst overtone ͑middle͒, and the ﬁrst overtone multiplied

by the quantum yield for photodissociation given in Eq. ͑6͒ ͑bottom͒. In all

three spectra, the stick spectrum is convoluted with a Gaussian with a full

width half maximum of 70 cm⁻and all plots at 300 K are normalized so

that the maximum value of the stick and convoluted spectra in ͑a͒ are 100,

while the spectra at 233 K in ͑b͒ and ͑c͒ are normalized so that the peak

1

−

1

maxima at ϳ6300 cm have the same relative intensities.

requires that the excited HOONO dissociate to OH+NO₂.

We expect that intramolecular vibrational-energy redistribu-

tion ͑IVR͒ and dissociation will be fast and assume that any

=

3 state, with signiﬁcant contributions from states up to nЉ

7. This band is composed of hot bands primarily from

excited state above D will predissociate and contribute to

0

states that sample the cis-perp HOONO geometry. Additional

−

1

the action spectrum. Thus, in our quantum yield calculation

side bands to the blue of the 6900 cm peak, labeled E, F,

and G, arise from ⌬n= +2 transitions, originating from the

ground state ͑G͒, the state with nЉ=1 ͑F͒, and a pileup of

states with nЉϾ1 ͑E͒.

we count states that are prepared with energy above D .

0

−

1

Experimentally, the 2␯_O_Hstate lies 490 cm below the

dissociation threshold, based on the estimation of D ͑cis-cis

0

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

130.102.42.98 On: Sat, 22 Nov 2014 01:33:04

104311-10

McCoy et al.

J. Chem. Phys. 122, 104311 ͑2005͒

TABLE I. Frequencies, relative absorption and action intensities at 300 K, and absolute transition strengths for

the strongest transitions in the calculated cis-cis HOONO spectrum, discussed in the text.

I

vЉ,nЉ;vЈ,nЈ

nЉ

nЈ

␯͑cm⁻¹

͒

Band

a

I

b

I

ACTIONvЉ,nЉ;vЈ,nЈ

b

͑cm D ͒

−1

2

ABSvЉ,nЉ;vЈ,nЈ

0

1

2

3

0

1

2

3

4

0

1

2

3

1

2

3

4

5

2

6309

6467

6617

6804

6847

6866

6881

6907

6918

7246

A

B

C

1.000

0.166

0.036

0.020

0.032

0.021

0.031

0.058

0.043

0.031

0.104

0.048

0.036

0.020

0.009

0.021

0.031

0.058

0.043

0.031

0.3744

0.3839

0.2752

0.2219

0.0121

0.0491

0.2342

0.6353

0.6077

0.0115

D

G

a

As identiﬁed in Figs. 7͑b͒ and 7͑c͒.

Deﬁned in Eqs. ͑5͒ and ͑6͒. The values are normalized so that for the 0� 0 transition in the absorption, the

b

intensity is unity.

The overtone action spectrum at 233 K is shown as a

dashed line in Fig. 7͑c͒. We have again normalized the spec-

tra so that the origin peaks at 6310 cm have the same in-

tensity. In the action spectrum case, in contrast to the absorp-

tion spectrum, there is strikingly little change in the relative

band intensities with changing temperature.

40 Torr; the background empty-cell ringdown spectrum has

been subtracted. The ordinate is the absorption in units of

ppm per pass in the cavity ringdown cell ͑where the effective

pass length is estimated to be about 50 cm͒. No attempt was

made to measure an exact absorption cross section.

−

1

−

1

Reference spectra of HONO ͑centered at 6680 cm , as-

3

0

signed as the 2␯ band of cis-HONO by Guilmot et al. ͒ and

1

−

1

HONO ͑centered at 6530 cm , assigned as the 2␯ −␯

band by Feierabend et al. ͒ are shown in Fig. 8͑b͒. Above

6850 cm , the HONO2 2␯1 band dominates, and no subtrac-

tion is possible. The spectra have been scaled in intensity to

best ﬁt matching bands in the discharge spectrum.

2

1

9

E. Experimental results

31

−

1

The cavity ringdown absorption spectrum of products

from the discharge reaction H+NO is shown in Fig. 8͑a͒.

