Rotationally resolved photoelectron spectroscopy of hot N2 formed in the photofragmentation of N2O

Article abstract of DOI:10.1063/1.1370078

A study of the rotationally resolved photoelectron spectroscopy of hot N₂ formed in the photofragmentation of N₂O was performed. Resonance enhanced multiphoton ionization with photoelectron spectroscopy was used for the study. Two-photon excitation to the lowest Rydberg state followed by one-photon ionization was used for the photoelectron detection of the rotatinally hot N₂ states. Ejection of electrons in p and f partial waves dominated one-photon ionization.

Full text of DOI:10.1063/1.1370078

Rotationally resolved photoelectron spectroscopy of hot N 2 formed in the
photofragmentation of N 2 O
Anouk M. Rijs, Ellen H. G. Backus, Cornelis A. de Lange, Maurice H. M. Janssen, Kwanghsi Wang, and Vincent
McKoy
Citation: The Journal of Chemical Physics 114, 9413 (2001); doi: 10.1063/1.1370078
View online: http://dx.doi.org/10.1063/1.1370078
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/114/21?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Single-field slice-imaging with a movable repeller: Photodissociation of N2O from a hot nozzle
J. Chem. Phys. 141, 054201 (2014); 10.1063/1.4891469
The role of Rydberg states in photoionization of NO2 and (NO+, O−) ion pair formation induced by one VUV
photon
J. Chem. Phys. 139, 044311 (2013); 10.1063/1.4811713
Femtosecond time-resolved photoelectron-photoion coincidence imaging of multiphoton multichannel
photodynamics in N O 2
J. Chem. Phys. 128, 204311 (2008); 10.1063/1.2924134
Rotationally resolved photoionization dynamics of hot CO fragmented from OCS
J. Chem. Phys. 116, 2776 (2002); 10.1063/1.1434993
Detection of “ended” NO recoil in the 355 nm NO 2 photodissociation mechanism
J. Chem. Phys. 111, 5287 (1999); 10.1063/1.479806
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11

JOURNAL OF CHEMICAL PHYSICS
VOLUME 114, NUMBER 21
1 JUNE 2001
Rotationally resolved photoelectron spectroscopy of hot N2 formed
in the photofragmentation of N2O
Anouk M. Rijs
Laboratory of Physical Chemistry, University of Amsterdam, Nieuwe Achtergracht 127-129, 1018 WS,
Amsterdam and Laser Center and Department of Chemistry, Vrije Universiteit, De Boelelaan 1083,
1081 HV Amsterdam, The Netherlands
Ellen H. G. Backus and Cornelis A. de Langea)
Laboratory of Physical Chemistry, University of Amsterdam, Nieuwe Achtergracht 127-129, 1018 WS,
Amsterdam, The Netherlands
Maurice H. M. Janssen
Laser Center and Department of Chemistry, Vrije Universiteit, De Boelelaan 1083, 1081 HV Amsterdam,
The Netherlands
Kwanghsi Wang and Vincent McKoy
Arthur Amos Noyes Laboratory for Chemical Physics, California Institute of Technology, Pasadena,
California 91125
͑Received 13 February 2001; accepted 14 March 2001͒
The photoionization dynamics of rotationally hot molecular nitrogen are studied employing
resonance enhanced multiphoton ionization in combination with photoelectron spectroscopy.
Photodissociation of N2O at ϳ203 nm results in highly rotationally excited N2 fragments in
ϩ
g
1
X 1͚ (N , ϭ0,1) states and O atoms in the excited D state. Photoelectron detection of the
v
Љ Љ
2
ϩ
1
rotationally hot N states is performed by a two-photon excitation to the lowest a
͚
Rydberg
Љ
2
g
state followed by one-photon ionization. The large number of observed rotational levels, from N
Ј
ϩ
1
ϭ49 up to N ϭ94, results in improved rotational parameters for a
͚
( ϭ0). In addition,
v
Ј
Ј
Љ
g
ϩ
1
experimental and theoretical rotationally resolved photoelectron spectra of the a ͚ (
v
Љ
Ј
g
ϭ0,1;N ) state are presented. In these spectra only ⌬NϭNϩϪN ϭeven transitions are observed,
Ј
Ј
with a dominant ⌬Nϭ0 peak and rather weak ⌬NϭϮ2 peaks. The one-photon ionization is
dominated by ejection of electrons in p and f partial waves. The agreement between experimental
and calculated spectra is excellent. © 2001 American Institute of Physics.
͓DOI: 10.1063/1.1370078͔
I. INTRODUCTION
with Cs symmetry. Excited electronic states, thought to play
a role in the photofragmentation of N2O around 200 nm, are
Nitrous oxide (N2O) plays an important role in the
chemistry of the earth’s atmosphere. In the lower atmo-
sphere, nitrous oxide is a prominent greenhouse gas, and in
the stratosphere it contributes to the catalytic depletion of
ozone.1,2 the photolysis of atmospheric N2O by absorption of
solar radiation in the 200 nm range forms oxygen atoms in
1
Ϫ
1
A ͚ and B ⌬ ͑in Cϱ␯ symmetry͒, correlating to 1 1A and
˜
˜
Љ
2 1A , 2 1A , respectively ͑in C symmetry͒.4 The photodis-
Ј
Љ
s
sociation of N2O in the spectral region near 200 nm occurs
1
1
˜
mainly via the one-photon forbidden repulsive B ⌬ (2 A )
Ј
state, leading to N2 (X 1͚ϩ) and O (1D2).5,6 The 2 1A state
Ј
g
3
1
is highly anisotropic and has a minimum energy in the
the excited D2 state, and N2 in its electronic ground state.
The production of these very reactive oxygen atoms is an
important reason why even relatively small quantities of
N2O are still atmospherically relevant. In the stratosphere
N2O can subsequently undergo extensive oxidation reactions
with excited O(1D), to produce N2 and O2 , as well as NO,
and higher nitrogen oxides. In fact, photo-oxidation reactions
are an important source of the formation of stratospheric ni-
trogen oxides (NOx), which are now known to be crucial in
the ozone cycle. They play an important role in catalytic
ozone destruction processes.
