The photodissociation dynamics of NO2 at 308 nm and of NO 2 and N2O4 at 226 nm

THE JOURNAL OF CHEMICAL PHYSICS 124, 064309 ͑2006͒

The photodissociation dynamics of NO₂at 308 nm and of NO₂

and N₂O₄at 226 nm

M. Brouard,^a͒R. Cireasa,^b͒A. P. Clark, T. J. Preston, and C. Vallance

The Physical and Theoretical Chemistry Laboratory, The Department of Chemistry, University of Oxford,

South Parks Road, Oxford, OX1 3QZ, United Kingdom

͑Received 10 October 2005; accepted 21 December 2005; published online 10 February 2006͒

Velocity-map ion imaging has been applied to the photodissociation of NO₂via the ﬁrst absorption

band at 308 nm using ͑2+1͒ resonantly enhanced multiphoton ionization detection of the atomic

O͑³P_J͒ products. The resulting ion images have been analyzed to provide information about the

speed distribution of the O͑³P_J͒ products, the translational anisotropy, and the electronic angular

momentum alignment. The atomic speed distributions were used to provide information about the

internal quantum-state distribution in the NO coproducts. The data were found to be consistent with

an inverted NO vibrational quantum-state distribution, and thereby point to a dynamical, as opposed

to a statistical dissociation mechanism subsequent to photodissociation at 308 nm. Surprisingly, at

this wavelength the O-atom electronic angular momentum alignment was found to be small.

Probe-only ion images obtained under a variety of molecular-beam backing-pressure conditions, and

corresponding to O atoms generated in the photodissociation of either the monomer, NO₂, or the

dimer, N₂O₄, at 226 nm, are also reported. For the monomer, where 226 nm corresponds to

excitation into the second absorption band, the kinetic-energy release distributions are also found to

indicate a strong population inversion in the NO cofragment, and are shown to be remarkably

similar to those previously observed in the wavelength range of 193–248 nm. Mechanistic

implications of this result are discussed. At 226 nm it has also been possible to observe directly O

atoms from the photodissociation of the dimer. The O-atom velocity distribution has been analyzed

͓DOI: 10.1063/1.2166631͔

I. INTRODUCTION

states are accessed in the photon absorption step, and during

the subsequent evolution of these excited molecular states

into products? Can the dissociation be described by statisti-

cal methods ͑which assume a randomization of energy

throughout the internal modes at some critical point along

the reaction pathway͒, or do state-to-state, dynamical effects

make an important contribution to experimentally observed

results? How must one’s picture of the dissociation be modi-

ﬁed as the photolysis energy is changed?

Photodissociation of the nitrogen dioxide molecule has

been the source of considerable interest for many years. The

reason is essentially twofold, the photochemistry of the ni-

trogen oxide family ͑NO_x͒ being of great importance in at-

mospheric and combustion chemistry, and the photolysis of

NO₂itself having become a key benchmark system in the

study of fundamental photodissociation dynamics.

A number of factors have contributed to the popularity

of the photolysis process,

Since the seminal photofragment studies of Busch and

Wilson,^1,2a vast body of data, collected over a wide range of

photodissociation energies, has been amassed. The UV dis-

sociation of NO₂at wavelengths longer than about 250 nm

NO₂+ h␯ → NO͑²⌸_⍀;␷,J_NO,⌳͒ + O͑³P_J͒,

͑1͒

⌬H₀^ꢀ= 300.6 kJ mol⁻¹,

2

˜

occurs via excitation of the parent molecule from its X A₁

2

˜

ground state to a surface of mixed A B₂/X A₁electronic

character.^3–11A rapid nonadiabatic transition¹²through a

conical intersection then transfers the excited-state popula-

tion back onto the ground state, where decay into products

occurs. Product state distributions suggest that at low excess

energies ͑down to wavelengths around 350 nm͒, this unimo-

lecular decay can be described, at least on average, using

statistical theories.^13–26However, as evidenced, for example,

through highly inverted NO photofragment vibrational

distributions,^27–32at higher energies, where the dissociation

becomes more rapid, such models cease to be valid. In par-

ticular, population of the product vibrational degree of free-

dom appears to be determined somewhat earlier along the

as a focus for dynamical studies. The NO₂molecule is a rare

example of a stable yet simple open-shell molecule, possess-

ing a number of low-lying electronic states and an energetic

threshold to dissociation at a wavelength ͑around 398 nm͒

that has allowed the photolysis to be studied over a broad

range of excitation energies ͑and consequently product ex-

cess energies͒. Much attention has focused on the precise

nature of the dissociation pathway, particularly subsequent to

excitation into the ﬁrst absorption band. What electronic

a͒

Electronic mail: mark.brouard@chem.ox.ac.uk

Permanent address: Laser Department, National Institute of Laser, Plasma

b͒

and Radiation Physics, P.O. Box MG-36, Bucharest, Romania.

0021-9606/2006/124͑6͒/064309/15/$23.00

124, 064309-1

064309-2

Brouard et al.

J. Chem. Phys. 124, 064309 ͑2006͒

reaction coordinate than the rotational ͑and electronic͒ de-

grees of freedom.^17,19,21A visualization has emerged

whereby each product vibrational level is treated in isolation,

with the transition state used to determine the product rota-

tional distributions corresponding to each such channel tight-

ening as the excess energy is increased.^17,19,21,24This picture

is further supported through complementary time-resolved

studies of the dissociation rates.^7,24,33–39

the experiments employed a mixture of 5% NO₂seeded in

He at a backing pressure of ϳ2 bars, which was expanded

through a pulsed nozzle ͑General valve͒ with a 0.8 mm-diam

oriﬁce, and collimated by a 1-mm-diam skimmer. Additional

experiments were also performed with up to 5% O₂in the

gas mixture, to reduce residual levels of NO in the molecular

beam. The rotational temperature of the beam was deter-

mined to be ϳ50 K. Further downstream, the molecular

beam was passed through a 2 mm hole in the repeller plate

of the velocity mapping ion optics assembly and crossed

5 cm away from the nozzle exit by two counterpropagating

laser beams. The photolysis radiation was provided by a

Lambda Physik EMG103 excimer laser operating at 193 nm,

while the probe radiation was obtained by frequency dou-

bling the output of an excimer-pumped dye laser ͑Lambda

Physik LPD series͒. The time delay between the two laser

pulses was ϳ10 ns. Two planoconvex lenses of 30 cm focal

length were used to focus the radiation onto the molecular

beam. The O͑³P_J͒ photofragments were probed by ͑2+1͒

resonantly enhanced multiphoton ionization ͑REMPI͒ via the

Excitation of NO₂at wavelengths between 200 and

250 nm takes place via a second absorption band,^{29,31,32,40–52}

thought to involve an electronic state of ²B₂

symmetry,^{41,45–48,51}usually termed the D B₂or 2 B₂state.

At wavelengths close to 249 nm ﬂuorescence is

observed^40,41,43,51with a lifetime of around 42 ps,⁴⁰but this

decreases rapidly with vibronic level,⁵¹due to homogeneous

2

˜

51

2

˜

predissociation possibly by the lower A B₂state. Below

244 nm O͑¹D͒ production becomes energetically allowed,

and is produced with a near constant quantum yield of

around 50% between 244 and 193 nm.^41,48,52Although the

photodissociation dynamics in this region of the NO₂absorp-

tion spectrum are less well characterized than for the ﬁrst

absorption band, there is clear evidence for the production of

vibrationally highly excited NO at wavelengths between 248

and 193 nm.^32,46,47,52Both the O-atom velocity distribution

obtained at 212.8 nm ͑Ref. 46͒ and the vibrational popula-

tion distribution observed at 193 nm ͑Ref. 52͒ show evi-

dence of bimodality. Excitation at the latter wavelength may

involve a higher electronic state than the 2 ²B₂state, but the

similarities between the NO photofragment vibrational popu-

lation distribution at 193 nm ͑Ref. 52͒ and the corresponding

O͑³P_J͒ velocity distribution at 212.8 nm ͑Ref. 46͒ are strik-

ing.

3p ³P← P_Jtransitions near 225 nm. The probe laser energy

3

was typically around 500 ␮J pulse⁻¹. During image acquisi-

tion, the probe laser wavelength was scanned over the Dop-

pler proﬁle of the O͑³P_J͒ transitions in order to ensure an

equal detection sensitivity for all the product velocities. The

oxygen ions were velocity mapped onto an imaging detector

consisting of a pair of 40 mm chevron double microchannel

plates ͑MCPs͒ coupled to a P47 phosphor screen ͑DelMar

Ventures͒. The image on the phosphor was captured by an

intensiﬁed charge-coupled device ͑CCD͒ camera ͑Photonic

Science͒, electronically gated to the ﬂight time of the de-

tected ions, and sent to a personal computer ͑PC͒ for signal

processing ͑thresholding, event counting,⁷⁰and accumula-

tion͒. Images were averaged over 20 000 laser shots. Veloc-

ity calibration of the ﬁnal images was achieved using images

of O͑³P_J͒ from the photodissociation of O₂, the energetics

for which are well characterized.

Studies of vector correlations in this system focus prin-

cipally on the ﬁrst absorption band, and are generally con-

sistent with the dissociation mechanism outlined

above,^{2,13,15,16,25,26,53–60}with some studies providing addi-

tional information about excited-state lifetimes,^15,25,56,59and

possible ﬂuctuations in state-to-state rate coefﬁcients.^24,61–63

The alignment of the atomic O͑³P_J͒ ͑Jꢀ0͒ product has been

speciﬁcally investigated only once previously⁴⁶at a photoly-

sis wavelength of 212.8 nm. Evidence of atomic angular mo-

mentum polarization at longer wavelengths is mixed, with

one study suggesting alignment is present,⁵⁴and another

ﬁnding it to be absent.⁵⁵

In the present work we describe a new investigation of

the photofragmentation dynamics of NO₂at 308 and 226 nm

using velocity-map ion imaging. The following section de-

scribes the experimental and data analysis techniques em-

ployed. Section III presents the results from the study, while

in Sec. IV they are discussed in the light of previous experi-

mental and theoretical works. With Sec. V the paper con-

cludes with a brief summary of our principal ﬁndings.

In order to extract information on O͑³P_J͒ alignment, im-

ages were obtained in four geometries, labeled HH, HV, VH,

and VV according to the polarization of the pump and probe

lasers lying parallel ͑H͒ or perpendicular ͑V͒ to the image

plane. Pairs of images were collected simultaneously by us-

ing a photoelastic modulator to switch the polarization of the

probe laser on alternate laser shot. This procedure not only

reduced the errors due to experimental drift during the mea-

surements but also enabled the measurement of the total an-

gular momentum alignment, ͗A₂₀͘, from the measured inten-

sity ratios.⁷¹This alignment parameter is required to

normalize the image intensity prior to data analysis. In addi-

tion to this procedure, images were recorded on alternate

pairs of shots with and without ﬁring the photolysis laser.

This procedure thus generated “pump-probe” and “probe-

only” ion images to be accumulated concurrently, and sub-

sequently subtracted if necessary to remove background

probe-only signal from the pump-probe images.