2

The spectrum was recorded at 300 K and a total pressure of

The

300–6800 cm , is shown in Fig. 8͑c͒; only one feature re-

mains after subtraction of HONO and HONO , a band at

subtracted

cavity

ringdown

spectrum,

−

1

6

2

−

1

6

370 cm . The subtracted spectrum to the blue of this band,

−

1

from 6400 to 6800 cm , is noisier, primarily due to difﬁcul-

ties with background subtraction and to mirror reﬂectivity.

No additional bands are evident in the absorption spectrum

in this region; any absorption bands must have intensities on

the order of 10% or less of the origin band. No additional

−

1

absorption was observed up to 6850 cm .

−

1

The band at 6370 cm is roughly 30–50% of the inten-

sity of the cis-HONO band, a result similar to previous ob-

servations of cis-cis HOONO and cis-HONO band intensities

2

in the fundamental region. As this is the only peak observed

and has reasonable intensity, we assign this band to the ori-

gin of the cis-cis HOONO 2␯_O_Hspectrum. This peak has no

resolvable structure but has a slightly asymmetric ﬂat-top

proﬁle that does not match either Gaussian or Lorentzian line

−

1

shapes. The bandwidth, 60 cm ͑FWHM͒, is approximately

0% wider than predicted by a rotational contour simulation

5

6

for this band.

V. COMPARISON OF CALCULATED

AND EXPERIMENTAL SPECTRA

FIG. 8. Direct absorption spectrum of cis-cis HOONO recorded by cavity

ringdown of OH+NO system, reaction path length LϷ50 cm. ͑a͒ Spectrum

2

of products formed in the discharge reaction. ͑b͒ Reference spectra of

HONO and HNO₃and scaled to match intensities in ͑a͒. ͑c͒ Spectrum of

cis-cis HOONO obtained by subtracting reference spectra from discharge

spectrum ͑curve͒, compared to the smoothed action spectrum of cis-cis

HOONO ͑Ref. 6͒ in the same spectral region.

The two-dimensional calculations of the ﬁrst overtone

spectrum of cis-cis HOONO predict that the intensity pattern

in the action spectrum will differ signiﬁcantly from the ab-

sorption spectrum, due to quantum yield effects. By detect-

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-11

Spectrum of cis-cis HOONO

J. Chem. Phys. 122, 104311 ͑2005͒

ing the CRDS spectrum of this region, we are now in a

position to test this prediction and to evaluate the extent to

which the coupling of just two modes, the OH stretch and

HOON torsion, can account for the previously observed ac-

tion spectrum.

direct absorption spectrum, within our signal-to-noise. Thus,

the intensity patterns of the experimental action and absorp-

tion spectra differ qualitatively, in accord with our predic-

tions.

Unfortunately, background absorption by the 2␯ band

OH

−

1

To compare experimental and calculated spectra, we

must shift the absolute frequency scale of the calculated ab-

of nitric acid at 6944 cm overwhelms any signal above

6850 cm and prevents us from detecting any HOONO ab-

sorption. Consequently, we cannot rule out the presence of

−

1

−

1

sorption and action spectra by 55 cm in order to match the

−

1

−1

origin in the calculated spectrum with the 6365 cm

͑

the strong 6970 cm band seen in the action spectrum. Thus,

−

1

6370 cm ͒ peaks seen in the experimental action ͑absorp-

complete veriﬁcation by experiment of the effects of quan-

tum yield on the overtone spectrum must await experiments

that can observe this band directly.

tion͒ spectra. Such a shift is well within the uncertainties of

the ab initio calculation as well as uncertainties stemming

from neglect of vibrational couplings with the seven other

modes.

VI. DISCUSSION

In the case of the action spectrum, the strongest feature

−

1

A. Limitations of calculated spectra

in the experimental spectrum is the broad band 600 cm to

the blue of the origin. A band at a similar shift and with large

relative intensity ͑D͒ is also seen in the calculated spectrum.

Examining the stick spectrum that underlies this feature, we

ﬁnd that this band has contributions from a series of overlap-

ping ⌬n= +1 hot band transitions from torsionally excited

molecules. The most intense of these originates from the nЉ

Our two-dimensional model captures much of the essen-

tial physics of this system through consideration of only the

coupling between the HOON torsion and the OH-stretch mo-

tions; however, like any reduced dimensional model, it has

limitations. Speciﬁcally, we know that HOONO has nine de-

grees of freedom, three of which have relatively low frequen-

cies. Further, just as motion in the HOON torsion will break

the intramolecular hydrogen bond in the equilibrium con-

ﬁguration, so will motion along the heavy atom torsion and

the in-plane breathing mode. Our calculations indicate that

the two torsional modes have the same symmetry and nearly

identical frequencies. The lowest torsional levels are thus

likely to be highly mixed.