Franck–Condon region at a bending angle of about 130°.5 As
the fragments separate, a strong torque will be exerted on the
N2 photofragments, resulting in rotationally ‘‘hot’’ N2 . Sev-
eral experimental techniques, including resonance-enhanced
multiphoton ionization ͑REMPI͒, either in conjunction with
mass-resolved ion detection,7 or ion imaging,5,8–12 have been
used to probe the product channels. All these studies of the
photodissociation of N2O have focused on elucidating the
details of the various photodissociation pathways, and on
measuring the energy and angular distributions of N2 .
REMPI of N2 is not a trivial matter, and relatively few
excited states have so far been utilized as stepping stones in
an (nϩm) REMPI process. The reasons for this are the high
ionization energy of ϳ15.6 eV, the large energy separation
between the electronic ground state and most excited states,
The molecule N2O is linear (Cϱ␯) in its electronic
ground state. Upon absorption of one or more photons, the
molecule may undergo a transition to a bent configuration
a͒
Electronic mail: deLange@fys.chem.uva.nl
0021-9606/2001/114(21)/9413/8/$18.00
9413
© 2001 American Institute of Physics
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11

9414
J. Chem. Phys., Vol. 114, No. 21, 1 June 2001
Rijs et al.
and the strict optical selection rules which arise from the
electron spectroscopy ͑PES͒, possesses unique capabilities.
The rules of energy and momentum conservation allow the
experimentalist to determine the internal energies of the ions
formed in the two-step photoabsorption process. The deter-
mination of ion internal energies ͑electronic, vibrational, and
rotational͒ is a true asset of the method, and laser photoelec-
tron spectroscopy offers very significant advantages, which
are not easily matched in its broad applicability by other
techniques.17–19
When the resolution of the electron spectrometer used is
sufficient, ion rotational levels can be resolved, and ion ro-
tational branching ratios can be measured. From angular mo-
mentum considerations very detailed information about the
photoionization dynamics of simple molecules can be ob-
tained, especially if these experiments can be compared with
the results of high-level ab initio theoretical calculations.20,21
With the REMPI method, a single rotational level of the
resonant intermediate state is selected, so that only a few
rotational levels of the ion are accessed. Rotational resolu-
tion in the ion is easily achieved, if the rotational constants
of the ionic state under consideration lead to ionic rotational
spacings large compared to the spectrometer resolution. This
is usually the case for diatomic hydrides, and a plethora of
rotationally resolved PE studies have been performed on
such light molecules. Representative cases are stable mol-
ecules such as diatomic hydrogen22–25 and its isotopomer
D2 ,26,27 and several short-lived diatomic hydride radicals
such as OH,28 NH,29,30 and SH.31,32
symmetry restrictions associated with this homonuclear di-
1
atomic molecule. The lowest Rydberg state a
͚
ϩ , located
Љ
g
about ϳ12 eV above the electronic ground state, and ac-
cessed by a two-photon (␭ϳ203 nm͒ allowed excitation of
an electron in the highest occupied 3␴ orbital to the
g
4␴g(nϭ3) Rydberg orbital, was used for the first time in a
͑2ϩ1͒ REMPI study13 as a convenient stepping stone. The
ϩ
g
1
a
͚
state represents the lowest member of a Rydberg
Љ
series converging upon the X 2͚ ionic ground state of N2ϩ .
ϩ
g
In this work the dominant Q branch was used in the inter-
ϩ
g
ϩ
g
1
mediate excitation a
͚
X 1͚ to study the subse-
Љ
quent photoionization process. From the Q-branch photoex-
citation information on the relative vibrational and rotational
populations of the electronic ground state can be obtained
directly. In a similar study high-resolution ͑2ϩ1͒ REMPI
spectroscopy was performed on the ͑0,0͒, ͑1,1͒, and ͑2,2͒
ϩ
g
ϩ
g
1
transitions of the a
͚
X 1͚ two-photon absorption
Љ
in N2 .14,15 In addition, accurate vibrational and rotational
ϩ
g
1
constants of the excited a
͚
Rydberg state were obtained
Љ
based on rotational levels up to J ϭ39.14
Ј
The experimental vibrational and rotational energy dis-
tributions of N2 , generated in the photofragmentation of
N2O, have been studied by also employing ͑2ϩ1͒ REMPI
ϩ
g
ϩ
g
1
via the a ͚ (␯ )
X 1͚ ͑␯Љ͒ transitions.7 The Q
Љ
Ј
branches were again found to dominate strongly over O and
S branches. It was concluded that the N2 fragments are pre-
dominantly formed in their electronic and vibrational ground
state, but with high rotational energies. The rotational energy
distribution peaks with a single maximum at rotational level
The present paper is concerned with rotationally re-
solved laser photoelectron spectroscopy of hot N2 photofrag-
ϩ
g
1
ments using the a
͚
intermediate state. The rotational
Љ
J ϭ74. Several peaks above J ϭ77 were also observed, but
Љ
Љ
constant of Nϩ2 in the X 2͚ ionic ground state is only
ϩ
g
remained unassigned. The authors suggested that these peaks
1.922 321 cmϪ1 33 However, the high rotational excitation of
.
ϩ
g
1
might be due to perturbations in the N a
͚
state.7
Љ
2
the N2 fragments following photodissociation of the N2O
parent accesses ionic rotational levels with high Jϩ, and
hence leads to a situation where the ionic rotational spacings
exceed the spectrometer resolution. The results of our experi-
ments are compared to those of advanced ab initio quantum-
chemical calculations, a strategy that has proved its worth in
many cases,28,30,31 and which gives detailed information
about the dynamics of the photoionization process. In addi-
tion, improved values of rovibrational parameters of the
In recent ion imaging experiments of the photodissocia-
tion of N2O at Sandia Livermore and VU Amsterdam5,9–11
state-specific angular and velocity distributions for the indi-
vidual rovibrational levels of the N2 photofragment and O
(1D) atom were obtained. In these studies rotationally ‘‘hot’’
N2 fragments were observed, again with a rotational distri-
bution peaking at J ϭ74. Furthermore, initial assignments of
Љ
ϩ
g
ϩ
1
the a
͚
X 1͚ transitions above J ϭ77 (␯ ϭ0),
Ј Ј
g
Љ
which were previously ascribed to perturbations in the
ϩ
1
ϩ
g
1
a
͚
Rydberg state are obtained.