II. METHOD

A. Experimental procedures

Attempts were made to obtain separate REMPI spectra

to determine the relative spin-orbit populations arising from

the photolysis. For this purpose, the total signal output of the

phosphor screen was sent to a boxcar averager, gated at the

The experiments were carried out using a standard

velocity-map⁶⁴ion imaging⁶⁵apparatus, which has been de-

scribed in detail previously.^57,66–69Brieﬂy, the majority of

064309-3

Photodissociation dynamics of NO₂

J. Chem. Phys. 124, 064309 ͑2006͒

appropriate arrival time. In principle the integrated signal

corresponding to each spin-orbit state could be used to de-

termine the ratio of the photofragment populations. However,

as discussed in Sec. III, these measurements were compro-

mised by the weak pump-probe signal for O͑³P₁͒ and

O͑³P₀͒, compared with the probe-only signal, and by an

overlap between the REMPI transitions used for O͑³P₁͒ de-

tection, and ͑1+1͒ REMPI transitions in NO. The pump-

probe signal for O͑³P₀͒ was so weak compared to the probe-

only signal that satisfactory ion images for this fragment

could not be obtained.

and angular momentum distributions of the product in the

ﬁnal or “detection” reference frame.

Equation ͑3͒ takes the form of a Fourier cosine series, in

which the coefﬁcient of each term is a sum over products of

“geometrical factors” f₀^Kand “dynamical factors” F^K₀. The

geometrical factors depend only on experimental geometry

and are easily calculated, while the dynamical factors are

functions of a set of alignment parameters which deﬁne the

scattering dynamics. Analytical expressions for the f^K₀and F₀^K

factors have been given in previous publications.^57,66,67Note

that simple relationships exist between each of the com-

monly used sets of alignment parameters ͑e.g., the bipolar

moments b^K₀͑k₁,k;␷͒ introduced in the semiclassical treat-

ment by Dixon,⁷³the polarization parameters a^k_q͑p͒ used by

Rakitzis et al.,^74,75and the alignment anisotropy parameters

of Bracker et al.⁷¹͒. Here we use the alignment anisotropy

B. Data analysis

The method used to extract dynamical information from

the velocity-map images is identical to that described in pre-

vious papers on the photodissociation of N₂O ͑Ref. 67͒ and

SO₂.⁶⁹Brieﬂy, the laboratory frame scattering distribution

P͑␷,⍀ ,⍀_j͒ of the O͑³P ͒ product following polarized laser

parameters ␤, s₂, ␣₂, ␥₂, and ␩ since these are appropriate

2

for

a

full

quantum-mechanical

treatment

of

photodissociation.⁷¹

␷

J

The Hertel-Stoll scheme⁷⁶is used to convert the ͑poten-

photolysis of NO₂may be expanded semiclassically in

spherical harmonics,

k

q

tially͒ complex quantities ␳ into the real quantities ␳

,

q

Љ

␳ ͑␷,⍀ ͒C^*͑⍀ ͒,

͑2͒

k

q

P͑␷,⍀ ,⍀ ͒ =

͚

␷

j

␷

j

kq

1

k

q +

k

+q

k

−q

q

Љ

k,q

␳

͓͑− 1 ͒␳ ͑␷_p,␾ ͒ + ␳ ͑␷_p,␾ ͔͒,

1

^{͑␷ ,␾ ͒ =}ͱ₂

p

T

Љ

where ␷ is the product speed and ⍀ =͑␪ ,␾ ͒ and ⍀_j

␷

ഛ q ഛ k

͑4͒

Љ

=͑␪_j,␾ ͒ are the laboratory frame polar coordinates of the

j

product velocity vector ͑i.e., the scattering angle͒ and total

and

angular momentum vector, respectively. The expansion coef-

k

q

ﬁcients ͑or rotational moments͒ ␳ ͑␷,⍀ ͒ are functions of

k

0+

k

␷

␳ ͑␷_p,␾ ͒ = ␳ ͑␷_p,␾ ͒.

͑5͒

T

0

both velocity and scattering angle, and the spherical har-

monic components depend only on the angular momentum

polar coordinates ⍀_j. The laboratory frame is deﬁned such

that the z axis lies along the polarization vector and the x axis

along the propagation vector of the photolysis light. The ﬁrst

step in obtaining the rotational moments of a velocity-map

image is to express the distribution in Eq. ͑2͒ in terms of

coordinates in a new reference frame, known as the time-of-

ﬂight ͑TOF͒ frame, in which z lies along the time-of-ﬂight

axis. ͑x is still deﬁned to lie along the photolysis laser propa-

gation direction.͒ The image rotational moments are then ob-

tained simply by integrating the distribution along the time-

of-ﬂight axis, mirroring the compression of the ion cloud

along this axis during the experiment. A second rotation, to a

detection frame, is often required to allow use of expressions

in the literature for the rotational line strengths ͑see, for ex-

ample, Ref. 72͒. The ﬁnal expression for the rotational mo-

ments of the images may then be written as

Using linearly polarized pump and probe radiation, and prob-

ing the O͑³P_J͒ photofragments via ͑2+1͒ REMPI, the ana-

lytical form of the images obtained depends only on the ␳⁰₀,

2

␳²₀, and ␳ rotational moments. For the four experimental

2+

geometries used the appropriate expressions for the images

are^66,71

P₂

0

2

0

ͱ

I_HV͑␷_p,␾ ͒ = I_VV͑␷_p,␾ ͒ = ␳ ͑␷_p,␾ ͒ + 5 ␳ ͑␷_p,␾ ͒,

T

P₀

ͱ

2 P₀

5 P₂

0

2

I_VH͑␷_p,␾ ͒ = I_HH͑␷_p,␾ ͒ = ␳ ͑␷_p,␾ ͒ −

͓␳ ͑␷_p,␾ ͒

T

0

2

ͱ

+ 3␳ ͑␷_p,␾ ͔͒,

͑6͒

T

2+

where P₂/P₀are line strength factors for the probe ͑2+1͒

ͱ

REMPI transition, taking the values 0, 1/2, and − 7/10 for

the J=0, 1, and 2 spin-orbit components of O͑³P_J͒.^77,78Ex-

1

k

q

k +q

e^{iq ␾}

͓1 + ͑− 1͒

͔

Ј

T

1

k

␳ ͑␷_p,␾ ͒

=

͚

͚ ͚

T DET

plicit expressions for the ␳ ͑␷_p,␾ ͒ appearing in these equa-

Љ

T

q

4␲

K=0,2

k

1

tions have been given by Bracker et al.⁷¹

q

Ј

ϫ f^K͑k ,k,q ,q ,R,R ͒F^K͑k ,k,q ;␷ ͒,

These expressions may be used in conjunction with Eqs.

͑3͒ and ͑4͒ to obtain analytical expressions for the Fourier

moments of the measured images. As described previously,⁶⁷

here we have used a basis set made up of a sum of Gaussian

functions to describe the speed dependence of the various

angular momentum polarization parameters. These may be ﬁt

to the Fourier moments extracted from the experimental data

in order to obtain the ͑velocity-dependent͒ alignment param-

eters characterizing the scattering distribution, using the

Ј Љ

Ј

1

p

0

͑3͒

in which ␷ and ␾ are the radial and angular coordinates of

p

T

the image, and ␪_Tis the angle that the product velocity vector

makes with the time-of-ﬂight axis. R=͑␣,␤,␥͒ and R

Ј

=͑␣ ,␤ ,␥ ͒ are Euler angles for the frame transformations,

Ј

Ј Ј

deﬁned in Table I of Ref. 57. The indices k , q and k, q

Ј

Љ

1

denote the spherical harmonic components of the velocity

064309-4

Brouard et al.

J. Chem. Phys. 124, 064309 ͑2006͒

methodology described in Ref. 67. The experimental Fourier

moments, c_n͑␷ ͒, of the ion images, I͑␷_x,␷ ͒␷_p=I͑␷_p,␾ ͒,

p

y

T

are deﬁned

2␲

c ͑␷ ͒ = N

I͑␷_x,␷ ͒cos͑n␾ ͒␷_pd␾_T,

͑7͒

͵

n

p

y

T

0

where the normalization constant N is equal to 1 when n

=0 and 2 when nϾ0. With linearly polarized pump and

probe radiation, n is restricted to even terms.

Four sets of pump-probe and probe-only images were

collected for each experimental geometry and spin-orbit state

of O͑³P_J͒ and used to determine the alignment parameters.

Some data for individual pump-probe geometries ͑identiﬁed

in the text to follow͒ were also analyzed using Abel inver-

sion methods described fully elsewhere.⁷⁹Error estimates for

the speed distributions and alignment parameters are given

as the standard deviation in the parameters returned from ﬁts,

and were determined using a Monte Carlo procedure de-

scribed previously.⁵⁷

III. RESULTS

3

A. NO₂+h␯„308 nm…\NO+O„ P₂…

The summed and background-subtracted velocity-map

ion images for the J=2 products obtained in these experi-

ments are shown in Fig. 1. Also presented are the Fourier

moments extracted from the images, and the ﬁts to these data

obtained using the analysis procedure described in Sec. II B.

Both the zeroth-order ͑large and positive in sign͒ and the

second-order ͑small and negative in sign͒ moments are ex-

tremely similar in shape and magnitude for the two experi-

mental geometry pairs HH/HV and VH/VV. This invariance

in the proﬁle of the images to the probe laser polarization

provides a visual indication that the angular momentum

alignment of O͑³P₂͒ in this system is very small. The fourth-

order moment in the HH geometry, which in principle can

arise only from terms describing angular momentum polar-

ization, is also very close to zero. Note that the second-order

moment in the VV geometry, and the fourth-order moments

in the HV, VH, and VV geometries, must be zero for the

pump-probe process of interest.⁵⁷

FIG. 1. ͑Color online͒ Background-subtracted images ͑left panels͒ and Fou-

rier moments ͑right panels͒ for O͑³P₂͒ products following NO₂photolysis at

308 nm. The experimental geometries are, from top to bottom, HH, HV, VH,

and VV. Experimental Fourier moments are in bold ͑—͒ and the ﬁtted mo-

ments dashed ͑----͒. Only those moments which are nonzero by symmetry

are shown. The fourth-order moment, c₄͑␷ ͒, is only nonzero for the HH

p

geometry, and the second-order moment, c₂͑␷ ͒, is zero for the VV geom-

p

etry. Note that the second-order moments are negative in the HH and HV

geometries, and all the zeroth-order moments, c₀͑␷ ͒, are positive. The

p

straight horizontal lines mark zero on the y axis.

should provide conﬁrmation of the results obtained from the

Fourier moment analysis method. It may be seen from Fig. 2

͑dashed line͒ that these data are generally in excellent agree-

ment with the parameters extracted from the full Fourier mo-

The speed distribution extracted from the analysis is

shown as the dashed line in Fig. 2. As expected, the oxygen-

atom products are generated with a broad spread of speeds,

peaking around 1900 m s⁻¹, and extending out towards the

energetic limit at around 2700 m s⁻¹. The speed-dependent

spatial anisotropy is also shown in Fig. 2. ␤͑␷͒ exhibits small

oscillations around a mean value of 1.0 0.2 ͑obtained by

averaging over the speed distribution͒. As expected from the

appearance of the images and their corresponding Fourier

moments, the terms parametrizing the angular momentum

polarization are very small. The speed-averaged polarization

moments, alongside their limiting values, are presented in

Table I. It is immediately obvious that within the experimen-

tal error, almost no angular momentum polarization is ob-

served.