=

3 state. Transitions originating from states with nЉഛ7 carry

intensities that are at least 15% of the intensity of the most

intense band in this progression.

There are two additional features to the blue of the large

1

6

970 cm⁻peak in the experimental action spectrum, at 7130

−

1

and 7240 cm . Similar bands to the blue of the large band

are seen in the calculated spectrum ͑F and G͒ and arise from

⌬

n= +2 transitions from the states with nЉ=1 and 0, respec-

We expect that there will also be signiﬁcant anharmonic

couplings between the OH-stretch vibration and the other

two low-frequency vibrational modes. Furthermore, Fermi

resonances between the OH stretch and the OOH bend are

often evident in overtone spectra. As a result, there are fea-

tures in the spectrum that cannot be accounted for in this

model. We note that parallel to these studies Stanton and

tively. The intensity of the G band was observed by Fry et

6

al. to increase slightly with decreasing temperature, in

agreement with the change seen in the calculated spectra

from 300 to 233 K. The remaining features that are seen in

both the observed and calculated action spectra are the peaks

−

1

at 6500 and 6640 cm . These are assigned to the n=1� 1

1

3

͑

B͒ and n=2� 2 ͑C͒ sequence bands, respectively.

co-workers have been using perturbation theory to calcu-

While many of the peaks in the experimental spectrum

late frequencies and intensities of transitions in the 2␯ re-

OH

−

1

are reproduced in the calculated spectrum, there are experi-

mental features that are not found in the calculation. Of the

major bands, only the band at 6730 cm⁻in Fig. 1 is not

reproduced by these reduced dimensional calculations. This

band is most likely a combination or sequence band involv-

ing one of the seven vibrational modes that are not included

in this model. Based on the position of this band and the

harmonic frequencies of cis-cis HOONO, this band is most

likely a combination band that involves another mode.

In contrast to the action spectrum, the calculated absorp-

tion spectrum consists of a single strong band at the origin,

with the other bands possessing intensities Ͻ5% of the ori-

gin. The experimental CRDS absorption spectrum, shown in

gion in the spectrum. The 6730 cm band that is not as-

signed using this two-dimensional model appears to be

accounted for as a different combination band in their model.

In addition to missing seven vibrational degrees of free-

dom, the calculated relative intensities of the action spectrum

1

are also very sensitive to the choice of D . The experiment of

0

1

8

Konen et al. pinpoints D with high precision, but since the

0

dissociation limit falls in the middle of the overtone spec-

trum, even small shifts in D or in the predicted vibrational

0

frequencies will alter the calculated intensity patterns. Our

treatment of the quantum yield is also simple; we consider

only which vibrational levels are above or below dissocia-

tion. We neglect effects due to dissociation lifetimes and col-

lisional relaxation or excitation, processes which lead to

pressure dependent quantum yields that may be important

very close to threshold. Finally, we neglect rotations, which

couple to the torsion, but also can lead to centrifugal barriers

and hence inﬂuence dissociation yields.

−

1

Fig. 8͑c͒, conﬁrms these predictions, up to 6850 cm . In

Fig. 8͑c͒, the experimental CRDS absorption spectrum is

compared to the previously reported action spectrum. As ex-

pected, the absorption spectrum has only one strong band, at

−

1

6

370 cm . This band is nearly identical to the band ͑A͒ in

the action spectrum at essentially the same frequency and is

assigned as the origin. None of the features present in the

action spectrum at 6505, 6640, or 6730 cm are seen in the

In spite of the above comments it is quite remarkable

how well a two-dimensional model based entirely on

ab initio energies and frequencies, with the dissociation en-

−

1

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-12

McCoy et al.

J. Chem. Phys. 122, 104311 ͑2005͒

ergy as the only experimentally constrained parameter, repro-

duces nearly all of the major features in the recorded action

spectrum.