Љ
g
a
͚
intermediate state,7 revealed the production of N2 in
Љ
In Sec. II a brief description of the experimental setup is
its first excited vibrational state (␯ ϭ1). By combining the
Љ
O (1D) and N2 measurements, velocity-resolved determina-
tions of the O (1D) orbital alignment and angular distribu-
tions were obtained. An analysis of the observed dynamical
effects was carried out,5,9,11 based on the results of theoreti-
cal calculations of the potential energy surfaces of N2O.4,16
These results were employed to estimate the branching ratio
given and in Sec. III a brief outline of the theoretical formu-
lation is given. The measured and calculated spectra are pre-
sented and discussed in Sec. IV. The conclusions are sum-
marized in Sec. V.
II. EXPERIMENT
between the two bent excited states. For the highest J levels
Ј
͑above J ϭ74), corresponding to the highest degree of
Ј
A description of the laser system and the ‘‘magnetic
bottle’’ spectrometer has been reported in great detail
recently.34 Therefore, the experimental setup is described
only briefly here.
The laser system consists of a XeCl excimer laser ͑Lu-
monics HyperEx 460͒, operating at a 30 Hz repetition rate,
which pumps a Lumonics HyperDye 500 dye laser ͑band-
width 0.08 cmϪ1), operating on Rhodamine B. The dye laser
bending of the parent N2O, the photodissociation occurs al-
5
1
most completely via the 2 1A (B ⌬) state.
˜
Ј
REMPI in combination with mass-resolved ion detection
has the disadvantage that no experimental information can be
obtained about the final state of the ion. Albeit experimen-
tally more demanding, REMPI in conjunction with kinetic-
energy-resolved electron detection, also termed laser photo-
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11

9415
J. Chem. Phys., Vol. 114, No. 21, 1 June 2001
Photofragmentation of N2O
output ͑around 610 nm͒ is frequency doubled using a Lu-
Second, for several resonances photoelectron spectra are
measured at a fixed excitation wavelength. For calibration of
both the wavelength and the photoelectron kinetic energies,
well-known resonances of krypton atoms in combination
with the oxygen atom resonances were used.
*
monics HyperTrack 1000, with an angle-tuned KD P crystal.
A zero-order half-lambda waveplate ͑CVI͒ is used to rotate
the polarization of the fundamental laser beam by 90°, while
maintaining the polarization of the frequency-doubled beam.
The rotated fundamental and the frequency-doubled laser
beams are sum-frequency mixed in an angle-tuned BBO
crystal to generate ultraviolet light around 203 nm. The fun-
damental and the harmonic beams are spatially separated by
a Pellin–Broca prism to yield a vertically polarized 203 nm
beam. The typical pulse energy used in these experiments is
0.1 ␮J. During the wavelength scans the power of the 203
nm laser beam is monitored with a photodiode. The laser
light is focused into the ionization region of the ‘‘magnetic
bottle’’ spectrometer by a quartz lens with a focal length of
25 mm. In the present experiments the ‘‘magnetic bottle’’
spectrometer is used exclusively in the electron detection
mode, with a collection efficiency of ϳ50%.
The rotationally hot N2 photofragments are generated in
situ by one-photon dissociation of neat N2O ͑Matheson͒. The
N2O gas is effusively introduced into the ‘‘magnetic bottle’’
spectrometer. The photodissociation of N2O as well as the
subsequent ͑2ϩ1͒ REMPI of the ‘‘hot’’ N2 fragments, is
performed with the same laser photons of 203 nm.
Two kinds of experiments are performed. First, the reso-
nant excited states are located by scanning the excitation
wavelength and monitoring the electron current of N2 , se-
lecting a small kinetic-energy window of photoelectrons.
III. THEORY AND NUMERICAL DETAILS
The general theory of molecular REMPI processes used
in the present study has been described previously.35 Here
we present just a brief outline of some essential features as it
is used to obtain rotationally resolved photoelectron spectra.
In the present case, ionization originating from each of the
ϩ
g
1
(2Jϩ1) magnetic sublevels of the a
͚
state of N2 forms
Љ
an independent channel. The total cross section ␴ for ioniza-
tion of a J level of the intermediate state leading to a Jϩ
level of the ion can be written as
2
␴ϰ
␳
͉Clm M ,M ϩ͒
J
͉
,
͑1͒
͑
͚
lm
M M
J
J
J
ϩ
M
,M
J
J
where ␳
is the population of a specific MJ level of the
J
M M
J
intermediate state. The coefficients Clm(MJ ,MJϩ) of Eq. ͑1͒
are related to the probability for photoionization of the MJ
level of the intermediate state leading to the MJϩ level of the
ionic state. The Clm(MJ ,MJϩ) coefficients, which explicitly
consider the spin coupling associated with multiplet-specific
final state wave functions and a Hund’s case ͑b͒ coupling
scheme, have the form
4␲
1/2
Clm M ,M ϩ͒ϭ
ͱ
2Nϩϩ1͒ 2Nϩ1͒ 2Jϩ1͒ 2Jϩϩ1͒ 2Sϩ1͒
͑
͓͑
͑
͑
͑
͑
͔
J
J
3
N
S
J
Nϩ
Sϩ
Jϩ
Sϩ 1/2
S
Nϩ
N
Nt
mt
P
ϫ
Ϫ1͒ 2N ϩ1͒
͑
t
͑
ͩ
ͪͩ
ͪͩ
ͪͩ
ͪ
͚
M
MS ϪMJ Mϩ MS
ϩ
ϪMJ
ϩ
MS
ϩ
M
l
ϪMS ϪMϩ
M
␴
Nt
1
l
Nϩ
N
Nt
Nt
1
ϩ
Ϫ⌳ Ϫ␮
˜
ϫ
Ϫ1͒
Il␭␮
,
͑2͒
͑
ͩ
ͪ
ͩ
ͪͩ
ͪ
͚
Ϫmt
␮
Ϫm
Ϫ⌳ϩ
⌳
␭
Ϫ␭
t
␮
Ϫ␭
␮
t
0
t
␭
with
PϭMϩϩMJϩϪMϩ␮0ϪNϪNϩϩsϩ 12,
vibrationally averaged photoelectron matrix element between
the resonant state and the photoelectron continuum wave
͑3͒
͑4͒
function, and ␮ and ␮ are the light polarization index in the
0
molecular and laboratory frame, respectively. The symbols
with superscript ϩ are related to the same quantities in the
ion. In Eq. ͑5͒, we have employed single-center expansions
of a partial wave component of the photoelectron orbital,
Il␭␮ϭ Ϫi͒lei␩ dR ␹ R͒rl␭␮ R͒␹ R͒,
˜
l
*
͑
͑
͑
͑
␷
i
͵
ϩ
fi
␷
l␭␮
l,l ␭
l ␭
␺
(Ϫ)(r) and of the initial orbital ␾ (r ), where ionization
Ј
i
klm
ˆ
r
R͒ϭ
g
k,r,R͒Y
r ͒
͉
rY1␮
͉
͑
͑
͑
Ј
͗
͚
fi
Ј
Ј
l ,l
Ј
0
originates, i.e.,
ˆ
r ͒ ,
͑
Ј
ϫ␾ r,R͒Y
l ,␭
0 0
͑5͒
͑
͘
il
0
͑Ϫ͒
klm
ˆ
␺
ϭ
g
k,r͒Dl Yl␭ r͒,
͑6͒
͑
͑
͚
l,l ␭
m␭
Ј
l ␭
where ␭ and m are projections of the photoelectron’s angular
momentum l in the molecular and laboratory frames, respec-
tively, ⌳ is the projection of electronic orbital angular mo-
mentum along the internuclear axis, S is the total spin, R
Ј
with
1/2
2
͑Ϫ͒
f,k
͑Ϫ͒
il␺
͑7͒
*
r͒Y k͒,
͑ ͑
klm lm
⌿
ϭ
ͩ ͪ
͚
˜
␲
denotes a dependence on internuclear distance, Il␭␮ is the
lm
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11

9416
J. Chem. Phys., Vol. 114, No. 21, 1 June 2001
Rijs et al.