FIG. 2. Comparison between the speed ͑left͒ and anisotropy ͑right͒ distri-

butions returned for the O͑³P₂͒ products of the 308 nm photodissociation of

NO₂, obtained using a full Fourier moment analysis ͑—͒, and a simpler Abel

inversion of ion images collected in the HH geometry ͑----͒, which ignores

rotational alignment effects.

The absence of angular momentum polarization renders

the three-dimensional ͑3D͒ distribution of oxygen products

cylindrically symmetric, and the Abel inversion method

064309-5

Photodissociation dynamics of NO₂

J. Chem. Phys. 124, 064309 ͑2006͒

TABLE I. Speed-averaged angular momentum alignment parameters for

O͑³P₂͒ products of NO₂photodissociation at 308 nm. Numbers in parenthe-

ses give the experimental error in the ﬁnal decimal place. The right column

gives the limiting values of the alignment parameters ͑Ref. 71͒.

Alignment parameter

This study

Limiting values

s₂

−0.003͑2͒

−0.008͑6͒

−0.004͑6͒

0.006͑7͒

−0.20 to +0.10

−0.10 to +0.15

−0.21 to +0.21

−0.20 to +0.20

␣

2

␩

2

␥

2

ment analysis. A slightly larger average value of 1.2 is ob-

tained for the translational anisotropy using an Abel

inversion of the data collected in the HH experimental ge-

ometry ͑shown as the smooth line on the right of Fig. 2͒,

compared with that obtained using the Fourier inversion

method. The small discrepancy between these values may

arise from several factors, but most likely reﬂects the fact

that the latter method uses information from only one experi-

mental geometry ͑i.e., one quarter of the total data collected͒.

3

B. NO₂+h␯„308 nm…\NO+O„ P₁…

The summed and background-subtracted velocity-map

ion images for the J=1 products are shown in Fig. 3. Along-

side are the extracted Fourier moments for these images and

the ﬁts returned to the data by the Fourier moment analysis

procedure. It is clear, both from the images themselves and

from their corresponding moments, that the data in this case

possess a pronounced bimodality. This bimodality was found

FIG. 4. ͑a͒ Speed distribution returned by Fourier moment analysis of the

O͑³P₁͒ images arising from NO₂photolysis at 308 nm. ͑b͒ Speed distribu-

tions returned by Abel inversion of the O͑³P₁͒ images in the HH ͑—͒ and

HV ͑----͒ geometries.

to persist even in the presence of added O₂to the gas mix-

ture, suggesting that the outer ring of the images did not arise

solely from NO contamination in the NO₂gas sample. The

speed distribution extracted from Fourier moment analysis of

the images conﬁrms this bimodality, and is shown in Fig.

4͑a͒. The lower-velocity component is similar to that ex-

tracted for the O͑³P₂͒ products. It begins at around

500 m s⁻¹, but is somewhat broadened compared to that for

O͑³P₂͒, apparently peaking at ϳ2300 m s⁻¹, and possibly

extending slightly beyond the energetic limit for single pho-

ton dissociation, which corresponds to a maximum velocity

of around 2680 m s⁻¹. The oscillations in the data probably

result from the increased experimental noise in this detection

channel ͓the signal being much reduced as compared to the

O͑³P₂͒ channel due to a lower population in this state͔. The

component in the high-velocity region of the distribution

cannot arise from a single photon dissociation of NO₂at

308 nm, since it lies well beyond the aforementioned ener-

getic cutoff.

FIG. 3. ͑Color online͒ Background-subtracted images ͑left͒ and Fourier mo-

ments ͑right͒ for O͑³P₁͒ products subsequent to NO₂photodissociation at

308 nm. Experimental geometries are, from top to bottom, HH, HV, VH,

and VV. Experimental Fourier moments are in bold ͑—͒ and the ﬁtted mo-

ments dashed ͑----͒. See also the caption to Fig. 1.

The presence of the high-velocity ͑multiphoton͒ peak

complicates the analysis, and the HH and HV images were

Abel inverted in order to extract an alternative, approximate

measure of the speed and anisotropy distributions. The speed

064309-6

Brouard et al.

J. Chem. Phys. 124, 064309 ͑2006͒

FIG. 6. Speed-dependent alignment anisotropy parameters s₂͑␷͒ ͑—͒, ␣₂͑␷͒

FIG. 5. Speed-dependent spatial anisotropy parameters ␤͑␷͒ returned by

Fourier moment analysis ͑—͒ and Abel inversion ͑----͒ of the O͑³P₁͒ images

arising from NO₂photolysis at 308 nm. Note that the analysis assumes a

single photon dissociation mechanism, whereas energetically O-atom prod-

ucts possessing velocities greater than ϳ3000 m s⁻¹must originate from a

multiphoton process, and are therefore not shown. See text for details.

͑----͒, ␥₂͑␷͒ ͑¯¯͒, and ␩ ͑␷͒ ͑· – · –͒ returned by Fourier moment analysis

2

of the O͑³P₁͒ images shown in Fig. 3. Note that, as with Fig. 5, the analysis

assumes a single photon dissociation process throughout, and therefore is

not appropriate for O-atom velocities greater than 3000 m s⁻¹͑not shown͒.

tained at high velocities, discussed above. It should be

stressed that the Fourier moment analysis method in the form

described in Sec. II B does not explicitly account for the

existence of multiphoton dissociation processes. As will be

discussed further in Sec. IV, it seems likely that a multipho-

ton process is responsible for the high-velocity component of

the distribution, and involves at least one photon from the

probe laser. Thus, the ion images and the derived spatial

anisotropy of the O͑³P₁͒ atoms are inﬂuenced by the polar-

ization vector of the probe laser radiation, and both the Abel

inversion and Fourier moment analyses are likely to break

down in this case.

distribution is shown in Fig. 4͑b͒. While the overall nature of

the distribution closely resembles that in Fig. 4͑a͒, the high-

velocity component is somewhat less intense than when the

Fourier moment analysis is used. Again, the distribution for

the low-velocity component peaks at higher speed than was

found for the O͑³P₂͒ products.

The spatial anisotropy parameters obtained via the two

analysis methods are shown in Fig. 5. Agreement between

the two distributions is excellent in the low-velocity ͑one

photon͒ regime ͑that lying between 1000 and 2700 m s⁻¹, as

displayed in the ﬁgure͒, indicating that for the 308 nm pho-

tolysis component leading to O͑³P₁͒, angular momentum po-

larization is extremely small. Averaging over the speed range

of this peak, an anisotropy of ␤=1.3 0.2 is obtained from

the Fourier moment ﬁt, which is slightly higher than was

measured for the O͑³P₂͒ product channel. The apparent spa-

tial distribution of the products in the high-velocity region

͑not shown in Fig. 5͒ is more complicated. The apparent ␤͑␷͒

parameter at these high speeds is strongly dependent on the

analysis method employed, the Fourier moment method re-

turning an effectively velocity-independent spatial aniso-

tropy of around 0.9 in this region, and the Abel-inverted

image exhibiting a distribution characterized by a ␤͑␷͒ pa-

rameter of greater than 2. The latter value is nonphysical for

a single photon dissociation in which product angular mo-

mentum alignment is assumed to be zero ͑see further discus-

sion in Sec. IV͒.

3

C. NO₂+h␯„226 nm…\NO+O„ P_J…

The primary function of the images collected with only

the probe laser in operation was to allow for a background

subtraction of the contribution made to the total photolysis

signal by photodissociation at the REMPI detection wave-

lengths. However, this procedure also generates an additional

data set that is amenable to analysis. Representative images

collected on the probe-only transitions for O͑³P₂͒, O͑³P₁͒,

and O͑³P₀͒ are presented in Fig. 7. These data were collected

with O₂added to the gas mixture, which in this case helped

to suppress background signals from NO contaminants. A

relatively isotropic central feature is evident for all the spin-

The angular momentum alignment parameters returned

by the Fourier moment ﬁt are shown in Fig. 6. In the low-

velocity ͑one photon͒ region displayed in the ﬁgure, the

alignment parameters oscillate around zero and suggest that

angular momentum polarization is very small ͓although their

residual values are somewhat higher than for the O͑³P₂͒

case͔. In the high-velocity region, not shown in the ﬁgure,

the returned, apparent values for the alignment parameters

FIG. 7. Probe-only velocity-map ion images of O͑³P₀͒ ͑left͒, O͑³P₁͒

͑middle͒, and O͑³P₂͒ ͑right͒, generated subsequent to photolysis of NO₂at

226 nm. The data were obtained with a mixture of 10% NO₂in He.

s₂͑␷͒ and ␩ ͑␷͒ become relatively large, a fact that explains

2

the discrepancy between the effective ␤͑␷͒ parameters ob-

064309-7

Photodissociation dynamics of NO₂

J. Chem. Phys. 124, 064309 ͑2006͒

orbit products, but this component is dominant in the case of

O͑³P₂͒. The O͑³P₀͒ and O͑³P₁͒ products show a particularly

pronounced bimodal structure, with the fast component

showing clear evidence of spatial anisotropy.

The probe-only images for the O͑³P₁͒ and O͑³P₂͒ prod-

ucts provide insufﬁcient data to extract angular momentum

alignment information, and this must be assumed to be zero

͑unlike in the pump-probe data above, where a full data

analysis concluded this to be the case͒. Ahmed et al.⁴⁶found

evidence for angular momentum alignment in the O͑³P₁͒

products of the 212.8 nm photodissociation of NO₂. If simi-

lar angular momentum polarization effects were to occur at

226 nm, this would result in some error in the Abel-inverted

speed distributions and translational anisotropies for O͑³P₁͒

and O͑³P₂͒. However, it is reassuring to note that the re-

turned speed distribution for the O͑³P₁͒ photofragment ͑see

below͒ is very similar to that for the O͑³P₀͒ products, which

necessarily cannot possess angular momentum alignment,

suggesting that the effects of any alignment on the returned

velocity distributions for O͑³P₁͒ and O͑³P₂͒ are not that

large in the present case.

The speed distributions returned from an Abel inversion

of the O͑³P₀͒, O͑³P₁͒, and O͑³P₂͒ images are presented in

Fig. 8. Bimodal structure is present for all three spin-orbit

products, with the ratio of the fast component to the slow

component much greater for the O͑³P₀͒ and O͑³P₁͒ channels

than for O͑³P₂͒ ͑as can be seen from the raw images͒. Close

inspection of the probe-only data revealed day-to-day varia-

tions in the ratio of the fast to slow components of the speed

distributions. Variations in laser power over the range of

0.5–1.5 mJ were found to have no marked effect on the im-

age proﬁles. However, the relative concentration of NO₂in

the NO₂/He gas mixture was found to inﬂuence strongly the

resulting images. Shown in Fig. 8 are the speed distributions

returned for the O͑³P₀͒, O͑³P₁͒, and O͑³P₂͒ photolysis prod-

ucts as the NO₂concentration was successively reduced,

from 10% ͑by partial pressure͒, to 1%, and then to a nomi-

nally trace quantity ͑obtained by evacuating the gas line sev-

eral times and reﬁlling with He only͒. It should be stressed

that the true variation in reagent partial pressure will not be

close to as dramatic as indicated, since a signiﬁcant quantity

of NO₂is likely to remain within the gas system adsorbed

onto surfaces even under evacuation. Nonetheless, it is im-

mediately evident that the relative intensity of the slow com-

ponent in the overall O͑³P₁͒ speed distributions, in particu-

lar, increases as the concentration of NO₂in the gas line

becomes greater.

