The effect of parent zero-point motion on the ND 2 ( ? ) rotational state distribution in the 193.3 nm photolysis of ND 3

Article abstract of DOI:10.1016/S0009-2614(00)00605-9

The vibrational and rotational product-state distributions of ND₂(?²A₁) has been probed following the photodissociation of ND₃ at 193.3 nm by time-resolved Fourier Transform infrared emission spectroscopy. The dynamics of the bond cleavage are inferred from the product state distributions by comparison with an earlier study of the photodissociation of NH₃. The degree of excitation about the minor rotational b/c-axes of the product is attributed to the amount of zero-point energy of the parent molecule in the ν₄ H-N-H (D-N-D) scissors bending coordinate of the NH₃/ND₃(?) predissociative state. A bimodal ND₂(?²A₁) distribution is observed for rotation about the primary a-axis, analogous to the NH₂ fragment formed in the photodissociation of NH₃.

Full text of DOI:10.1016/S0009-2614(00)00605-9

7
July 2000
Chemical Physics Letters 324 Ž2000. 240–248
www.elsevier.nlrlocatercplett
ž ˜ /
The effect of parent zero-point motion on the ND A rotational
2
state distribution in the 193.3 nm photolysis of ND3
Jonathan P. Reid, Richard A. Loomis , Stephen R. Leone),2
1
JILA, National Institute of Standards and Technology and UniÕersity of Colorado, Department of Chemistry and Biochemistry,
UniÕersity of Colorado, Boulder, CO 80309-0440, USA
Received 31 March 2000; in final form 8 May 2000
Abstract
2
The vibrational and rotational product-state distributions of ND Ž A˜ A . has been probed following the photodissociation
2
1
of ND at 193.3 nm by time-resolved Fourier Transform infrared emission spectroscopy. The dynamics of the bond cleavage
3
are inferred from the product state distributions by comparison with an earlier study of the photodissociation of NH . The
3
degree of excitation about the minor rotational brc-axes of the product is attributed to the amount of zero-point energy of
the parent molecule in the n H–N–H ŽD–N–D. scissors bending coordinate of the NH rND Ž A˜ . predissociative state. A
4
3
3
2
bimodal ND Ž A˜ A . distribution is observed for rotation about the primary a-axis, analogous to the NH fragment formed
2
1
2
in the photodissociation of NH . q 2000 Elsevier Science B.V. All rights reserved.
3
1
. Introduction
spectroscopy w1x. By probing the emission from the
2
electronically excited NH Ž A˜ A . fragments, the
2
1
Probing the product state distribution following
partitioning of energy into the rotational and bending
vibrational levels of the product was determined. In
particular, the dynamics leading to rotation around
both the a-axis and brc-axes were explored in a
supersonic jet-cooled expansion of the parent species,
which minimizes the influence of parent rotational
excitation on the bond cleavage dynamics. This ex-
tended the results of an earlier study of dissociation
at room temperature, in which the degree of rota-
tional excitation about the primary rotational a-axis
was probed w2x.
the photofragmentation of a molecular species is a
powerful method for characterizing the mechanism
by which bond cleavage occurs. Recently, we exam-
ined the dissociation mechanism of ammonia, NH ,
3
1
XX
from its first excited electronic state Ž A˜ A . by
2
time-resolved Fourier transform infrared ŽFTIR.
)
Corresponding author. Fax: q1-303-492-5504; e-mail:
srl@jila.colorado.edu
1
National Research Council Research Associate, National In-
Other studies have probed in detail both experi-
mentally Žsee, e.g., Refs. w3–6x and references
therein. and theoretically w7–9x the energy partition-
stitute of Standards and Technology. Current address: Department
of Chemistry, Campus Box 1134, Washington University, One
Brookings Drive, Saint Louis, MO 63130-4899, USA.
Staff member, Quantum Physics Division, National Institute
of Standards and Technology.
2
ing in the ground electronic state photoproducts,
2
NH Ž X˜ B .. These yield complementary informa-
2
1
0
009-2614r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved.
PII: S0009-2614Ž00.00605-9

J.P. Reid et al.rChemical Physics Letters 324 (2000) 240–248
241
˜
w x
tion to the studies that probe the A state fragments
temperature of 120"15 K about the minor axes 1 .