TABLE I. Comparison of rotational constants (cmϪ1) of the lowest excited
superimposed on top of a smooth continuum, especially in
the region around the peak of the absorption band ϳ182 nm.
Around 203 nm the spectrum is not showing any structure at
room temperature. We expect that under these conditions,
over the 1.5 nm range studied, the photodynamics of N2O in
the effusive beam will not be strongly influenced.
ϩ
1
Rydberg state a
͑see Ref. 12͒.
͚
(
ϭ0) of N from this work and from the literature
vЈ
2
Љ
g
Constants
Present work (cmϪ1
)
Ref. valuesa (cmϪ1
)
98 840.41͑9͒
1.91439͑4͒
6.143͑2͒
98 840.59͑0.12͒
1.9143͑2͒
6.6͑2͒
v00
B0Ј
D0Јϫ106
The molecular ground and excited states involved in our
experiments, as well as the lowest ionic state, all have ͚
symmetry. Therefore, these states can be conveniently
treated in a Hund’s case ͑b͒ coupling ͓Eq. ͑2͔͒, in which we
only focus on the end-over-end rotation N , N , and Nϩ.
aSee Ref. 13.
(-)
f,k
Љ
Ј
where ⌿ is the photoelectron continuum orbital. Equation
Rotational structure in the photoelectron spectra of a homo-
nuclear diatom, ionized in a ͑2ϩ1͒ multiphoton absorption
process, is subject to rotational selection rules ͓Eq. ͑8͔͒.39
Selection rules predict that for one-photon ionization from
͑2͒ yields the selection rule
⌬Nϩlϭodd,
͑8͒
for photoionization of these ͚–͚ transitions. Since the 4␴
g
ϩ
g
1
the a ͚ ( ;N ) levels, only ⌬Nϩlϭodd transitions are
v
Љ
Ј
Ј
orbital contains mainly s character, ͑94.22% s, and 5.58% d͒,
the p wave component of the photoelectron matrix element
should be dominant ͑there are no even partial waves͒, so
only even ⌬N transitions are expected.
allowed, where ⌬NϭNϩϪN , and l is a partial wave com-
Ј
ponent of the photoelectron. In an atomic-like picture, re-
moval of the 4␴g(nϭ3) Rydberg electron is expected to
lead to p (lϭ1) partial waves and, hence, only ⌬Nϭeven
transitions would be expected in the photoelectron spectra.
Moreover, additional selection rules for homonuclear di-
atomics, based on inversion symmetry, also predict odd par-
ϩ
g
1
For the wave function of the a
͚
resonant state, we
Љ
use the improved virtual orbital method36 in which the core
ϩ
2
orbitals are taken to be those of the fully relaxed ͚ ion
g
and the 4␴ resonant orbital is obtained as an eigenfunction
g
ϩ
g
1
of the static-exchange potential of the core. The orbital basis
used in these calculations consists of a ͓4,3͔ contraction of
the ͑9,5͒ primitive Gaussian basis37 augmented by one s
function (␣ϭ0.08), one p function (␣ϭ0.075), and two d
functions (␣ϭ1.4836, and 0.4691͒ on the nitrogen atom. At
the molecular center, we add three s functions (␣ϭ0.06,
0.02, and 0.008͒, three p functions (␣ϭ0.035, 0.012, and
0.005͒, and two d functions (aϭ0.12 and 0.025͒. With this
basis and choice of wave functions, we obtain a self-
consistent-field energy of Ϫ108.532 032 9 a.u. at Rϭ2.12
tial waves for the one-photon ionization from the a
͚
to
Љ
ϩ
the X 2͚ ionic state.39
g
Conservation of angular momentum also requires ⌬N
ϭlϩ1,l,...,ϪlϪ1, where l is an angular momentum com-
ponent of the photoelectron. Combined with selection rules
͓Eq. ͑8͔͒, the allowed transitions for a p wave are limited to
⌬Nϭ0,Ϯ2, while for an electron leaving in an f wave ⌬N
ϭ0,Ϯ2,Ϯ4 holds. All our observations agree with these pre-
dictions.
In the assignments of our photoelectron spectra, the bo-
ϩ
g
1
a.u. for the a
͚
state. In this basis set, the ground state
son nuclear spin statistics of 14N 14N, the most abundant iso-
Љ
Hartree–Fock energy is Ϫ108.976 61 a.u. at the experimen-
tal equilibrium geometry of Rϭ2.068 a.u. The energy differ-
ence between the calculated ground and excited state ener-
gies is 12.09 eV. This is in good agreement with the
ϩ
g
1
topomer of N , must also be considered. In the a
͚
ex-
Љ
2
cited states of molecular nitrogen, even values of N can
Ј
only exist in conjunction with nuclear spin wave functions
which are symmetrical under permutation of the spins. Con-
experimental ϭ0 band origin of 12.25 eV ͑see also
v
versely, odd values of N can only exist in combination with
Ј
Table I͒.
nuclear spin functions which are antisymmetrical under per-
For the final state we assume a frozen-core Hartree–
Fock model in which the core orbitals are taken to be those
mutation. The ratio of symmetrical to antisymmetrical
ϩ
nuclear spin functions is 2:1. In the X 2͚ ionic ground state
ϩ
g
2
g
of the ͚ ion and the photoelectron orbital is a solution of
of molecular nitrogen, even Nϩ levels exist in combination
with symmetrical, odd Nϩ levels with antisymmetrical
nuclear spin functions. Therefore, in our experiments either
even or odd Nϩ levels are expected to be accessed in the
ionic state.