The strong partial pressure dependence observed appears

to indicate a product channel arising from the dissociation of

a higher-order cluster of NO₂. NO₂is well known to exist in

equilibrium with its dimer, and the dissociation of N₂O₄it-

self has been the subject of a number of dynamical

studies.^32,80–85In the region of the probe laser wavelength,

the absorption cross section of this species is between one

and two orders of magnitude greater than that of the

monomer.^86,87Consequently, even small N₂O₄concentra-

tions may contribute signiﬁcantly to the observed results in

the probe-only channel if a dissociation pathway forming

O͑³P_J͒ fragments existed. Note that this is unlikely to be the

FIG. 8. Speed distributions extracted from probe-only velocity-map ion im-

ages of O͑³P₀͒ ͑top͒, O͑³P₁͒ ͑middle͒, and O͑³P₂͒ ͑bottom͒ obtained at

226 nm, and at various NO₂partial backing pressures: ͑—͒ high pressure;

͑----͒ medium pressure; ͑¯¯͒ low pressure. See text for further details.

case in the pump-probe dissociation channel at 308 nm,

since the absorption cross sections are comparable in magni-

tude at this wavelength. Estimates of the contribution of

N₂O₄to the probe-only signal can be made based on the

known equilibrium constant between NO₂and N₂O₄, their

relative absorption cross sections at the probe

wavelength,^86,87and the approximate partial pressures of

NO₂employed, and these are in broad agreement with the

observations shown in Fig. 8. Interestingly, the contribution

of the dimer to the total probe-only signal is very dependent

on O͑³P_J͒ spin-orbit state, a point that will be returned to in

Sec. IV C.

Abel inversion of the images suggests that the low speed

component, associated with photodissociation of N₂O₄, is

064309-8

Brouard et al.

J. Chem. Phys. 124, 064309 ͑2006͒

characterized by a low value of ␤ around 0.1. By contrast, as

is clearly visible from the raw O͑³P₀͒ ion images, the fast

component is much more anisotropic, and is characterized by

a ␤ parameter close to unity.

IV. DISCUSSION

3

A. NO₂+h␯„308 nm…\NO+O„ P₂…

1. Speed distribution

The distribution of O͑³P₂͒ product speeds presented in

Fig. 2 reﬂects the manifold of populated rovibronic levels

within the undetected NO coproducts. The O͑³P₂͒ and

O͑³P₀͒ recoil energy distributions obtained by Demyanenko

et al.⁵⁵are the closest in photolysis wavelength ͑338.9 nm͒

to compare with the present work, and it is striking how

different the longer wavelength distributions are from those

obtained here at 308 nm. Literature reports of the NO inter-

nal energy disposal at 308 nm, which can be thought of as

bridging the high and low excess energy regimes under

which the dissociation has been studied, are currently some-

what contradictory.^19,20,28In particular, there is some dis-

agreement as to whether the NO vibrational distribution at

this ͑relatively high͒ excess energy can be described by sta-

tistical methods,¹⁹or rather is best thought of as being under

dynamical control.^20,28NO rotational populations within the

various energetically accessible ͑0ഛ␷ഛ3͒ vibrational levels

are, on the other hand, relatively well characterized at this

wavelength.^19,28Due to the difﬁculties associated with deter-

mining vibrational populations by laser-induced ﬂuorescence

͑LIF͒, it seems reasonable to suppose that the rotational dis-

tributions reported in these studies are likely to be more ro-

bust than the corresponding vibrational populations. In order

to investigate the deposition of vibrational energy in the NO

products implied by the current data, we have therefore cho-

sen to assume that one or other of the rotational distributions

previously reported at this wavelength^19,28represent a rea-

sonable approximation to the true NO rotational populations

in the present data. A series of “vibrational envelopes”—

whose individual rotational distributions are taken directly

either from Refs. 28 or 19—were constructed to model the

internal energy deposition in the NO coproduct, and there-

fore reproduce the speed distribution generated in the present

experiment. The vibrational populations were used as adjust-

able parameters in a ﬁt of the vibrational envelopes to the

experimental speed distribution. The resulting ﬁt assuming

the rotational distributions of Knepp et al.¹⁹is shown as the

dashed line in Fig. 9. The “vibrational envelopes” used as

basis functions in the ﬁtting are shown as dotted lines to-

wards the bottom of the ﬁgure. The magnitudes of the basis

functions are obtained by a simple genetic algorithm ﬁt of

the overall product speed distribution to that measured in the

current study.

FIG. 9. Fit ͑----͒ to the O͑³P₂͒ speed distribution following 308 nm pho-

tolysis ͑—͒, using vibrational basis functions ͑¯¯͒ constructed from the

data of Knepp et al. ͑Ref. 19͒ ͑see text for further details͒. From right to left

͑high product speed, low internal excitation to low product speed, high

internal excitation͒, these functions correspond to the ␷=0, ␷=1, ␷=2, and

␷=3 NO products, respectively.

et al. was used to generate the vibrational envelopes,²⁸are

shown in Fig. 10 and Table II. Note that the two alternative

vibrational distributions are in reasonable agreement, al-

though the preferred population distribution, that obtained

using the data of Knepp et al.,¹⁹is somewhat hotter than that

obtained using the rotational populations of Zacharias et al.²⁸

Also shown in the ﬁgure for comparison are the previously

reported vibrational populations from these two LIF

studies,^19,28and from the Fourier transform infrared ͑FTIR͒

experiment of Doughty et al.²⁰In contrast to the vibrational

populations reported in Refs. 28 and 19, the present data

show an inverted population distribution, with very little NO

borne in its lowest ␷=0 vibrational level. Broadly speaking,

the relative populations in ␷Ͼ0 are in reasonable agreement

with the results of Knepp et al.¹⁹One problem reported in

The best ﬁt to the extracted speed distribution was ob-

tained when the rotational populations used to construct the

basis functions were taken from the experimental results of

Knepp et al. ͑which, as in the current study, made of use of

a jet-cooled reactant beam͒.¹⁹The returned relative NO vi-

brational populations for this ﬁt, and for the best ﬁt achieved

when the results of the room-temperature study of Zacharias

FIG. 10. Comparison between NO vibrational distributions obtained in this

work ͑—͒ subsequent to NO₂photodissociation at 308 nm, with those ex-

tracted in previous studies ͑----͒. Labeling is as follows: -᭝-, current work,

rotational populations taken from Zacharias et al. ͑Ref. 28͒; -᭞-, current

work, rotational populations taken from Knepp et al. ͑Ref. 19͒; --᭿-- Za-

charias et al. ͑Ref. 28͒; --

ࡗ

-- Knepp et al. ͑Ref. 19͒; --b-- Hancock et al.

͑Ref. 20͒.

064309-9

Photodissociation dynamics of NO₂

J. Chem. Phys. 124, 064309 ͑2006͒

TABLE II. The ﬁrst two columns show estimated vibrational population distributions determined from ﬁts to

the experimentally determined O͑³P₂͒ velocity distribution shown in Fig. 2 assuming the rotational distributions

measured previously either by Knepp et al. ͑Ref. 19͒ or Zacharias et al. ͑Ref. 28͒. The three columns on the

right show the vibrational distributions obtained by Knepp et al. ͑Ref. 19͒, Zacharias et al. ͑Ref. 28͒, and

Doughty et al. ͑Ref. 20͒.

Vibrational

level

Fit

͑Knepp͒

Fit

͑Zacharias͒

Knepp^a

Zacharias^b

Hancock^c

0

1

2

3

0.07

0.40

0.13

0.07

0.51

0.35

0.07

0.55

0.22

0.13

0.10

0.29

0.28

0.19

0.22

¯

0.28

0.38

0.34

^aReference 19.

^bReference 28.

^cReference 20.

both the earlier LIF experiments was the presence of cold

NO as a contaminant in the NO₂source—the vast majority

of this would be in its lowest vibrational state and could

therefore lead to artiﬁcially high levels of ␷=0 NO “prod-

ucts.” The present results are in reasonable agreement with

the results of Doughty et al. ͑who were unable to extract the

population in ␷=0͒,²⁰although the latter work reported a

higher proportion of NO͑␷=3͒ products than observed here.

It is possible that the present estimate for the yield of ␷=3 is

particularly sensitive to the assumed rotational distribution in

lower vibrational levels. Qualitative inspection of Fig. 9 sug-

gests that the population in ␷=3 might be as much as double

the present estimate if the ␷=2 rotational distribution em-

ployed to generate the ﬁtting functions was somewhat colder

than the one reported by Knepp et al.¹⁹Such an enhancement

in ␷=3 population would bring the current vibrational

distribution into much better agreement with those of

Doughty et al.²⁰

It should also be noted that in each of the aforemen-

tioned previous studies of the 308 nm photodissociation of

NO₂, the NO species was measured directly, and therefore

provided a weighted average over each of the three O͑³P_J͒

spin-orbit levels populated. If the three oxygen spin-orbit

states were to be correlated with rather different ensembles

of rovibrational quanta in the corresponding NO coproducts,

the current results would no longer merit direct comparison

with these studies. We return to this point in Sec. IV B.

O͑³P₂͒+NO͑␷,J_NO͒ product channels, and as such any dis-

cussion of its functional form is necessarily complicated. In

fact, even at 2500 m s⁻¹, the present results are not directly

comparable with those detecting the NO͑␷=0,N=30͒

species.^56,57As may be seen from Fig. 9, low rotational lev-

els of the ﬁrst vibrationally excited state of NO already over-

lap the ␷=0 products at this speed. Only at the far extremities

of the product speed range do single NO vibrational levels

contribute to the distribution, and may the NO rotational

level dependence of the anisotropy properly assessed. Unfor-

tunately, it is just at these extremes of the distribution where

the intensity is small, and extraction of reliable translational

anisotropy parameters most problematic.