This can be rationalized by considering the effect of
vibrational zero-point motion in the parent molecule
on the dissociation dynamics. Specifically, the rota-
tional excitation is attributed to a mapping of the
zero-point bending motion in the n4 H–N–H scis-
and provide insights into the competition between
adiabatic and non-adiabatic electronic state dynam-
ics. The electronically excited NH Ž A˜ A . parent
correlates with the ground state NH Ž X˜ . product
1
XX
3
2
2
species at planar geometries while the excited state
NH Ž A˜ . product correlates with the ground state
sors bending coordinate of the NH Ž A˜ . predissocia-
tive state onto the NH Ž A˜ .qH photofragments w7x.
2
3
1
NH Ž X˜ A .. At non-planar geometries, however, the
3
1
2
1
X
total symmetry of these states is the same Ž A. and
an avoided crossing of the states occurs. Thus, the
parent molecule can dissociate diabatically at planar
geometries to the ground state product, while at
non-planar geometries, adiabatic dissociation leads to
The motion in this bending coordinate allows the
trigonal symmetry of the parent molecule to be
broken and permits a torque to be exerted about the
brc rotational axes of the product as the dissociating
bond stretches. This hypothesis, which is based on
theoretical calculations w7x, is tested further here by
the excited electronic state NH Ž A˜ . fragment and a
non-adiabatic dissociation leads to a ground state
2
probing the rotational excitation of ND Ž A˜ . about the
minor axes when ND is photodissociated in a jet-
2
fragment NH Ž X˜ ..
2
3
In our earlier study on the photodissociation of
cooled expansion. By deuterating the parent
molecule, the zero-point energy is reduced and hence
NH , a bimodal rotational distribution was observed
3
2
about the a-axis of the NH Ž A˜ A . product w1x. We
the zero-point bending amplitude. The ND Ž A˜ .
2
1
2
suggested that the origin of the bimodal distribution
is competition between two distinct mechanisms of
dissociation that sample different regions of the po-
tential energy surface. In addition, we suggested that
the competition between tunneling and direct disso-
ciations may play a significant role in determining
the product energy partitioning. The Franck–Condon
ŽFC. region accessed in the excitation step with a
brc-axes rotational temperature is found to be 81"
11 K. The lower degree of rotational excitation for
ND Ž A˜ . compared to NH Ž A˜ . is believed to be a
2
2
result of the smaller amplitude of the zero-point
motion of ND Ž A˜ . compared to NH Ž A˜ ..
3
3
Although experiments have also been performed
to probe the ground electronic state fragments in the
photodissociation of ND , ND H and NH D w6,10–
3
2
2
pyramidal geometry is classically bound along the
dissociation coordinate and dissociation can only
occur by tunneling through the barrier to bond cleav-
age. In contrast, as the molecule evolves away from
bent pyramidal geometries, the dissociation barrier
diminishes and the molecule can dissociate directly
at a planar geometry. To further probe the role of
tunneling and direct dissociations, we present here a
12x, there have been no studies of deuterated ammo-
nia to determine the energy partitioning of the frag-
ments in the excited electronic state. We recently
undertook such a comprehensive study, and the fully
and partially deuterated parent molecules provide a
wealth of information on the photofragmentation dy-
namics of ammonia, especially when compared with
the dissociation of non-deuterated NH . In this paper
3
study of the photodissociation dynamics of ND . By
we focus on the results of the photodissociation of
3
deuterating the parent molecule we anticipate a
change in efficiency for any tunneling dissociation
ND at a wavelength of 193.3 nm.
3
that may occur. In probing the ND Ž A˜ . fragment, we
2
observe a qualitatively similar bimodal rotational
distribution, but with quantitative differences in the
relative populations of high and low rotational states.
Jet-cooling the parent molecules to ;10 K shows
that the rotational energy about the minor brc-axes
does not arise simply as a consequence of carry-over
of parent rotational excitation to products, but may
2
. Experimental
A detailed description of the experimental appara-
tus has been given in a previous publication and only
a brief summary will be presented here w1x. Studies
are performed both at room temperature, with a
continuous flow of ammonia, and jet-cooled, with a
piezoelectric pulsed valve. The triggering of the ArF
have a dynamic origin. For jet-cooled NH , the
3
NH Ž A˜ . product was found to have a rotational
2

2
42
J.P. Reid et al.rChemical Physics Letters 324 (2000) 240–248
excimer laser, which initiates the photodissociation,
and the opening of the pulsed valve for the jet-cooled
studies, are both synchronized to the position of the
moving mirror of a commercial continuous-scan
Fourier Transform infrared ŽFTIR. spectrometer. Fol-
improved conditions used in our study, we have
concluded that the rotational temperature of the par-
ent molecules is also -10 K. In the experiments
presented below, a concentration of 2% ND3 in
helium was used with a backing pressure of ;67
kPa.