As seen in Fig. 7͑c͒, the calculated action spectra at 233

K and 300 K show no signiﬁcant change in relative band

intensities with decreasing temperature, demonstrating that

−

1

18

the value of D =6860 cm obtained by Konen et al. is

0

6

consistent with the results of Fry et al. The surprising simi-

B. Interpretation of the action spectrum

larity of the two plots in Fig. 7͑c͒ can be explained by the

observation that the action spectrum is largely due to hot

bands of cis-cis HOONO. As the temperature is reduced

from 300 to 233 K, transitions originating out of the more

populous states, nЉ=0 or 1, have smaller quantum yields,

while transitions originating out of higher energy torsionally

excited states nЉϾ1, decrease in intensity as their popula-

tions decrease. The reductions in quantum yields for the

lower frequency nЉ=0 or 1 transitions nearly exactly match

the decrease in the higher frequency bands excited from tor-

sionally hot states. Thus, the overall intensity of the action

spectrum decreases almost identically across all bands, and

the relative intensities of most of the bands remain un-

changed.

This work clariﬁes the physical basis for the experimen-

tally observed ﬁrst overtone action spectrum of cis-cis

1

,6

HOONO. We ﬁnd that most of the peaks in the observed

action spectrum ͑Fig. 1͒ are hot bands, i.e., are transitions

from initial states that are vibrationally excited. The ﬁrst

three prominent peaks are ⌬n=0 sequence bands. Of these,

the n=0� 0 and n=1� 1 must arise from hot cis-cis

HOONO molecules with vibrational excitation in modes

other than the OH stretch or the HOON torsion, in order to

result in a ﬁnal state that is above the dissociation threshold.

While the photon energy of the broad D band exceeds the

−

1

dissociation energy of 6860 cm , this band is also composed

of transitions from torsionally excited molecules, primarily

There are two notable exceptions to the consistency in

band intensity ratios with changing temperature. The F and

G bands in the calculated spectrum are predicted to increase

in intensity when the temperature is decreased from 300 to

⌬

n= +1 from higher nЉ states.

In previous work, the strength of the 6970 cm band

−

1

was puzzling. We ﬁnd that several effects contribute to the

intensity of this band. First, the intensity of this band does

not result from borrowing intensity from the zeroth order

2

33 K. The increase in the intensity of these transitions in the

calculated spectrum reﬂect the fact that they are assigned to

the ⌬n= +2 transition from the v=1 and 0 levels and are

transitions that have unit quantum yield for dissociation,

within our model. The experimental data support this, as the

2

␯_O_Hstate, as is often seen in overtone spectra, e.g., in Fermi

resonance states. This is evidenced by the clean nodal pat-

terns in the vibrational wave functions associated with the

states that are indicated by the horizontal lines in Fig. 5. As

is seen in Table I, the ⌬n= +1 transitions from higher level

torsional states, e.g., n=4� 3 and n=5� 4 actually have

larger transition strengths than the origin band. The origin of

this enhancement is due in part to the large derivative of the

C component of dipole moment with ␶. In addition, compar-

ing the states on the adiabatic surfaces with zero and two

quanta in the OH stretch, plotted in Fig. 5, we ﬁnd that the

states that are just above the cis-perp shelf differ in the num-

ber of quanta in the torsion by one. This reﬂects the deepen-

ing of the cis-cis well with increasing OH-stretch excitation.

This leads to an increase in the magnitude of the matrix

elements of the dipole moment operator, and from that the

intensities of these ⌬n= +1 transitions. Second, there is an

unusually high density of states starting at nЉ=2, which

stems from the opening of the potential at the cis-perp shelf.

Thus, while we do not ﬁnd any states that are truly bound in

a cis-perp well, the probability density of the states just

above this shelf indicates that the vibrationally averaged

structure of these molecules is effectively the cis-perp geom-

etry. These are the states that predominantly contribute to the

D band.

−

1

band at 7230 cm in the experimental action spectrum also

increased in intensity with decreasing temperature from 273

6

to 193 K. The unassigned band in the experimental action

spectrum also showed an increase in intensity with decreas-

ing temperature, consistent with its assignment as a band

from the ground state to a combination level of 2␯_O_Hwith

excitation in at least one of the seven vibrations not included

in the present model.