3. Angular momentum alignment

The present results indicate that the O͑³P₂͒ fragments of

NO₂photodissociation at 308 nm do not show angular mo-

mentum alignment. Although there are no studies with which

these results bear direct comparison, the conclusions drawn

here are similar to those of Demyanenko et al., who sug-

gested that atomic polarization was not present in dissocia-

tion between 372 and 388 nm.⁵⁵These results are somewhat

surprising, given that the spin-orbit states of O͑³P_J͒, particu-

larly at longer wavelengths, are produced with nonstatistical

populations,⁴⁴and that polarized oxygen atoms have been

observed subsequent to photodissociation via the second ab-

sorption band.⁴⁶

Although we have no quantitative explanation for the

absence of O-atom polarization in the photodissociation of

NO₂via the ﬁrst absorption band, consideration of the orbital

correlation diagram provides some clues as to the origin of

the behavior. At linearity, the O͑³P͒+NO͑²⌸͒ products cor-

relate with doublet and quartet states of the NO₂molecule

with ⌺⁺, ⌺⁻, ⌸, and ⌬ symmetries.^6,8The ⌸ state is formed

by bringing the ﬁlled p-orbital on the oxygen atom up to the

NO molecule at linearity, and this state is therefore the most

repulsive. The ⌺⁺, ⌺⁻, and ⌬ states are formed by bringing a

half-ﬁlled p-orbital on the oxygen atom up to the NO mol-

ecule. On bending, the symmetries of these states are re-

2. Spatial anisotropy

The spatial anisotropy parameters obtained in the photo-

dissociation at 308 nm have been explored in previous stud-

ies by Brouard et al.⁵⁶and Bass et al.⁵⁷The values reported

for the ␤ parameter were 1.4 0.1 ͑under jet-cooled condi-

tions͒ and 1.22 0.04 ͑at room temperature͒. At ﬁrst glance,

the current results seem to be rather low, but it is important

to note that the previous results reported the spatial aniso-

tropy for a single NO quantum state ͑␷=0, N=30͒.^56,57In

fact, the present data show that at the O͑³P₂͒ speed correlat-

ing with these products ͑around 2500 m s⁻¹͒, the anisotropy

reported here agrees well with the previously determined NO

quantum-state resolved values.

duced to A , A , and A +A , respectively. In the asymptotic

Ј

Љ

Ј

Љ

region, the ⌺⁺state is the lowest in energy, and at nonlinear

conﬁgurations this state correlates adiabatically with the

ground 1 ²AЈ state of NO₂.^6,8It is worthwhile comparing

The ␤͑␷͒ parameter observed at any given velocity rep-

resents a weighted average over the various overlapping

064309-10

Brouard et al.

J. Chem. Phys. 124, 064309 ͑2006͒

these state correlations with those for other systems that dis-

sociate via states of ⌺ symmetry, and for which O-atom po-

larization has been observed, for example, SO₂͑Ref. 69͒ and

N₂O.⁶⁷Unlike these other systems, for NO₂the ⌸ character

of the NO coproduct means that twice as many A and A

states are generated when the half-ﬁlled O-atom orbital is

directed towards the diatomic molecule ͑these states being

the aforementioned ⌺⁺, ⌺⁻, and ⌬ at linearity͒ compared

with the case in SO₂and N₂O. Thus, on bending, a more

dense manifold of crossings and avoided crossings will be

present for NO₂than for SO₂or N₂O, possibly leading to

scrambling of the O-atom alignment. Why the alignment

should be seen subsequent to absorption via the second ad-

sorption band remains a mystery, although it is perhaps sig-

niﬁcant that, in that case, dissociation may not proceed ulti-

mately on the ground-state potential-energy surface.

=1.00:0.23:0.04͒, and that the O͑³P₀͒ products are part-

nered exclusively by NO in ␷ =0. The latter assumption en-

Ј

sures that our estimate for the average vibrational population

distribution is the coldest that is compatible with the O͑³P₁͒

and O͑³P₂͒ data ͓the population in O͑³P₀͒ is so small that in

any event it has little inﬂuence on the averaged distribution͔.

The resulting spin-orbit averaged NO coproduct

populations are then predicted to be P͑␷=0:1:2:3͒

=0.15:0.38:0.35:0.12. These estimated vibrational popula-

tions are closer to the previous results from the LIF

studies.^19,28Nevertheless, in spite of the uncertainties in-

volved in averaging the present results over O-atom spin-

orbit state, it is important to stress that a clear conclusion

from the present work is that the NO coproducts must be

produced with a vibrational population inversion. This result

suggests a dynamical, as opposed to a statistical, interpreta-

tion of the dissociation mechanism. The vibrational popula-

tion inversion is likely to reﬂect changes in bond length

and/or vibrational frequency during excitation and dissocia-

tion, with the latter occurring on a time scale that is prompt

relative to that for the statistical redistribution of internal

energy in the excited NO₂molecule. At the excess energies

supplied in the present work, it should also be borne in mind

that dissociation may not only occur exclusively via the

ground electronic state potential-energy surface but may also

Ј

Љ

3

B. NO₂+h␯„308 nm…\NO+O„ P₁…

1. Speed distribution

The speed distribution for the O͑³P₁͒ products shown in

Fig. 4 raises several questions. The lower-velocity peak rep-

resents the pump-probe signal from dissociation at 308 nm,

but while it is generally in reasonable accord with the O͑³P₂͒

data, the distribution is somewhat broader, and clearly peaks

at higher velocity, than that of the O͑³P₂͒ products. Mean-

while, the feature at high velocity, which is absent entirely in

the O͑³P₂͒ images, cannot energetically arise from photoly-

sis at 308 nm.

2

˜

take place directly on the excited A B₂͑2 AЈ͒ electronic

state.

While the scenario in which different spin-orbit states of

the O atom correlate with different NO coproduct vibrational

populations is an appealing one, this interpretation of the

O͑³P₁͒ results should be treated with some caution. The dif-

ferences in the spin-orbit state-resolved velocity distributions

observed here seem quite large compared with the data ob-

tained by Hsieh et al. at 355 nm,⁵⁴and the data of Demy-

anenko et al. at 338.9 nm.⁵⁵At the higher recoil velocities

accessed in the present work, one might expect smaller dif-

ferences in the speed distributions for the three spin-orbit

states compared with those observed at 338.9 and

355 nm.^67,69It is possible that another phenomenon is re-

sponsible for the broadening of the O͑³P₁͒ nascent product

velocity distribution. One additional source of such broaden-

ing is the space charging effect. It should be noted that a

large ion signal was detected at the mass corresponding to

NO⁺at the REMPI wavelength for detection of the O͑³P₁͒

͓and O͑³P₀͔͒ products. This could have resulted in the large

ion density in the interaction region necessary to cause this

effect.

A consideration of the potential sources of the high-

velocity products observed in these experiments provides

some evidence that space charging may be responsible, at

least in part, for the broadening of the O͑³P₁͒ speed distri-

bution with respect to that for the O͑³P₂͒ products. As has

been indicated already, this component cannot arise from a

single photon dissociation process, lying beyond the physical

limit for photolysis at 308 nm. The probe-only dissociative

contribution at 226 nm, which could give rise to products

with these velocities, has already been subtracted from the

total signal. A further requirement for the high-velocity com-

ponent is that it must involve resonant detection of O⁺ions.

The differences between the speed distributions for the

two detected oxygen spin-orbit states in the low speed

͑308 nm photolysis͒ region may reﬂect a number of factors.

First, as noted above, it is possible that the dynamics of the

dissociation do genuinely lead to atomic products that, de-

pending on their spin-orbit level, correlate to different quan-

tum states of the NO cofragments. In fact, in the core-

sampling experiments of Hsieh et al.,⁵⁴which used REMPI

detection to establish the speed distributions of each of the

O͑³P_J͒ spin-orbit states following dissociation at 355 nm,

such variations were indeed observed. Their study revealed

all three O͑³P_J͒ spin-orbit states to be correlated with differ-

ent

ratios

of

NO͑␷=0͒:NO͑␷=1͒

and

NO͑␷

=1,low J_NO͒:NO͑␷=1,high J_NO͒ coproducts. Similar spin-

orbit state speciﬁc behavior has also been observed by De-

myanenko et al.⁵⁵at selected wavelengths as short as

338.9 nm. In the current experiments, the consequence

would be that a larger quantity of NO͑␷=0͒ would be formed

in conjunction with O͑³P₁͒ products than with O͑³P₂͒ prod-

ucts. A ﬁt to the low speed component of the O͑³P₁͒ velocity

distribution, along the same lines as that shown in Fig. 9 for

O͑³P₂͒ and assuming the rotational population distributions

determined by Knepp et al., yields a vibration population

P͑␷=0:1:2:3͒=0.40:0.35:0.15:0.09. We may use these

NO coproduct vibrational populations, together with those

for the O͑³P₂͒ products, to estimate the spin-orbit averaged

NO vibrational populations. In making this estimate, we

have assumed an average of the O-atom spin-orbit

populations of Miyawaki et al.⁸⁸at 266 nm, ͑P͑J=2:1:0͒

=1.00:0.17:0.03͒,

and

337 nm

͑P͑J=2:1:0͒

064309-11

Photodissociation dynamics of NO₂

J. Chem. Phys. 124, 064309 ͑2006͒

One is therefore left to consider what combination of pho-

tons from the 308 nm laser and the 226 nm laser could lead

to further generation of oxygen products. The NO A͑²⌺⁺͒

←X͑²⌸͒ electronic transition, its 0-0 band lying at around

the wavelength of the probe laser wavelength of 226 nm, is

known to be a highly efﬁcient process.⁸⁹It therefore seems

likely that a large quantity of electronically excited NO

would be generated if the probe laser wavelength were to

coincide with a resonant transition in this band. Indeed, the

wavelengths of the O͑³P₁͒ and O͑³P₀͒ REMPI transitions

are near resonant with a number of A←X transitions in the

0-0 band originating from low rotational levels in NO.^90,91

These levels would be populated signiﬁcantly by photodis-

sociation of NO₂at 308 nm. It is also possible that a very

small NO residual impurity remains in the molecular-beam

expansion, even when O₂is added to the gas mixture. The

electronically excited NO species thereby produced may then

absorb a further photon that will cause either dissociation or

ionization,

detector would then result either from ͑2+1͒ REMPI of the

atomic oxygen products formed in Eq. ͑8͒ or alternatively

via photodissociation of the NO⁺ions produced in Eq. ͑10͒.

͑A further two-probe laser photons would be required to dis-

sociate NO⁺into N+O⁺.͒ Processes ͑8͒ and ͑9͒ generate

O-atom products with velocities of 4120 and 1886 m s⁻¹, and

the former is quite close to the observed secondary peak in

the velocity distribution, allowing for the effects of space-

charge broadening. The sign and approximate magnitude of

the anisotropy parameter for process ͑8͒ ͑Ref. 89͒ is also

consistent with present observations. The space charging ef-

fect which precluded the collection of images at this probe

laser wavelength in the experiments of Bakker et al.,⁸⁹might

also cause deviations between the low-velocity features for

the O͑³P₁͒ images, as opposed to the O͑³P₂͒ data, and could

potentially blur the nascent product distribution signiﬁcantly.

In summary, the energetics of the components in the

speed distribution, their energy broadening, and the observa-

tions of previous work are consistent with the operation of a

second, multiphoton dissociative channel described above. It

is also conceivable that this process might be the cause of

some of the discrepancies we have observed at 308 nm be-

tween the two spin-orbit states detected in this study. How-

ever, irrespective of these potential complications, it is worth

reiterating that a ﬁrm conclusion from the present work is

that subsequent to the photodissociation of NO₂at 308 nm

NO is produced with a vibrational population inversion.