lowing the laser pulse, emission from the electroni-
cally excited ND fragments is collected and imaged
from the gas flow into the FTIR spectrometer with a
In both studies, with a room-temperature continu-
ous reagent flow and with a jet-cooled expansion, no
time dependence is observed in the product rota-
tional level distribution from 20 ns out to 2 ms after
2
CaF lensrmirror telescope arrangement. The emis-
2
sion is modulated by the interferometer and recorded
by a detector. Boxcar integrators are used to record
several time windows of the fluorescence at each
interferogram position and the time history of the
entire interferogram is constructed by pulsing the
piezoelectric pulsed valve and the laser once for each
interferogram position. A total of 16 coadditions are
recorded for each spectrum. The time-dependent
modulated signal is stored on a computer following
analog-to-digital conversion.
the photolysis pulse. The ND Ž A˜ . excited state rovi-
2
brational populations obtained under these conditions
represent the nascent values prior to any collisional
relaxation.
3
. Spectroscopy and product state analysis
2
2
The ND Ž A˜ A
X˜ B . electronic emission is
Numerous experimental and theoretical investiga-
2
1
1
y1
˜
˜
observed over the spectral range 6000–15 000 cm
.
tions of the NH A
X emission spectrum have
2
To acquire a complete spectrum over this broad
frequency range, the experiment is repeated under
been published Žsee, e.g., Refs. w15–17x.. However,
˜
˜
the emission spectrum of ND A X is much less
2
identical experimental conditions with two detectors,
well characterized. Indeed, the relevant term energies
and transition probabilities are unknown for the tran-
sitions observed in this study that originate from the
electronically excited nascent product levels popu-
lated. By comparison with the spectroscopy of NH2 ,
the observed spectral lines are attributed to rotational
transitions within the vibrational bands originating
y1
an InSb detector, for the range 6000–12 000 cm
,
and a Si avalanche photodiode for the range 10 000–
y1
1
5 000 cm . A matched detector is used to monitor
the total unmodulated emission intensity, and the
modulated signal is then corrected for variations in
the laser power and gas flow using this second
normalization channel. A complete spectrum with a
X
0
0
0
0
0
from y s0, 2 , 2 , 2 , 2 and 2 , where the
2
0
1
2
3
4
y1
Nyquist frequency of 15 798 cm is acquired with a
superscript denotes the vibrational level of the ex-
cited electronic state and the subscript that of the
ground state. The rovibrational selection rules for all
of the electronic transitions used in this spectral
analysis follow those of a symmetric prolate top w18x;
y1
spectral resolution of 0.2 cm
.
Extensive experiments were performed earlier to
characterize the jet-cooled expansion for NH3 w1x.
Only the conclusions crucial to the study presented
here are mentioned. It is important to ensure that the
dilution of ammonia in a helium buffer gas is suffi-
cient to avoid clustering of the parent molecules. In
addition, the backing pressure of the pulsed valve
must be sufficiently low to avoid clustering. With an
orifice diameter of 750 mm, no evidence of cluster-
ing was observed with concentrations of (2% NH3
in He and with backing pressures of less than 100
kPa. Similar experiments with an expansion of NH3
in Ar and an orifice of 1 mm showed that the
rotational temperature of the jet-cooled parent
molecule was -10 K w13,14x. Therefore, with the
the change in the total rotational angular momentum,
D Ns"1, equals the change in rotational quanta
about the a-axis, i.e. D NsD K s"1. Rotational
a
angular momenta about the axes of rotation other
than the a-axis, referred to collectively as rotation
X
about the brc-axes, is given by those values of N
X
X
for which N )K . Therefore, using the notation for
a
absorption, the emission spectrum is dominated by
P
XX
XX
a
R
XX
vibrational bands with PŽN , K . and RŽN ,
XX
K . rotational lines. The ‘P’ and ‘R’ superscripts
a
refer to D K sy1 and D K sq1, respectively,
a
a
and the ‘P’ and ‘R’ branches refer to D Nsy1 and

J.P. Reid et al.rChemical Physics Letters 324 (2000) 240–248
243
D Nsq1. Each rotational transition is further split
by spin-rotation coupling into two components, F1
F1 and F2 F2.