C. Implications for the HOONO yield

in the OH+NO reaction

2

Bean et al. determined the HOONO/HONO yield in

2

reaction ͑1͒ by measuring the integrated band intensities of

the respective fundamental modes, and converting these to

densities using the ratio of CCSD͑T͒/AUG-cc-pVTZ cross

sections computed in the double harmonic approximation for

2

cis-cis HOONO. Thus, they made the assumption that all

molecules of both species were absorbing in the ␯_O_Hband.

This is in fact true for HONO ͑the main sequence band is

2

within the envelope of the fundamental͒. However, the cur-

rent calculations ͓Fig. 7͑a͔͒ demonstrate that the strong

stretch-torsion coupling shifts all of the HOON torsional hot

The current results also resolve a puzzling discrepancy.

6

Fry et al. observed no change in the major features in the

−

1

experimental action spectrum as they lowered the tempera-

ture. However, they expected that bands with photon ener-

bands ͑sequence or otherwise͒ more than 80 cm away from

the origin of the fundamental band. Thus, integration of the

absorption at the origin band excludes states with nЉജ1.

Based on the harmonic oscillator partition function, the mea-

surement given by Bean et al. must underestimate the cis-cis

HOONO population by Ϸ20%.

gies below the dissociation limit D would appear weaker at

0

lower temperatures. Using statistical RRKM calculations,

they predicted that the quantum yield would fall off rapidly

for h␯ϽD . The origin band intensity was estimated to de-

0

crease by over a factor of 2 relative to the D band, which lies

above D₀.

These considerations lead us to the tentative conclusion

that the low pressure HOONO/HONO₂branching ratio

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

104311-13

Spectrum of cis-cis HOONO

J. Chem. Phys. 122, 104311 ͑2005͒

should be signiﬁcantly higher than the 7.5% measured by

Bean et al. ͑300 K and 13 Torr͒. To quantify fully this

effect, more accurate integrated intensities, taking into ac-

count anharmonicity of the OH stretch, must be computed

sional excitation up to n=6 contribute to the spectrum, indi-

cating that perturbation theory is unlikely to describe com-

pletely the action spectrum.

2

The cis-perp geometry plays an important role in the

spectroscopy of cis-cis HOONO. A pileup of torsionally ex-

cited ⌬n= +1 transitions in a cis-perp-like geometry, with an

effectively free OH stretch and with large intrinsic intensi-

ties, accounts for the strongest band near 7000 cm⁻¹in the

action spectrum. This is surprising in light of the high energy

of the torsionally excited ground states, but is explained in

this study as resulting from a combination of effects: a pileup

of multiple states at the cis-perp geometry, the anomalously

large intrinsic intensity of the OH stretch in the cis-perp

geometry, and the suppression of the origin band which

would otherwise dominate the spectrum.

for both HOONO and HONO . Furthermore, all of the other

2

modes must be included. Coupling with other modes, nota-

bly the low-frequency bending/breathing mode ␯₇and

OONO torsion ␯ will alter the frequency shifts. However, if

8

these modes also lead to large shifts in their respective hot

bands, then the ratio could be even larger.

D. Mechanism for broadening of the overtone

spectrum

The current results shed additional light on the spectral

broadening observed in the overtone spectrum of cis-cis

HOONO. Our CRDS measurement demonstrates that the ori-

gin band of the ﬁrst overtone measured in direct absorption

is only slightly narrower than the same band detected by

The calculations are highly sensitive to the dissociation

energy and are in excellent agreement with the experimental

18

D determined by Konen et al. This resolves the apparent

0

inconsistency between this value of D and the ﬁnding of Fry

6

0

action spectroscopy. Fry et al. noted that the width of the

6

et al. that the relative intensities of bands in the 2␯ region

OH

2

6

␯_O_Hovertone origin band in the action spectrum ͑at

of the action spectrum appear to be independent of tempera-

ture. As the temperature is lowered, we predict that torsional

hot band intensities in the absorption spectrum decrease sig-

niﬁcantly. In the action spectrum, this decrease in torsional

hot band intensities is nearly exactly matched by a drop in

the quantum yield for HOONO dissociation of the bands at

lower photon energies, which require additional energy in

other modes to dissociate.

−

1

365 cm ͒ is Ϸ50% broader than a simulated rotational

−

1

contour. The laser linewidth in both cases was ϳ4 cm .