NO͑²⌸͒ + h␯͑226 nm͒ → NO͑²⌺⁺͒

+ h␯͑308 or 226 nm͒

→ N͑⁴S͒ + O͑³P͒,

͑8͒

͑9͒

NO͑²⌸͒ + h␯͑226 nm͒ → NO͑²⌺⁺͒

+ h␯͑308 or 226 nm͒

→ N͑²D͒ + O͑²P͒,

2. Spatial anisotropy and alignment parameters

NO͑²⌸͒ + h␯͑226 nm͒ → NO͑²⌺⁺͒

As mentioned in Sec. III B, the presence of multiphoton

processes in the high-velocity region of the O͑³P₁͒ velocity

distribution complicates the analysis of the angular distribu-

tions and angular momentum alignment. Nevertheless, at the

low speeds characterizing the component arising from single

photon photolysis at 308 nm, the alignment is again found to

be extremely small—in accord with the data from the O͑³P₂͒

products presented in Sec. III A. The spatial anisotropy ob-

tained from the Fourier moment analysis for these products

is well reproduced by a simple Abel inversion procedure, and

is also of similar magnitude to that for the O͑³P₂͒ channel.

The interpretation of the translational anisotropy of the

high speed component of O͑³P₁͒ has already been touched

on in the preceding section. At second photon wavelengths

around 339 nm, process ͑8͒, generating O͑³P₂͒, is known to

possess a positive ␤ parameter of around unity.⁸⁹The angular

distribution in the outer high-velocity ring observed here for

O͑³P₁͒ at second photon wavelengths around 308 nm is con-

sistent with this ﬁnding, though the distribution is severely

broadened by space-charge effects. In light of this broaden-

ing, more detailed analysis is probably unwarranted.

+ h␯͑308 or 226 nm͒ → NO⁺.

͑10͒

The two-photon dissociation process shown in Eq. ͑8͒ has

been observed by Bakker et al.⁸⁹In their experiments, the

photolysis laser was tuned to the 226.19 nm, which lies in

the center of the Q₁₁+P₁₁bandhead. Although the focus of

the study was on subsequent dissociation of the excited

A͑²⌺⁺͒ state at 339 nm, an intense ring was also observed in

their velocity-map imaging study that corresponded to disso-

ciation at the probe laser wavelength ͓in this case the wave-

length of the O͑³P₂͒͑2+1͒ REMPI transition at 225.7 nm͔. A

leakage signal that corresponded to the production of large

quantities of NO⁺was also present and was assigned to the

process indicated in Eq. ͑10͒. Interestingly, the same authors

noted that at the O͑³P₁͒͑2+1͒ REMPI detection wavelength,

an accidental resonance ͑or near resonance͒ in the NO

A͑²⌺⁺͒-X͑²⌸͒ transition generated large quantities of re-

sidual NO⁺through ͑1+1͒ REMPI, with resulting space

charge effects that precluded the collection of images for this

spin-orbit product.

3

C. NO₂+h␯„226 nm…\NO+O„ P_J…

In the light of this information, it seems probable that a

multiphoton process, possibly incorporating ﬁrst the disso-

ciation of NO₂by the photolysis laser at 308 nm followed by

resonant excitation of NO to the A state, and then photodis-

sociation or photoionization at either 308 or 226 nm, is re-

sponsible for the high-velocity peak observed in these ex-

periments. Formation of the product ions incident on the

1. Photodissociation at high NO₂partial pressures

As noted in Sec. III, analysis of the probe-only images at

226 nm is complicated by photodissociation of N₂O₄, par-

ticularly under conditions of high partial backing pressures

of NO₂. A number of studies of N₂O₄photolysis have been

conducted at a dissociation wavelength of 193 nm.^80–82The

064309-12

Brouard et al.

J. Chem. Phys. 124, 064309 ͑2006͒

experiments of Kawasaki et al.⁸⁰and of Sisk et al.⁸¹detected

emission from electronically excited NO₂that was assigned

by the latter researchers to a dissociation channel that yielded

electronically excited NO₂͑1 ²B₂͒+NO₂͑1 ²B₂͒ products. A

more recent study at 193 nm by Mueller et al.⁸²used photo-

fragment translational spectroscopy to observe two-product

dissociation channels, which were tentatively assigned

2

4

2

˜

to

NO₂͑X A₁͒+NO₂͑1 B₂/1 A₂͒

and

NO₂͑X A₁͒

+NO₂͑2 ²B₂͒ products, respectively. In this experiment, a

slight broadening of peaks in the NO⁺and O⁺TOF spectra

relative to the corresponding peak in the NO⁺₂spectrum was

concluded to arise from secondary dissociation of some of

the NO₂products into O+NO fragments. ͑No evidence of

NO formation had previously been observed in the earlier

study of Kawasaki et al.⁸⁰͒ The observations of Mueller et al.

are consistent with the current results, which suggest that

O͑³P_J͒ products are generated via the process,

N₂O₄+ h␯͑226 nm͒ → NO₂+ NO + O͑³P_J͒,

͑11͒

on the time scales of the present experiments. Once scaled

for the different photon energies employed, the O-atom ve-

locity distribution observed in the present work at 226 nm is

similar to the fast component of the NO₂, NO, and O veloc-

ity distributions observed in the work of Mueller et al.,⁸²and

2

4

˜

assigned to production of NO₂͑X A₁͒+NO₂͑1 B₂/1 A₂͒.

Note that the slow velocity component observed by Mueller

et al.⁸²assigned to the channel leading to NO₂͑X A ͒

2

˜

1

+NO₂͑2 ²B₂͒, is energetically inaccessible at the present

wavelength of 226 nm.

In their FTIR studies of emission from NO and NO₂

following the photodissociation of NO₂/N₂O₄at 248 nm,

Morrell et al. have drawn similar conclusions about the pos-

sible involvement of process ͑11͒.³²In these experiments, the

authors observed an emission band thought to be from NO

that became markedly more intense as the ambient condi-

tions were altered to increase the concentration of N₂O₄

͑while keeping the NO₂concentration approximately con-

stant͒. Although other processes could not be ruled out, it

was concluded that this emission band most likely arose

from direct dissociation of N₂O₄as indicated in Eq. ͑11͒. If

this were indeed the case, the quantum yield for the forma-

tion of NO in low vibrational levels by this mechanism ap-

peared to be signiﬁcant.

Evidence of direct NO formation from N₂O₄has also

been seen at various wavelengths around 200 nm by Parsons

et al.⁸³In these experiments, ﬂuorescence was observed from

NO͑A ²⌺⁺͒, which was suggested to be formed either by a

secondary photon absorption and subsequent dissociation of

the NO₂products of N₂O₄photolysis, or by absorption and

consequent emission of the NO product formed in Eq. ͑11͒.

A much earlier study of N₂O₄photolysis at 265 nm was also

seen to yield a large quantity of NO,⁸⁵with a quantum yield

of around 0.2 being suggested for the photolytic contribution

of Eq. ͑11͒ ͑the remainder of the NO being formed by sec-

ondary reaction of atomic oxygen fragments with NO₂͒.⁸⁴

It will be apparent from the above discussion that, al-

though previous studies of N₂O₄photodissociation do not

provide a universally consistent picture, the weight of

FIG. 11. ͑a͒ The velocity distribution for O͑³P₂͒ photofragments obtained

using high partial backing pressures of NO₂, and associated here with the

production of O atoms in the photodissociation of N₂O₄. ␷ and ␷ are the

12

13

recoil velocities of the NO₂fragments generated in step ͑12͒ and O͑³P͒ in

step ͑13͒, respectively. ͑b͒ ͑——͒ The kinetic-energy release distribution for

O͑³P₀͒ fragments obtained at a photolysis wavelength of 226 nm using a

low partial backing pressure of NO₂, and associated here with the photodis-

sociation of the NO₂monomer. Note that the data are plotted as a function

of NO internal energy. For comparison, the NO vibrational populations ob-

tained by Hancock and co-workers at 248 nm ͑¯

ࡗ

¯͒ ͑Ref. 32͒ and

193 nm ͑¯b¯͒ ͑Ref. 52͒, and the O͑³P₁͒ velocity obtained by Ahmed et

al. ͑----͒ ͑Ref. 46͒ at 212.8 nm are also shown. See text for details.

evidence^32,82–84seems to suggest that oxygen-atom products

may be formed either by direct photodissociation of the

dimer, or, more likely, by the secondary unimolecular disso-

ciation of internally excited primary NO₂photofragments. In

the light of this, together with the backing-pressure-

dependent data present here, it would appear that photolysis

of N₂O₄accounts for the low-velocity component seen in the

present probe-only images. The O͑³P_J͒ speed distributions

arising from photolysis of the NO₂monomer would thus

contain a single component that peaks at around 3000 m s⁻¹

͑see further discussion below͒.

Supporting evidence for the role of process ͑11͒ is pro-

vided by a more detailed consideration of the O͑³P₂͒ veloc-

ity distribution obtained at high NO₂backing-pressure con-

ditions, shown in Fig. 11͑a͒. This distribution can be crudely

characterized as a square-shaped distribution, centered at

064309-13

Photodissociation dynamics of NO₂

J. Chem. Phys. 124, 064309 ͑2006͒

1300 m s⁻¹with a maximum width of 700 m s⁻¹. The

193 nm data of Parsons et al. suggest that the production of

O͑³P͒ follows a two-step mechanism:

a preferential population of the lowest O͑³P₂͒ spin-orbit or-

bit channel is reminiscent of the near-adiabatic dissociation

observed for NO₂photodissociation via the 1 ²B₂state at

wavelengths around 355 nm,⁴⁴which, as discussed in Sec. I,

proceeds via internal conversion to the ground electronic

state. The spin-orbit populations observed here therefore pro-

vide some additional support for the notion that O-atom pro-

duction from N₂O₄follows the two-step mechanism de-

scribed above. By analogy with the long-wavelength

photodissociation of NO₂, it is suggested that the second step

in the dissociation of N₂O₄involves near threshold dissocia-

tion of a low-lying electronic state of NOЉ₂, perhaps the 1 ²B₂

state, subsequent to internal conversion to the ground elec-

tronic state. Secondary dissociation of NOЉ₂borne in the

1 ⁴B₂/1 ⁴A₂seems a somewhat less likely candidate, as dis-

sociation involving an excited quartet might be expected to

generate rather different spin-orbit propensities than for the

ground doublet state of NO₂.

N₂O₄+ h␯͑226 nm͒ → NO₂Ј + NOЉ₂,

NO₂Љ → NO + O͑³P_J͒.