London factors, AK Xa X, NXX , and theoretically predicted
rovibronic transition moments, R, according to w18x
I
A full spectral analysis of the vibrational bands
X
X
n
Ž
N ,Ka.
s
.
3
1
Ž .
0
0
0
0
0
0
2
2
, 2 , 2 , 2 and 2 will be given in a future
AKXX XX < R< n
1
2
3
4
, N
a
publication; only the information needed to elucidate
the dynamics occurring during the bond cleavage are
discussed here. The method of combination differ-
ences was used in order to ensure a correct assign-
ment, by comparison with the known ground elec-
tronic state term energies as determined by Mordaunt
et al. w6x Based on these ground state term energies,
The Honl–London factors, AK Xa X, NXX , appropriate for a
perpendicular transition are used w18x.
¨
XX
XX
a
XX
XX
a
Ž
N y1qK . Ž N qK
.
AKXX
XX
s
2
Ž .
, N
XX
XX
a
Ka
Ž
2 K q1
.
a
the excited state term energies are determined and
presented in Table 1. The assignment was facilitated
by comparing the product emission spectra from
both the room-temperature and jet-cooled dissocia-
tions. In the latter of these two spectra, the dramatic
cooling of the parent molecule greatly reduces the
congestion of the emission spectra by constraining
the parent rotation, and hence the product, to a
smaller subset of rotational states. An example of a
section of the emission spectra, both in a jet-cooled
experiment and in a room-temperature dissociation is
shown in Fig. 1.
The vibronic intensity factors, R, are not cur-
rently known for ND2 and so correction of the
populations for these factors cannot be made. By
analogy with the analysis performed on the fragment
NH2 from the photodissociation of NH , correction
3
by the vibronic intensity factors is not expected to
change the qualitative trends and conclusions of this
work, although the quantitative populations will be
susceptible to error. To illustrate the change that
inclusion of the vibronic intensity factors could make,
consider the most intense vibrational bands in the
0
1
0
2
NH emission spectrum, 2 and 2 , and the transi-
2
X
X
The relative populations, nŽN , K ., found within
tions from the lowest and highest rotational states in
both the P and R bands. The vibronic intensity
factors increase by a factor of ;2.3 for both the P
a
each ND Ž A˜ . rovibrational level can be in principle
2
P
XX
XX
a
determined from the intensities of the PŽN , K .
0
emission lines, I, transition frequencies, n, H o¨ nl–
and R branches of the 21 vibrational band and by
Table 1
X
X
a
X
2
Term energies for the states ND Ž A˜ , N sK , y s0. relative to the zero-point energy of the ground electronic state. The values quoted are
2
0
0
P
0
0
R
0
1
P
0
R
y1
taken from the 2
P bands, although values from the 2 R, 2
P and 2 R agree within 7 cm . The value for the ND Ž A˜ ,
1
2
X
X
X
2
0
P
1
N sK s0, y s0. level is taken from the 2 P band. These term energies are based on the experimental term energies estimated by
a
XX
XX
XX
X
X
Mordaunt et al. for ND Ž X˜ , N sK , y s0., which are quoted for completeness w6x. The N )K term energies are only quoted for those
2
a
a
X
X
a
states used in the analysis of the N )K populations
Ka
Term energyrcmy1
XX
XX
a
XX
X
X
a
X
2
ND Ž X˜ , N , K , y s0.
ND ŽN , K , y s0.
2
2
XX
XX
XX
XX
a
XX
XX
a
X
X
a
X
X
a
X
X
a
N sK
N sK q1
N sK q2
N sK
N sK q1
N sK q2
a
0
1
2
5
6
7
8
9
0
19
64
358
506
679
876
1096
11
40
96
421
581
765
973
1204
32
72
10478
10842
11195
12397
12835
13306
13804
14324
–
–
–
–
–
–
–
–
139
496
666
861
1080
1322
12896
13379
13884
–
12967
13458
–
–

2
44
J.P. Reid et al.rChemical Physics Letters 324 (2000) 240–248
transitions are observed with the resolution of the
spectrometer, as is seen in Fig. 1, and the average of
the F and F rotational level populations are re-
1
2
ported. Several vibrational bands originating from
the same excited n2 vibrational state are observed
X
X
and the average of the N sK populations is deter-
a
0
Õ
mined from all of the observed 2 XX bands. The total
X
X
a
X
2
integrated ND Ž A˜ , N sK , y s0. level population
2
is normalized to one.