¨
the one-electron Schrodinger equation. We obtain the photo-
electron orbitals used in these calculations numerically em-
ploying an iterative procedure to solve the Lippmann–
Schwinger equation associated with the one-electron
¨
Schrodinger equation. The radial grid used in the numerical
At 203 nm, the photodissociation of N2O produces rota-
tionally hot N2 photofragments, mostly in its vibrational
ground state.6,7 Figure 1 shows a two-photon excitation spec-
trum of the N2 photofragments in the two-photon energy
range 98 100–98 700 cmϪ1. The spectrum was obtained by
monitoring photoelectrons with a kinetic energy of about 2.8
integration of these equations extended out to 64 a.u. and
contained 800 points.
IV. RESULTS AND DISCUSSION
Since all the experiments are carried out with one laser
only, the photodissociation wavelength is not constant.
Therefore, both the internal state distribution and the total
yield of N2 may vary as the wavelength is scanned. N2O is
known to possess an absorption spectrum extending from
220 to 170 nm.38 The cross section shows distinct structure
eV. These photoelectrons arise from the ͑2ϩ1͒ photoioniza-
ϩ
g
tion process, starting from the X 1͚
(
ϭ0,1;J ) ground
v
Љ Љ
state to the X 2͚ϩ( ϩϭ0,1;Nϩ) lowest ionic state, via the
v
g
ϩ
1
a
͚ ( ϭ0,1;N ) intermediate Rydberg state. All the ob-
v
Ј Ј
g
Љ
served transitions are of Q type (⌬Nϭ0). The rotationally-
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11

9417
J. Chem. Phys., Vol. 114, No. 21, 1 June 2001
Photofragmentation of N2O
ϩ
g
1
FIG. 1. Two-photon excitation spectrum of the Q branch of the a
͚
Љ
ϩ
1
X
͚
transition of hot N2 photofragments in the two-photon energy
g
range 98 100–98 800 cmϪ1 ͑vacuum energies͒, obtained by monitoring pho-
toelectrons with a kinetic energy of approximately 2.8 eV.
resolved wavelength spectrum reflects the rotational distribu-
ϩ
g
1
tion of N in the a
͚
state, extending from N ϭ49 up to
Ј
Љ
2
N ϭ94. At low N (N ϭ49–77) an even/odd intensity al-
Ј
Ј
Ј
ternation is observed, with transitions originating from even
more intense due to nuclear spin statistics. The comb
N
Ј
above the spectrum in Fig. 1 presents the assignment of the
N rotational levels of the Q-branch transitions arising from
Ј
X 1͚ϩ( ϭ0,1;N ).
v
Љ
Љ
g
Our excitation spectrum is in good agreement with the
spectrum recorded by Hanisco and Kummel,7 by Neyer
et al.,9 and by Teule.11 Our spectrum is measured with an
effusive beam of pure N2O, while the other groups employed
pulsed molecular beams. Hanisco and Kummel7 used a mo-
lecular beam of He with 35% N2O, Neyer et al.9 a pure
beam, and Teule11 a hexapole state-selected seeded beam.
This indicates that the use of an effusive beam has a negli-
gible effect on the N2 distribution at low N levels (N
Ј
Ј
Ͻ77). However, below 98 375 cmϪ1 our spectrum shows
much more structure.
Towards the red end of the spectrum ͑below 98 375
cmϪ1) some irregularities in the spectrum arising from
ϩ
g
N2 X 1͚
(
ϭ0) appear. Hanisco and Kummel7 were not
v
Љ
able to assign these multiple peaks, and suggested that they
ϩ
g
1
might be due to local perturbations in the N a
͚
state.
Љ
2
Neyer et al.,9 using an ion-imaging technique, assigned these
ϩ
g
1
FIG. 2. ͑a͒ Fit of observed rotational level energies for the a
͚
(
Љ
vЈ
peaks in part as a superimposed structure of rotational states
ϭ0) Rydberg state, based on the rotational constants of the vibrationless
ϩ
arising from X 1͚
(
Љ
ϭ0) and overlapping transitions
v
Љ
ϩ
1
ground state X
͚
(
ϭ0) of N ͑Ref. 18͒, and on the improved constants
vЉ
2
g
g
from X 1͚ϩ( ϭ1;N ). By using REMPI in combination
ϩ
1
for the a
͚
(
ϭ0) state given in Table I. ͑b͒ Difference between ex-
Љ
vЈ
v
Љ
g
g
ϩ
g
1
perimental and calculated line positions for the a
state; ͑c͒ similar to ͑a͒ but for the a
͚
(
ϭ0) Rydberg
vЈ
Љ
with kinetic-energy-resolved electron detection, we were
able to obtain a much more structured and resolved excita-
tion spectrum, especially at the red end of the spectrum.
Therefore, we have been able to assign the perturbed peaks
ϩ
g
1
͚
(
ϭ1) Rydberg state; ͑d͒ similar
vЈ
Љ
ϩ
1
to ͑b͒ but for the a
͚
(
ϭ1) Rydberg state.
vЈ
Љ
g
of N ( ϭ0), N Ͼ77 up to N ϭ94. We also assigned the
overlapping transitions of N2 in the first vibrationally excited
state ͑1,1 transition͒.
The assignment of the peaks in Fig. 1 and the rotational
analysis were carried out using the published line positions
v
Ј
Ј
Ј
2
͑low N values͒ of Hanisco and Kummel,14 by assigning the
measured rotational lines in an iterative fashion. By using the
published ground state rotational constants,40 the excited
state rotational constants were determined by fitting all the
Ј
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11

9418
J. Chem. Phys., Vol. 114, No. 21, 1 June 2001
Rijs et al.