͑12͒

͑13͒

Little kinetic energy is released in step ͑13͒, such that the

velocity distributions of the O and NO secondary fragments

essentially mirror those of the nascent NO₂photofragments

born in step ͑12͒. If the same were true at 226 nm, the inter-

pretation of the present O-atom data would be that the cen-

tral velocity of ϳ1300 m s⁻¹must be associated with the

recoil velocity of the nascent NO₂products formed in step

͑12͒, while the 700 m s⁻¹width must be associated with the

maximum kinetic-energy release in step ͑13͒. The latter as-

signment, together with the conclusions of Doughty et al.

that at 248 nm the NO fragments in step ͑13͒ are produced in

low vibrational states, between ␷=1 and 3, implies that the

dissociating NOЉ₂molecule must possess an internal energy

of at most 335 kJ mol⁻¹͑the sum of the dissociation energy

of NO₂, an estimate of the vibrational energy of NO of

ϳ25 kJ mol⁻¹, and the maximum total kinetic-energy release

associated with an O-atom velocity of ϳ700 m s⁻¹͒. This

internal energy could reside in electronic or rovibrational de-

grees of freedom, or a combination of the two, and would be

2. Photodissociation at low NO₂partial pressures

As noted in the previous subsection, at low partial back-

ing pressures of NO₂the velocity distributions for O͑³P₁͒

and O͑³P₀͒ can be assigned as arising from photodissocia-

tion of NO₂monomers at the probe-only wavelength. The

kinetic-energy release distribution for O͑³P₀͒ under these

conditions is shown in Fig. 11͑b͒. The implicit distribution of

cofragment NO vibrational states at this wavelength, peaking

around ␷=5, is clearly highly nonstatistical in nature, as ex-

pected from the previous results at a variety of photon ener-

gies above and below 226 nm.^{29–32,46,47,52}

˜

consistent with production of NOЉ₂in either of the X, the

1 B₂, or the 1 ²B₁electronic states ͓recall that production of

NO₂͑2 ²B₂͒ at 226 nm can be excluded on energetic

grounds͔. The 1 ⁴B₂/1 ⁴A₂electronic states, favored by

Mueller et al.,⁸²might also be involved, but the electronic

excitation energy from the ground to these excited states

͑T₀͒, which is at present unknown, would have to be less

than about 3 eV. With the above interpretation of the O-atom

velocity distribution, a central velocity of ϳ1300 m s⁻¹cor-

responds a total kinetic-energy release in step ͑12͒ of about

80 kJ mol⁻¹. This leaves only 60 kJ mol⁻¹of internal energy

in the NO₂Ј cofragment of the dissociating NO₂Љ species, in-

sufﬁcient either for the former to also undergo secondary

dissociation or for it to be produced in an electronically ex-

cited state. Our conclusion, therefore, is that step ͑12͒ can be

Velocity-map ion images of the O͑³P₀͒ products of NO₂

monomer photodissociation at 226 nm have been reported

previously by Ahmed et al.⁴⁷The resulting speed distribution

and translational anisotropy are in excellent accord with the

present low partial pressure data, although the speed distri-

bution reported here is somewhat sharper than that obtained

previously. ͓Note, however, that the energy scale on Fig. 2͑a͒

of Ref. 47 is incorrect.͔ Interestingly, both the present work

and that of Ahmed et al. at 226 nm ͑Ref. 47͒ indicate that the

O-atom velocity distribution arising from photolysis of

monomer NO₂is unimodal. This behavior should be con-

trasted with the shorter-wavelength studies of Ahmed et al.⁴⁶

at 212.8 nm and Hancock and Morrison⁵²at 193 nm, both of

which showed evidence for a bimodal NO vibrational state

distribution arising from photofragmentation of the NO₂

monomer ͑see Sec. IV C͒. The data reveal the opening of a

new dissociation pathway at wavelengths between 226 and

212.8 nm, although inspection of the absorption spectrum in

this region suggests that this new mechanism must be as-

cribed to exit channel dynamics, rather than to excitation into

a different electronic state.

˜

reﬁned as leading to NO₂͑X͒+NO₂Љ, with NO₂Љ borne in ei-

2

4

˜

ther the X, the 1 B₂, the 1 B₁, or ͑possibly͒ the 1 B₂/1 A₂

states, with a total internal energy of around 335 kJ mol⁻¹.

A particularly interesting feature of the current probe-

only data is the variation in fast and slow O-atom velocity

components with spin-orbit state. The current data consis-

tently show a greater intensity of the low speed versus the

high speed component for the O͑³P₂͒ products ͑this is

clearly evident in Fig. 8͒. The conclusion must be that the

process described by Eq. ͑11͒ generates very different

O-atom spin-orbit populations compared with the direct NO₂

monomer photodissociation at 226 nm. The latter process is

known to generate O atoms with a near-statistical spin-orbit

population distribution⁴⁴͑in the ratio 1.00:0.50:0.17 for J

=2, 1 and 0͒. Analysis of the data shown in Fig. 8 suggests

that the spin-orbit populations arising from process ͑11͒ must

therefore be close to 1.00:0.03:0.02 for J=2, 1, and 0. Such

The O͑³P₀͒ kinetic-energy release distribution arising

from NO₂photolysis at 226 nm deserves special comment.

Two aspects of this distribution are of interest. First, as

shown in Fig. 11͑b͒, it is a very sharply peaked distribution,

suggesting that the NO cofragments are born in a narrow

range of rovibrational states centered around ␷ =4–6. Sec-

Ј

ond, as demonstrated by the comparison with previous work

at other wavelengths shown in Fig. 11͑b͒, apart from the

064309-14

Brouard et al.

J. Chem. Phys. 124, 064309 ͑2006͒

bimodal feature that appears at shorter wavelengths, the po-

sition of the main peak in the associated vibrational distribu-

tion appears remarkably insensitive to photolysis wave-

length. The sharp, inverted NO internal state distribution is

indicative of population through a Franck-Condon-type

mechanism, associated either with the excitation to the

two-step dissociation mechanism involving initial production

Ã

˜

of NO₂͑X͒+NO₂, followed by subsequent unimolecular

dissociation of internally excited NO^Ã₂, possibly via the

˜

ground X.

ACKNOWLEDGMENTS

2

˜

2 B₂͑D B₂͒ state, or with the homogeneous predissociation

51

2

We gratefully acknowledge the Royal Society for the

award of a Royal Society Fellowship to one of the authors

͑C.V.͒, and the EPSRC for a research grant.

˜

of the 2 B₂state, possibly by the 1 B₂͑A B₂͒ state. The

2 ²B₂state has a similar bond angle to the ground state, but a

considerably extended N–O bond length,⁴⁰which would fa-

vor initial excitation to vibrationally excited stretching levels

¹G. E. Busch and K. R. Wilson, J. Chem. Phys. 56, 3626 ͑1972͒.

²G. E. Busch and K. R. Wilson, J. Chem. Phys. 56, 3639 ͑1972͒.

³A. Delon, R. Georges, and R. Jost, J. Chem. Phys. 103, 7740 ͑1995͒.

⁴J. Liévin, A. Delon, and R. Jost, J. Chem. Phys. 108, 8931 ͑1998͒.

⁵A. Delon, S. Heilliette, and R. Jost, Chem. Phys. 238, 465 ͑1998͒.

⁶H. Katagiri and S. Kato, J. Chem. Phys. 99, 8805 ͑1993͒.

⁷S. Y. Grebenshchikov, C. Beck, H. Flöthman, R. Schinke, and S. Kato, J.

Chem. Phys. 111, 619 ͑1999͒.

2

˜ ˜

of the 2 B₂state. Indeed, the D→X dispersed ﬂuorescence

˜

spectrum from the ͑0, 0, 0͒ level of the D state shows a

pronounced progression in the ␯ symmetric stretching

1

mode, which peaks in intensity at ␷Љ=5,⁹²in remarkable

1

agreement with the peak in the NO vibrational populations

inferred here for the photodissociation of NO₂via the same

electronically excited state. One puzzling feature is that, if

the NO vibrational populations were established in the exci-

tation step, one might expect from simple Franck-Condon

grounds that the position of the peak in the vibrational dis-

tribution might shift with photodissociation wavelength,

which it clearly does not. The alternative explanation for the

sharply peaked, inverted NO vibrational distribution is that it

is associated with the homogeneous coupling between the

⁸V. Kurkal, P. Fleurat-Lessard, and R. Schinke, J. Chem. Phys. 119, 1489

͑2003͒.

⁹A. Delon, R. Jost, and M. Jacon, J. Chem. Phys. 114, 331 ͑2000͒.

¹⁰S. Mahapatra, H. Köppel, L. S. Cederbaum, P. Stampfuss, and W. Wen-

zel, Chem. Phys. 259, 211 ͑2000͒.

¹¹R. F. Salzgeber, V. Mandelshtam, C. Schlier, and H. S. Taylor, J. Chem.

Phys. 109, 937 ͑1998͒.

¹²F. Santoro and C. Petrongolo, J. Chem. Phys. 110, 4419 ͑1999͒.

¹³M. Mons and I. Dimicoli, Chem. Phys. 130, 307 ͑1989͒.

¹⁴U. Robra, H. Zacharias, and K. H. Welge, Z. Phys. D: At., Mol. Clusters

16, 175 ͑1990͒.

51

2

¹⁵V. P. Hradil, T. Suzuki, S. A. Hewitt, and P. L. Houston, J. Chem. Phys.

99, 4455 ͑1993͒.

˜

D B₂and probably the A B₂states. If the avoided cross-

ing between these two surfaces occurred at extended N–O

bond lengths, homogeneous predissociation would be ex-

pected to occur over a relatively narrow range of conﬁgura-

tion space, that would be insensitive to the photolysis wave-

length.

¹⁶J. A. Harrison, X. Yang, M. Rösslein, P. Felder, and J. R. Huber, J. Phys.

Chem. 98, 12260 ͑1994͒.

¹⁷M. Hunter, S. A. Reid, D. C. Robie, and H. Reisler, J. Chem. Phys. 99,

1093 ͑1993͒.

¹⁸D. C. Robie, M. Hunter, J. L. Bates, and H. Reisler, Chem. Phys. Lett.

193, 413 ͑1992͒.

¹⁹P. T. Knepp, A. C. Terentis, and S. H. Kable, J. Chem. Phys. 103, 194

͑1995͒.

²⁰A. Doughty, G. Hancock, and E. L. Moore, Chem. Phys. Lett. 274, 58

͑1997͒.

V. CONCLUSIONS

²¹S. A. Reid, D. C. Robie, and H. Reisler, J. Chem. Phys. 100, 4256

͑1994͒.

The photodissociation of NO₂/N₂O₄mixtures has been

studied at 308 and 226 nm using REMPI detection of the

O͑³P_J͒ photofragments, coupled with velocity-map ion im-

aging. At the longer photolysis wavelength, where only NO₂

photofragmentation is involved, the velocity distributions of

O͑³P₂͒ and O͑³P₁͒ products suggested a nonstatistical popu-

lation distribution over the internal rovibrational quantum

states of the NO cofragments. The atomic fragments were

produced with very little electronic angular momentum po-

larization. Probe-only O-atom ion images were also recorded

at wavelengths around 226 nm. At this wavelength, the

O-atom velocity distributions were found to be bimodal, with

the relative intensities of the two components depending sen-

sitively on the NO₂partial backing pressure used in the

molecular-beam expansion, consistent with the competitive

production of O atoms from both photolysis of NO₂and

N₂O₄. The fast O-atom component, arising from the photoly-

sis of the monomer, possessed a velocity distribution corre-

sponding to an inverted, and sharply peaked NO cofragment

vibrational distribution, consistent with previous work on the

dissociation of NO₂via the second absorption band. The

slow O-atom products, generated in the photolysis of the

dimer, were found to be produced with a highly nonstatistical

spin-orbit state population distribution, consistent with a

²²S. A. Reid and H. Reisler, J. Chem. Phys. 101, 5683 ͑1994͒.

²³A. Sanov, C. R. Bieler, and H. Reisler, J. Phys. Chem. 99, 13637 ͑1995͒.

²⁴S. A. Reid and H. Reisler, J. Phys. Chem. 100, 474 ͑1996͒.