X
X
a
To analyze those states with N )K , only the
K s6, 7, and 8, y s0 populations are used. Insuf-
ficient signal-to-noise precludes the use of transitions
X
a
X
2
X
a
originating from other K states for the determina-
tion of the energy partitioning about the minor rota-
tional axes. As in our previous publications, we
assume that the vibronic intensity factors are the
X
X
same for each transition in the series N 0K for
a
X
a
any particular K value. The populations of the
X
X
a
X
2
ND Ž A˜ , N 0K , y s0. levels are scaled so that
2
X
X
a
the populations in the N sK s6, 7, and 8 rotor
X
levels divided by the Ž2 N q1. state degeneracy
each equal one. In our previous examination of the
product state distribution following the dissociation
Fig. 1. A section of the photodissociation product emission spec-
2
trum of ND2 Ž A . from the photodissociation ND . Ža. Room-
1
3
temperature dissociation; Žb. jet-cooled. Note the significant sim-
plification of the spectra on jet-cooling the parent molecule.
0
2
;
1.8 for the 2 band on comparing the lowest
rotational state with the highest w1,19x. This leads to
transitions that are approximately a factor of 4 weaker
X
X
from the lowest N sK states relative to the high-
a
est. If a similar trend were to be observed in the
vibronic intensity factors for ND , the population in
2
X
X
a
the lowest N sK state would increase by a factor
of 4, relative to the highest N sK states, from that
X
X
a
calculated in our analysis.
X
X
X
Fig. 2. The relative ND2 Žy s0, N s K . population distribu-
2
a
The populations, uncorrected for the vibronic in-
tion obtained by the 193.3 nm photolysis of a room-temperature
X
X
tensity factors, of the product states ND Ž A˜ , N sK ,
2
a
effusive flow Žsquares. and jet-cooled expansion Žcircles. of ND3.
X
2
y s0. are shown in Fig. 2. A bimodal rotational
Representative error bars are given to show the upper-limit of the
X
X
distribution is observed with the population peaking
population of the N s Kas3 and 4 states that cannot be ob-
served within the signal-to-noise of this experiment. For these
levels, the larger error bars refer to the room-temperature study
and the smaller error bars to the jet-cooled, and they are posi-
tioned in energy to illustrate only the approximate term energies
of these states.
X
X
in N sK s0 and 7. Note that the intervening
a
X
X
transitions from N sK s3 and 4 are too weak to
a
be observed within the signal-to-noise of this experi-
ment. The F rF spin-rotation components of the
1
2

J.P. Reid et al.rChemical Physics Letters 324 (2000) 240–248
245
of NH , the energy partitioning about the minor axes
4. Discussion
3
was best described by a rotational temperature w1,2x.
A similar Boltzmann analysis has been performed
here to characterize the rotational energy about the
brc-axes and the results of this analysis are illus-
trated in Fig. 3. To calculate the relative rotational
energy about the brc-axes, the term energies mea-
sured by Mordaunt et al. w6x are used for the appro-
In comparison with the dissociation of NH , the
most striking difference in the photodissociation of
3
ND , as illustrated in Fig. 3, is the rotational temper-
3
atures about the minor rotational brc-axes of the
ND Ž A˜ . fragment in the room-temperature and jet-
2
cooled dissociations. The analysis shows that the
priate states in the ground electronic state. The rota-
tional temperature of the ND2 product about the
brc-axes is 265"15 K in the room temperature
rotational temperature of ND Ž A˜ ., when produced in
2
a room temperature flow of the parent species, is
similar to that of NH Ž A˜ . from the photodissociation
2
photodissociation of ND and 81"11 K in the
3
of NH3 Ž280"20 K., within their error bars w1x.
Since the parent molecule initially has a room tem-
perature rotational distribution, the similarity of these
product rotational temperatures suggests that the level
of excitation about these minor rotational axes is a
result of carry-over of the parent rotational excitation
into the product. However, when the parent molecule
is cooled to a temperature of -10 K, the product
ND2 or NH2 fragment has an elevated level of
rotational excitation that cannot simply be a result of
carry-over from the parent molecule, but must have
an origin in the dynamics occurring during the bond
cleavage. The jet-cooled dissociation of both NH3
jet-cooled studies.
and ND show this qualitatively similar trend in the
3
product energy partitioning.