TABLE II. Line positions (cmϪ1
)
of the rotational levels of the
( ϭ0) Q-branch transitions. The left column
vЉ
ϩ
ϩ
1
1
a
͚
(
ϭ0)
X
͚
Љ
vЈ
g
g
shows the measured line positions, the middle column the calculated line
positions, obtained with the improved rotational constants given in Table I.
The right column shows the differences between experimental and calcu-
lated line positions.
͑0,0͒ transition
Measured Q(N )
͑1,1͒ transition
Measured Q(N ) Fit Q(N )
N
Fit Q(N )
Ј
Ј
Ј
Ј
Ј
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
98 654.73
98 647.11
98 639.34
98 630.82
98 623.05
98 614.38
98 605.86
98 596.44
98 588.4
98 578.38
98 569.53
98 560.26
98 549.82
98 541.00
98 531.28
98 520.36
98 511.09
98 500.59
98 490.09
98 479.17
98 468.25
98 457.30
98 445.57
98 435.49
98 422.86
98 411.13
98 398.98
98 390.55
98 375.01
98 362.8
98 349.84
98 336.34
98 325.00
98 310.72
98 297.16
98 282.88
98 270.70
98 256.00
98 244.66
98 228.22
98 213.16
98 197.98
98 184.18
98 169.00
98 155.23
98 138.01
98 653.92
98 646.21
98 638.34
98 630.30
98 622.11
98 613.75
98 605.22
98 596.53
98 587.68
98 578.66
98 569.47
98 560.12
98 550.6
98 540.91
98 531.06
98 521.03
98 510.84
98 500.47
98 489.94
98 479.23
98 468.36
98 457.30
98 446.08
98 434.68
98 423.11
98 411.36
98 399.44
98 387.34
98 375.06
98 362.61
98 349.97
98 337.16
98 324.17
98 310.99
98 297.64
98 284.1
¯
¯
¯
¯
98 427.96
98 416.59
98 426.83
98 417.94
98 408.89
98 399.68
98 390.30
98 380.75
98 371.04
98 361.17
98 351.13
98 340.93
98 330.56
98 320.04
98 309.35
98 298.50
98 287.48
98 276.31
98 264.97
98 253.47
98 241.82
98 230.00
98 218.02
98 205.89
98 193.59
98 181.14
98 168.53
98 155.76
98 142.84
98 129.76
¯
¯
¯
¯
98 382.99
98 371.59
98 359.44
98 349.84
98 340.63
98 331.3
98 321.28
98 308.08
FIG. 3. Measured ͑left͒ and calculated ͑right͒ rotationally resolved photo-
electron spectra of rotationally excited N2 , generated in the photofragmen-
tation of N2O. The spectra are obtained by ͑2ϩ1͒ photoionization via the
ϩ
1
Q(72) transition of the intermediate a
͚
(
ϭ0) state.
vЈ
Љ
g
¯
98 285.34
¯
of the third order constant (HЈ) did not provide a significant
v
98 266.06
improvement of the rotational fit. The measured and calcu-
lated line positions are listed in Table II. The differences
between experimental and calculated line positions are pre-
sented in Fig. 2͑b͒ ͑0,0 transition͒ and 2͑d͒ ͑1,1 transition͒. A
¯
98 242.38
¯
98 219.88
98 207.25
local perturbation at N ϭ76( ϭ0) can be clearly seen in
v
Ј
Ј
¯
the difference plot in Fig. 2͑b͒. Also, perturbations at N
Ј
¯
ϭ87,90,93( ϭ0) can be observed. They may be caused
v
Ј
98 167.77
by the interfering vibrationally excited ͑1,1͒ Q-branch tran-
sitions. These line positions were left out in the least-
squares-fitting procedure. Also, somewhat larger deviations
are observed in the difference plot ͓Fig. 2͑d͔͒ of the ͑1,1͒ Q
branch.
¯
¯
98 129.01
¯
¯
¯
¯
¯
Figure 3 presents the measured ͓Fig. 3͑a͔͒ and calculated
͓Fig. 3͑b͔͒ photoelectron spectra obtained for ͑2ϩ1͒ photo-
ionization via the ͑0,0͒ Q͑72͒ rotational transition at
¯
¯
¯
¯
¯
¯
98 435.49 cmϪ1, leading to the X 2͚ϩ( ϩϭ0,Nϩ) ionic
98 270.38
98 256.47
98 242.38
98 228.11
98 213.64
98 199.00
98 184.16
98 169.13
98 153.92
98 138.51
¯
¯
v
g
¯
¯
ground state of Nϩ2 . The dominant peak in the photoelectron
¯
¯
spectrum at about 2.78 eV is the ⌬Nϭ0 transition of the Q
branch. The neighboring peaks arise from ⌬NϭϮ2 transi-
tions. No photoelectron peaks are observed for ⌬NϭϮ1
transitions due to the selection rules for N2 .34 The rotational
spacings between the ⌬Nϭ0 and ⌬Nϭϩ2 and Ϫ2 peaks
are Bϩ(4Nϩϩ6)͑70 meV͒ and Bϩ(4NϩϪ2)͑68 meV͒, re-
spectively. Here Bϩ is the ionic rotational constant and Nϩ
signifies the end-over-end rotation. Higher order rotational
contributions were left out because their effect is small com-
pared to the resolution of the magnetic bottle spectrometer
͑8–10 meV͒. The calculated spectrum is convoluted with a
full width at half maximum of 18 meV. Our measurements
indicate that in this wavelength region ionization only takes
¯
¯
¯
¯
¯
¯
¯
¯
¯
¯
¯
¯
¯
¯
rotational energies of the Q-branch transitions to the follow-
ing expression:
⌬E N ͒ϭ
ϩ BЈϪBЉ͒N N ϩ1͒
͑
vv
͑
͑
Ј Ј
Ј
v
Ј Љ
v
v
place to the X 2͚ϩ( ϩϭ0) ionic state. No transitions to
v
Ϫ DЈϪDЉ͒N N ϩ1͒2.