²⁵N. Changlong, L. Hua, and J. Pfab, J. Phys. Chem. 97, 7458 ͑1993͒.

²⁶R. P. Baker, M. L. Costen, G. Hancock, G. A. D. Ritchie, and D. Sum-

merﬁeld, Phys. Chem. Chem. Phys. 2, 661 ͑2000͒.

²⁷H. Zacharias, M. Geilhaupt, K. Meier, and K. H. Welge, J. Chem. Phys.

74, 218 ͑1981͒.

²⁸H. Zacharias, K. Meier, and K. H. Welge, in Energy Storage and Redis-

tribution in Molecules, edited by J. Hinze ͑Plenum, New York, 1983͒.

²⁹T. G. Slanger, W. K. Bischel, and M. J. Dyer, J. Chem. Phys. 79, 2231

͑1983͒.

³⁰C. B. McKendrick, C. Fotakis, and R. J. Donovan, J. Photochem. 20, 175

͑1982͒.

³¹J. McFarlane, J. C. Polanyi, and J. G. Shapter, J. Photochem. Photobiol.,

A 58, 139 ͑1991͒.

³²C. Morrell, C. Breheny, V. Haverd, A. Cawley, and G. Hancock, J. Chem.

Phys. 117, 11121 ͑2002͒.

³³H. Gaedtke, H. Hippler, and J. Troe, Chem. Phys. Lett. 16, 177 ͑1972͒.

³⁴M. Quack and J. Troe, Ber. Bunsenges. Phys. Chem. 83, 3462 ͑1972͒.

³⁵D. M. Wardlaw and R. A. Marcus, J. Chem. Phys. 83, 3462 ͑1985͒.

³⁶S. I. Ionov, G. A. Brucker, C. Jacques, Y. Chen, and C. Wittig, J. Chem.

Phys. 99, 3420 ͑1993͒.

³⁷S. I. Ionov, H. F. Davis, K. Mikhaylichenko, L. Valachovic, R. A. Beau-

det, and C. Wittig, J. Chem. Phys. 101, 4809 ͑1994͒.

³⁸S. J. Klippenstein and T. Radivoyevitch, J. Chem. Phys. 99, 3644 ͑1993͒.

³⁹P. I. Ionov, I. Bezel, S. I. Ionov, and C. Wittig, Chem. Phys. Lett. 272,

257 ͑1997͒.

064309-15

Photodissociation dynamics of NO₂

J. Chem. Phys. 124, 064309 ͑2006͒

⁴⁰K. E. J. Hallin and A. Merer, Can. J. Phys. 54, 1157 ͑1976͒.

⁴¹W. M. Uselman and E. K. C. Lee, J. Chem. Phys. 65, 1948 ͑1976͒.

⁴²L. Bigio, R. S. Tapper, and E. R. Grant, J. Phys. Chem. 88, 1273 ͑1984͒.

⁴³K. Tsukiyama, K. Shibuya, K. Obi, and I. Tanaka, J. Chem. Phys. 82,

1147 ͑1985͒.

⁶⁸B. Martínez-Haya, M. J. Bass, M. Brouard, C. Vallance, I. Torres, and J.

Barr, J. Chem. Phys. 120, 11042 ͑2004͒.

⁶⁹M. Brouard, R. Cireasa, A. P. Clark, T. J. Preston, C. Vallance, G. C.

Groenenboom, and O. S. Vasyutinskii, J. Phys. Chem. A 108, 7965

͑2004͒.

⁴⁴H. G. Rubahn, W. J. van der Zande, R. Zhang, M. J. Bronikowski, and R.

N. Zare, Chem. Phys. Lett. 185, 154 ͑1991͒.

⁷⁰B. Y. Chang, R. C. Hoetzlein, J. A. Mueller, J. D. Geiser, and P. L.

Houston, Rev. Sci. Instrum. 69, 1665 ͑1998͒.

⁴⁵J. W. Au and C. E. Brion, Chem. Phys. 218, 109 ͑1997͒.

⁴⁶M. Ahmed, D. S. Peterka, A. S. Bracker, O. S. Vasyutinskii, and A. G.

Suits, J. Chem. Phys. 110, 4115 ͑1999͒.

⁷¹A. S. Bracker, E. R. Wouters, A. G. Suits, and O. S. Vasyutinskii, J.

Chem. Phys. 110, 6749 ͑1999͒.

⁷²A. C. Kummel, G. O. Sitz, and R. N. Zare, J. Chem. Phys. 85, 6874

͑1986͒.

⁴⁷M. Ahmed, D. S. Peterka, and A. G. Suits, in Atomic and Molecular

Beams, edited by R. Campargue ͑Springer-Verlag, Berlin, 2001͒, p. 343.

⁴⁸F. Sun and R. F. C. G. P. Glass, Chem. Phys. Lett. 337, 72 ͑2001͒.

⁴⁹R. C. Richter, V. I. Khamaganov, and A. J. Hynes, Chem. Phys. Lett.

319, 341 ͑2000͒.

⁷³R. N. Dixon, J. Chem. Phys. 85, 1866 ͑1986͒.

⁷⁴T. P. Rakitzis, S. A. Kandel, A. J. Alexander, Z. H. Kim, and R. N. Zare,

J. Chem. Phys. 110, 3351 ͑1999͒.

⁷⁵T. P. Rakitzis and R. N. Zare, J. Chem. Phys. 110, 3341 ͑1999͒.

⁷⁶I. V. Hertel and W. Stoll, Adv. At. Mol. Phys. 13, 113 ͑1978͒.

⁷⁷M. C. H. N. van Vroonhoven and G. C. Groenenboom, J. Chem. Phys.

116, 1965 ͑2002͒.

⁵⁰H. S. Im and E. R. Bernstein, J. Phys. Chem. A 106, 7565 ͑2002͒.

⁵¹K. Tsuji, M. Ikeda, J. Awamura, A. Kawai, and K. Shibuya, Chem. Phys.

Lett. 374, 601 ͑2003͒.

⁵²G. Hancock and M. Morrison, Mol. Phys. 103, 1727 ͑2005͒.

⁵³M. Kawasaki, H. Sato, A. Fukuroda, T. Kikuchi, S. Kobayashi, and T.

Arikawa, J. Chem. Phys. 86, 4431 ͑1987͒.

⁷⁸Y. Mo and T. Suzuki, J. Chem. Phys. 109, 4691 ͑1998͒.

⁷⁹Imaging in Molecular Dynamics: Technology and Applications, edited by

B. J. Whitaker ͑Cambridge University Press, Cambridge, ͑2003͒.

⁸⁰M. Kawasaki, K. Kasatani, H. Sato, H. Shinohara, and N. Nishi, Chem.

Phys. 78, 65 ͑1983͒.

⁵⁴C. H. Hsieh, Y. S. Lee, A. Fujii, S. H. Lee, and K. Liu, Chem. Phys. Lett.

277, 33 ͑1997͒.

⁵⁵A. V. Demyanenko, V. Dribinski, H. Reisler, H. Meyer, and C. X. W.

Qian, J. Chem. Phys. 111, 7383 ͑1999͒.

⁸¹W. N. Sisk, C. E. Miller, and H. S. Johnston, J. Phys. Chem. 97, 9916

͑1993͒.

⁵⁶M. Brouard, P. O’Keeffe, D. M. Joseph, and D. Minayev, Phys. Rev. Lett.

86, 2249 ͑2001͒.

⁸²J. A. Mueller, M. L. Morton, S. L. Curry, J. P. D. Abbatt, and L. J. Butler,

J. Phys. Chem. A 104, 4825 ͑2000͒.

⁵⁷M. Bass, M. Brouard, A. P. Clark, and C. Vallance, J. Chem. Phys. 117,

8723 ͑2002͒.

⁸³B. F. Parsons, S. L. Curry, J. A. Mueller, P. C. Ray, and L. J. Butler, J.

Chem. Phys. 111, 8486 ͑1999͒.

⁵⁸M. Brouard, P. O’Keefe, and C. Vallance, J. Phys. Chem. A 106, 3629

͑2002͒.

⁸⁴H. S. Johnston and R. Graham, Can. J. Chem. 52, 1415 ͑1974͒.

⁸⁵H. H. Holmes and F. Daniels, J. Am. Chem. Soc. 56, 630 ͑1934͒.

⁸⁶T. C.Hall, Jr. and F. E. Blacet, J. Chem. Phys. 20, 1745 ͑1952͒.

⁸⁷A. M. Bass, A. E. Ledford, Jr., and A. H. Laufer, J. Res. Natl. Bur. Stand.,

Sect. A 80A, 143 ͑1976͒.

⁵⁹G. Hancock, P. J. Pearson, G. A. D. Ritchie, and D. F. Tibbetts, Phys.

Chem. Chem. Phys. 5, 5386 ͑2003͒.

⁶⁰V. K. Nestorov and J. I. Cline, J. Chem. Phys. 111, 5287 ͑1999͒.

⁶¹S. A. Reid, A. Sanov, and H. Reisler, Faraday Discuss. 102, 129 ͑1995͒.

⁶²O. L. A. Monti, H. Dickinson, S. R. Mackenzie, and T. P. Softley, J.

Chem. Phys. 112, 3699 ͑2000͒.

⁸⁸J. Miyawaki, T. Tsuchizawa, K. Yamanouchi, and S. Tsuchiya, Chem.

Phys. Lett. 165, 168 ͑1990͒.

⁶³J. Miyawaki, K. Yamanouchi, and S. Tsuchiya, J. Chem. Phys. 99, 254

͑1993͒.

⁸⁹B. L. G. Bakker, A. T. J. B. Eppink, D. H. Parker, M. L. Costen, G.

Hancock, and G. A. D. Ritchie, Chem. Phys. Lett. 283, 319 ͑1998͒.

⁹⁰J. Luque and D. R. Crosley, LIFBASE, Database and Simulation Program

͑v. 1.6͒; SRI International Report No. MP 99-009, 1999.

⁹¹NIST Atomic Spectra Database, Version 3, http://physics.nist.gov/

PhysRefData/ASD/index.html, 2005.

⁶⁴A. T. J. B. Eppink and D. H. Parker, Rev. Sci. Instrum. 68, 3477 ͑1997͒.

⁶⁵D. W. Chandler and P. L. Houston, J. Chem. Phys. 87, 1445 ͑1987͒.

⁶⁶M. J. Bass, M. Brouard, A. P. Clark, B. Martínez-Haya, and C. Vallance,

Phys. Chem. Chem. Phys. 5, 856 ͑2003͒.

⁶⁷M. Brouard, A. P. Clark, C. Vallance, and O. S. Vasyutinskii, J. Chem.

Phys. 119, 771 ͑2003͒.

⁹²K. Aoki, K. Hoshina, and K. Shibuya, J. Chem. Phys. 105, 2228 ͑1996͒.

Article Doi

The photodissociation dynamics of NO2 at 308 nm and of NO 2 and N2O4 at 226 nm

DOI: 10.1063/1.2166631

Source and publish data:

Authors:

Article abstract of DOI:10.1063/1.2166631

Full text of DOI:10.1063/1.2166631

Products guided by the article

R&D Labs maybe for 10102-43-9

Relevant to this article

Hot Product