Thus, the width of the 2␯_O_Hovertone origin band at room

temperature is intrinsic to absorption in the overtone band

and is not a consequence of using action spectroscopy.

Such broadening may have many causes in an action

spectrum: dissociation lifetime broadening, inhomogeneous

broadening ͑especially for bands below D , which are exclu-

0

The quantity of most relevance in atmospheric chemistry

sively hot bands͒, or IVR. Since we now ﬁnd that the line-

width is similar in the direct absorption spectrum, we con-

clude that the 2␯_O_Horigin band of cis-cis HOONO is

broadened predominantly by IVR processes.

is the branching ratio of HOONO relative to HONO in re-

2

action ͑1͒. This work suggests an upward adjustment of that

2

branching ratio previously reported by Bean et al. The re-

duced dimensionality of this model does not allow a new

quantitative estimate of the branching ratio, but does suggest

that further work is required to correctly account for cou-

pling to other modes and to assess the integrated cross sec-

tion of cis-cis HOONO in the fundamental region.

In contrast, for the trans-perp HOONO isomer, the ac-

tion spectrum of the overtone is ﬁt well by a simulated rota-

6

tional contour. The difference in widths of the cis-cis

HOONO versus trans-perp HOONO overtone bands indi-

cates that IVR is faster in cis-cis HOONO. IVR may be

facilitated in cis-cis HOONO by the presence of the intramo-

lecular hydrogen bond and the strong stretch-torsion cou-

pling. In trans-perp HOONO, the OH bond is out of plane

and less likely to relax quickly.

ACKNOWLEDGMENTS

This work was supported by the California Air Re-

sources Board ͑Contract No. 03-333͒, the National Science

Foundation ͑NSF Grant Nos. CHE-0200968 and ATM-

VII. CONCLUSIONS

0

432377͒, and the National Aeronautics and Space Adminis-

tration Upper Atmospheric Research Program ͑NASA Grant

No. NGT-11657͒. Experiments were performed in the labo-

ratory of Professor Paul O. Wennberg, and the authors grate-

fully acknowledge his support and interest. J.L.F. and

A.K.M. acknowledge support of NSF Graduate Research

Fellowships and a NASA Earth System Science Fellowship.

The authors thank the NASA Jet Propulsion Laboratory Su-

percomputing Project for computer time. They thank Stanley

P. Sander, John F. Stanton, Amit Sinha, and Marsha I. Lester

for helpful discussions and for sharing unpublished results

prior to publication.

We report the results of a theoretical and experimental

investigation of the dynamics that lead to the observed inten-

sity pattern in the OH ﬁrst overtone action spectrum of cis-

cis HOONO. We ﬁnd that coupling of the OH stretch to the

HOON torsion and quantum yield effects are primarily re-

sponsible for the features seen in the action spectrum.

Theoretical calculated absorption and action spectra, and

experimental measurement of the direct absorption spectrum

of cis-cis HOONO, conﬁrm that the 2␯_O_Horigin band lies

below D and is therefore suppressed in the action spectrum.

0

We are able to account semiquantitatively for most of the

features in the action spectrum and their relative intensities,

despite the reduced dimensionality of the calculation. Only

1

S. A. Nizkorodov and P. O. Wennberg, J. Phys. Chem. A 106, 855 ͑2002͒.

2

B. D. Bean, A. K. Mollner, S. A. Nizkorodov, G. Nair, M. Okumura, S. P.

−

1

one major band, at 6730 cm , is not predicted and likely

involves other vibrations not treated here. Levels with tor-

Sander, K. A. Peterson, and J. S. Francisco, J. Phys. Chem. A 107, 6974

͑2003͒.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

04311-14

McCoy et al.

J. Chem. Phys. 122, 104311 ͑2005͒

F. Stanton, J. Chem. Phys. ͑to be published͒.

9

3

H. Hippler, S. Nasterlack, and F. Striebel, Phys. Chem. Chem. Phys. 4,

959 ͑2002͒.

1

2

G. Purvis and R. Bartlett, J. Chem. Phys. 76, 1910 ͑1982͒.

K. Raghavachari, G. W. Trucks, J. A. Pople, and M. Head-Gordon, Chem.

Phys. Lett. 157, 479 ͑1989͒.