͑9͒
2
g
͑
͑
Ј
Ј
v
v
higher vibrational levels or transitions to other ionic states
The results of a least-squares-fitting procedure, including our
were observed. Photoelectron spectra of N2 produced in its
first vibrational level are also measured. These spectra show
similar behavior as observed for the ͑0,0͒ Q͑72͒ rotational
transition shown in Fig. 3, with one dominant peak (⌬N
new high N transitions, are shown in Fig. 2͑a͒ ͑0–0͒ tran-
Ј
sition͒ and 2͑c͒ ͑1–1 transition͒. The improved rotational
ϩ
g
1
constants for a ͚ ( ϭ0) are listed in Table I. Addition
v
Љ
Ј
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11

9419
J. Chem. Phys., Vol. 114, No. 21, 1 June 2001
Photofragmentation of N2O
ϭ0) and smaller ⌬NϭϮ2 peaks on either side. Ionization of
ϩ
g
ϩ
g
1
N via a
2
͚
(
ϭ1) takes place to X 2͚
(
ϩϭ1), again
v
v
Љ
Ј
showing ⌬ ϭ0 propensity.
v
Even though the calculated photoelectron spectra shown
in Fig. 3͑b͒ are for N ϭ72, the spectra originating from
Ј
other N in this experiment are similar since at these high N
Ј
levels, contributions from the 3-j symbols in Eq. ͑2͒ do not
play an important role, i.e., they have reached the high N
limit. In the calculated spectra, the ⌬NϭϮ2 peaks show
almost equal intensity due to the high N value and the small
difference in photoelectron kinetic energy. As can be seen in
Figs. 3͑a͒ and 3͑b͒ the agreement between measured and
calculated spectra is excellent. Both spectra show the same
important features, viz. a dominant peak resulting from the
⌬Nϭ0 transition, two small peaks from the ⌬NϭϮ2 tran-
sitions, and no peaks from the ⌬NϭϮ1,Ϯ3 transitions. Vir-
tually zero intensity was calculated for the ⌬NϭϮ4 transi-
tions, in agreement with the fact that these transitions are not
visible in the measured photoelectron spectra. The spectra
were calculated assuming alignment appropriate to a two-
photon excitation as well as for an isotropic population of
MJ levels. These two resulting spectra were very similar and
indistinguishable on the scale of our figures. Furthermore,
since the optical transition responsible for the dissociation of
N2O is parallel ͑about 97%, Ref. 5͒, the N2 molecule is
expected to begin rotating in a plane containing the polariza-
tion of the laser, and one would expect the population in the
MJϭ0 level to be dominant. To explore this effect we also
calculated the spectra assuming only the MJϭ0 level to be
populated. The ⌬Nϭ2/⌬NϭϪ2 peaks were somewhat
larger, 0.348/0.353, with this assumption (MJϭ0) than for
the case of two-photon induced alignment, 0.208/0.213 rela-
tive to a ⌬Nϭ0 peak intensity of 1. The fact that the experi-
mentally observed intensity for the ⌬NϭϮ2 peaks is
smaller than the value calculated assuming complete align-
ment in MJϭ0 cannot originate from a loss of such align-
ment by a single collision in the time between dissociation of
N2O and detection of N2 . This time is maximally the 20 ns
duration of the laser pulse and with the number density at 1
mTorr in the crossing region this would imply a vanishing
collision probability (ϳ10Ϫ6) for a typical gas collision
cross section. Whatever its cause, it seems that this expected
strong alignment in MJϭ0 of the N2 molecule does not
manifest itself in these experiments as agreement with spec-
tra calculated assuming isotropic distribution appears better.
As we have mentioned earlier, based on the atomic
model, we would predict rather weak ⌬NϭϮ2 and Ϯ4 in-
tensities for a photoelectron carrying mostly the p wave. In-
deed, this is the case for both measured and calculated spec-
tra, suggesting that the underlying photoelectron dynamics
should be very atomic-like. Close examination of the photo-
electron matrix elements, however, reveals that the photoion-
ization shows significant molecular character. Figure 4
FIG. 4. Magnitude of the partial wave components of the photoelectron
ϩ
D(lϪ)
͉
for photoionization of the a
͚
resonant state of
1
matrix element
N2 leading to the X
͉
Љ
g
ϩ
2
͚
state of Nϩ2 via the 4␴g→k␴ ͑a͒ and 4␴
u
g
g
→k␲ ͑b͒ ionization channels.
u
˜
magnitude of only one of the ␮ components of Il␭␮ of Eq.
͑4͒. One sees that at all energies the f (lϭ3) partial wave
matrix element is always larger than that of the p(lϭ1)
wave in both ionization channels. With the dominant s char-
acter of the 4␴ orbital, atomic-like expectations would sug-
g
gest that the p wave contribution should be larger than that of
the f wave. One would further expect, on the basis of con-
servation of angular momenta, that an f wave photoelectron
matrix element of this magnitude would lead to a strong
intensity for ⌬NϭϮ2 transitions. However, we see rela-
tively weak transitions for these peaks. This suggests that
interference among partial waves in both ionization channels
and the unfavorable 3-j symbols result in the weak intensities
seen for these ⌬NϭϮ2 peaks. These f partial waves arise
from the coupling of angular momenta of the photoelectron
due to torques associated with the molecular ion potential.
V. CONCLUSIONS
The photoionization dynamics of highly rotationally ex-
cited N2 fragmented from the photodissociation of N2O have
been studied by rotationally resolved photoelectron spectros-
shows the magnitude ͉ ͉ of the ͓incoming-wave normal-
D(lϪ)
1
copy. The photodissociation of N2O via the bent ⌬ excited
ized, see Eq. ͑6͔͒ partial wave components of the photoelec-
state results in rotationally hot N2 photofragments, mostly in
the vibrational ground level. A rotational distribution of N2
extending from N ϭ49 up to N ϭ94 is observed and as-
Љ Љ
signed for the vibrational ground level. Moreover, we were
able to assign some of the peaks at the red end of the exci-
tron matrix elements as a function of photoelectron kinetic
energy for the k␴ channel ͑a͒ and k␲ channel ͑b͒.
͉
D(lϪ)
͉
u
u
˜
˜
is defined as
͉
Il00
͉
for the k␴ channel and
͉
Il11
͉ for the k␲
u
u
channel where Il␭␮ is given in Eq. ͑4͒. Note that
͉
D(lϪ)
͉
is the
˜
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11

9420
J. Chem. Phys., Vol. 114, No. 21, 1 June 2001
Rijs et al.
16 M. H. M. Janssen, J. M. Teule, D. W. Neyer, D. W. Chandler, and G. C.
Groenenboom, Faraday Discuss. 113, 473 ͑1999͒.
tation spectrum as arising from N2 produced in the first vi-
brational level. The rotationally resolved photoionization
17 C. A. de Lange, in High Resolution Laser Photoionization and Photoelec-
tron Studies, edited by I. Powis, T. Baer, and C. Y. Ng ͑Wiley, New York,
1995͒, p. 195.
ϩ
1
spectra of the a
͚
Rydberg state of N2 leading to the
Љ
g
ϩ
X 2͚ ionic state of Nϩ2 , were studied both experimentally
g
18 C. A. de Lange, J. Chem. Soc., Faraday Trans. 94, 3409 ͑1998͒.
19 C. A. de Lange, Adv. Chem. Phys. 117, 1 ͑2001͒.
20 R. R. Lucchese, G. Raseev, and V. McKoy, Phys. Rev. A 25, 2572 ͑1982͒.
21 K. Wang and V. McKoy, J. Chem. Phys. 95, 4977 ͑1991͒.
22 S. T. Pratt, P. M. Dehmer, and J. L. Dehmer, J. Chem. Phys. 78, 4315
͑1983͒.
and theoretically. The agreement between measured and cal-
culated spectra is excellent.
ACKNOWLEDGMENTS
A.M.R. acknowledges the Holland Research School of
Molecular Chemistry for a Ph.D. fellowship. The authors
thank Ing. D. Bebelaar for technical support. Work at the
California Institute of Technology was supported by a grant
from the National Science Foundation ͑U.S.͒.
23 M. A. O’Halloran, S. T. Pratt, P. M. Dehmer, and J. L. Dehmer, J. Chem.
Phys. 87, 3288 ͑1987͒.
24 S. T. Pratt, E. F. McCormack, P. M. Dehmer, and J. L. Dehmer, J. Chem.
Phys. 92, 1831 ͑1990͒.
25 C. R. Scheper, W. J. Buma, C. A. de Lange, and W. J. van der Zande, J.
Chem. Phys. 109, 8319 ͑1998͒.
26 S. T. Pratt, P. M. Dehmer, and J. L. Dehmer, J. Chem. Phys. 92, 262
͑1990͒.
1 T. E. Graedel and P. J. Crutzen, Atmospheric Change: An Earth System
Perspective ͑Freeman, New York, 1992͒.
27 C. R. Scheper, M. F. Somers, and C. A. de Lange, J. Electron Spectrosc.
Relat. Phenom. 108, 123 ͑2000͒.
2 J. T. Houghton, L. G. M. Filho, D. J. Griggs, and K. Maskell, Stabilization
of Atmospheric Greenhouse Gases: Physical, Biological and Socio-
economic Implications, World Meteorological Organization, IPCC Tech-
nical Paper III ͑IPCC, Geneva, Switzerland, 1997͒.
28 E. de Beer, C. A. de Lange, J. A. Stephens, K. Wang, and V. McKoy, J.
Chem. Phys. 95, 714 ͑1991͒.
29 E. de Beer, M. Born, C. A. de Lange, and N. P. C. Westwood, Chem.
Phys. Lett. 186, 40 ͑1991͒.
3 K. F. Preston and R. F. Barr, J. Chem. Phys. 54, 3347 ͑1971͒.
4 D. G. Hopper, J. Chem. Phys. 80, 4290 ͑1984͒.
30 K. Wang, J. A. Stephens, V. McKoy, E. de Beer, C. A. de Lange, and N.
P. C. Westwood, J. Chem. Phys. 97, 211 ͑1992͒.
5 J. M. Teule, G. C. Groenenboom, D. W. Neyer, D. W. Chandler, and M.
H. M. Janssen, Chem. Phys. Lett. 320, 177 ͑2000͒.
31 J. B. Milan, W. J. Buma, C. A. de Lange, K. Wang, and V. McKoy, J.
Chem. Phys. 107, 2782 ͑1997͒.
6 J. W. Rabalais, J. M. McDonald, V. Scherr, and S. P. McGlynn, Chem.
Rev. 71, 73 ͑1971͒.
32 C. A. de Lange, in The Role of Rydberg States in Spectroscopy and Re-
activity, edited by C. Sandorfy ͑Kluwer Academic, Dordrecht, 1999͒, p.
457.
7 Th. F. Hanisco and A. C. Kummel, J. Phys. Chem. 97, 7242 ͑1993͒.
8 T. Suzuki, H. Katayanagi, Y. Mo, and K. Tonokura, Chem. Phys. Lett.
256, 90 ͑1996͒.
33 T. A. Miller, T. Suzuki, and E. Hirota, J. Chem. Phys. 80, 4671 ͑1984͒.
34 A. M. Rijs, E. H. G. Backus, C. A. de Lange, N. P. C. Westwood, and M.
H. M. Janssen, J. Electron Spectrosc. Relat. Phenom. 112, 151 ͑2000͒.
35 K. Wang and V. McKoy, Annu. Rev. Phys. Chem. 46, 275 ͑1995͒.
36 W. J. Hunt and W. A. Goddard, Chem. Phys. Lett. 3, 414 ͑1969͒.
37 T. H. Dunning, Jr., J. Chem. Phys. 53, 2823 ͑1970͒; 55, 3958 ͑1971͒.
38 G. Selwyn, J. Podolske, and H. S. Johnston, Geophys. Res. Lett. 4, 427
͑1977͒.
9 D. W. Neyer, A. J. R. Heck, and D. W. Chandler, J. Chem. Phys. 110,
3411 ͑1999͒.
10 D. W. Neyer, A. J. R. Heck, D. W. Chandler, J. M. Teule, and M. H. M.
Janssen, J. Phys. Chem. A 103, 10388 ͑1999͒.
11 J. M. Teuel, Ph.D. thesis, Vrije Universiteit, Amsterdam, 1997.
12 M. Ahmed, E. R. Wouters, D. S. Peterka, O. S. Vasyutinskii, and A. G.
Suits, Faraday Discuss. 113, 425 ͑1999͒.
13 K. R. Lykke and B. D. Kay, J. Chem. Phys. 95, 2252 ͑1991͒.
14 Th. F. Hanisco and A. C. Kummel, J. Phys. Chem. 95, 8565 ͑1991͒.
15 Th. F. Hanisco and A. C. Kummel, J. Phys. Chem. 96, 2982 ͑1992͒.
39 J. Xie and R. N. Zare, J. Chem. Phys. 93, 3033 ͑1990͒.
40 J. Bendtsen, J. Raman Spectrosc. 2, 133 ͑1974͒.
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:
131.187.254.4 On: Sat, 22 Nov 2014 17:10:11