4

20

N. M. Donahue, R. Mohrschladt, T. J. Dransﬁeld, J. G. Anderson, and M.

K. Dubey, J. Phys. Chem. A 105, 1515 ͑2001͒.

D. M. Golden, J. R. Barker, and L. L. Lohr, J. Phys. Chem. A 107, 11057

5

2

1

2

J. D. Watts, J. Gauss, and R. J. Bartlett, J. Chem. Phys. 98, 8718 ͑1993͒.

T. H. Dunning, J. Chem. Phys. 90, 1007 ͑1989͒.

R. A. Kendall, T. H. Dunning, and R. J. Harrison, J. Chem. Phys. 96, 6796

₆͑

2003͒.

J. L. Fry, S. A. Nizkorodov, M. Okumura, C. M. Roehl, J. S. Francisco,

and P. O. Wennberg, J. Chem. Phys. 121, 1432 ͑2004͒.

23

24^͑

1992͒.

7

8

M. P. McGrath and F. S. Rowland, J. Phys. Chem. 98, 1061 ͑1994͒.

H. H. Tsai, T. P. Hamilton, J. H. M. Tsai, M. van derWoerd, J. G. Harrison,

M. J. Jablonsky, J. S. Beckman, and W. H. Koppenol, J. Phys. Chem. 100,

D. E. Woon and T. H. Dunning, J. Chem. Phys. 98, 1358 ͑1993͒.

J. A. Pople, M. Head-Gordon, and K. Raghavachari, J. Chem. Phys. 87,

25

26⁵

968 ͑1987͒.

15087 ͑1996͒.

9

0

M. J. Frisch, G. W. Trucks, H. B. Schlegel et al., GAUSSIAN 98, Revision

K. N. Houk, K. R. Condroski, and W. A. Pryor, J. Am. Chem. Soc. 118,

3002 ͑1996͒.

A.9, 1998.

1

27

1

D. T. Colbert and W. H. Miller, J. Chem. Phys. 96, 1982 ͑1992͒.

Z. Bacic and J. C. Light, Annu. Rev. Phys. Chem. 40, 469 ͑1989͒.

See EPAPS Document No. E-JCPSA6-122-009510 for parameters of the

potential energy and dipole surfaces. A direct link to this document may be

found in the online article’s HTML reference section. The document may

also be reached via the EPAPS homepage ͑http://www.aip.org/pubservs/

epaps.html͒ or from ftp.aip.org in the directory /epaps/. See the EPAPS

homepage for more information.

R. D. Bach, M. N. Glukhovtsev, and C. Canepa, J. Am. Chem. Soc. 120,

75 ͑1998͒.

2

8

9

7

11

K. Doclo and U. Rothlisberger, Chem. Phys. Lett. 297, 205 ͑1998͒.

R. S. Zhu and M. C. Lin, J. Chem. Phys. 119, 10667 ͑2003͒.

J. F. Stanton ͑private communication͒.

B. M. Cheng, J. W. Lee, and Y. P. Lee, J. Phys. Chem. 95, 2814 ͑1991͒.

W. J. Lo and Y. P. Lee, J. Chem. Phys. 101, 5494 ͑1994͒.

I. B. Pollack, I. M. Konen, E. X. J. Li, and M. I. Lester, J. Chem. Phys.

1

2

3

4

5

6

3

0

1

119, 9981 ͑2003͒.

J. M. Guilmot, M. Godefroid, and M. Herman, J. Mol. Spectrosc. 160,

387 ͑1993͒.

K. J. Feierabend, D. K. Havey, and V. Vaida, Spectrochim. Acta, Part A

1

7

J. Matthews, A. Sinha, and J. S. Francisco, J. Chem. Phys. 120, 10543

2004͒.

I. M. Konen, I. B. Pollack, E. X. J. Li, M. I. Lester, M. E. Varner, and J.

₈͑

60, 2775 ͑2004͒.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

30.102.42.98 On: Sat, 22 Nov 2014 01:33:04

1

Article Doi

Role of OH-stretch/torsion coupling and quantum yield effects in the first OH overtone spectrum of cis-cis HOONO

DOI: 10.1063/1.1859273

Source and publish data:

Authors:

Article abstract of DOI:10.1063/1.1859273

Full text of DOI:10.1063/1.1859273

Products guided by the article

R&D Labs maybe for 7782-77-6

Relevant to this article

Hot Product