Theoretical calculations by Dixon w7x have shown
that zero-point motion in the n4 H–N–H scissors
bending coordinate of the NH Ž A˜ . predissociative
3
state has an amplitude of "118. By considering the
mapping of a ground-state Gaussian wavefunction in
this coordinate onto the product states, Dixon pre-
XX
XX
XX
a
dicted the population in NH Ž X˜ , y s0, N , K .
2
XX
XX
states with N )K . For NH , we confirmed that
a
3
the rotational energy partitioning about the brc-axes
X
X
X
was indeed very similar for NH Ž A˜ , N )K , y s0.
2
a
2
to that predicted from the range of zero-point motion
geometries sampled during the dissociation w1x, as
calculated by Dixon for fragmentation to the ground
electronic state product w7x. A schematic, which illus-
trates this dissociation mechanism, is shown in Fig.
4. Dissociation from the equilibrium geometry in the
n4 H–N–H scissors bending coordinate of the parent
leads to a molecule that retains C2 Õ symmetry during
the elongation of the bond being cleaved. By con-
trast, zero-point motion in this vibrational mode
permits the molecule to evolve away from this equi-
X
2
X
X
a
Fig. 3. Boltzmann-type plots of ND2 Žy s0, N 0 K . popula-
tions Žfilled squares. following Ža. the room-temperature photodis-
sociation of ND , and Žb. the jet-cooled photodissociation. In both
cases, the best fit Boltzmann temperatures are illustrated by the
solid line. For comparison, the similar Boltzmann plots for the
photodissociation of NH3 are included Žopen circles. and the best
3
fit Boltzmann temperature is given by the dashed line.

2
46
J.P. Reid et al.rChemical Physics Letters 324 (2000) 240–248
Fig. 4. Schematic showing the origin of the rotational excitation about the minor brc-axes. A change in the equilibrium geometry that
corresponds to the approximate zero-point motion in the ND Ž A˜ . n coordinate breaks the symmetry of the parent molecule and would
3
4
classically generate a force about the ND Ž A˜ . brc-axes.
2
librium geometry, thus breaking the C2 Õ symmetry,
with the N–H bond elongating in a geometry for
shown in Fig. 2 are not corrected for the vibronic
intensity factors, R, of Eq. Ž2. for the ND rotational
2
which the H-atom moves out of the s symmetry
transitions. However, by analogy with similar transi-
Õ
plane. Simple classical arguments show that this
would result in a torque and excitation about the
minor-rotational axes.
tions in the spectroscopy of NH , the inclusion of
this factor will most likely enhance the populations
2
in the lowest K states relative to the populations in
a
With the qualitative explanation for the excitation
about the minor rotational axes being the same, we
now consider the quantitative values for the NH2
the high Ka states. Although the populations in the
lowest Ka levels are sufficient for transitions origi-
nating from these states to be observed, the bimodal-
ity of the rotational distribution precludes the obser-
and ND product brc-axes rotational temperatures.
2
The ND Ž A˜ . parent molecule has less zero-point
vation of the intermediate states with K s3 and 4
3
a
energy in the n4 D–N–D scissors bending mode
due to the very small H o¨ nl–London line strength
factors for these transitions. A further similarity be-
than the NH Ž A˜ . molecule and therefore less ampli-
3
tween the bimodal rotational distributions of NH Ž A˜ .
tude in the angular bending coordinate. Although the
extent of the energy difference cannot be determined
for the excited electronic states of the two molecules,
from the ground state harmonic frequencies, the
2
and ND Ž A˜ . is that, following jet-cooling of the
2
parent molecules, the breadth of the rotational distri-
bution narrows considerably. At room temperature a
y1
fraction of the total population is observed in ND Ž A˜ ,
zero-point levels differ in energy by 238 cm , with
2
y1
X
X
a
X
2
energies of 867 and 629 cm
for NH3 and ND3,
N sK s8 and 9, y s0. where both states lie
respectively w11x. Thus, there is a qualitative explana-
tion for the different brc-axes rotational tempera-
tures of the ND and NH . In the former case, the
above the energetically accessible limit for a cold
parent molecule. This fraction drops significantly
when the thermal rotational energy of the parent
molecule is reduced by jet-cooling.
2
2
parent ND molecule has less zero-point energy and
3
a smaller bending amplitude and, thus, less torque is
induced about the minor rotational axes as the N–D
bond cleaves. This is reflected in the lower brc-axes
In our previous paper on the photodissociation of
NH , we suggested that the bimodality of the rota-
3
tional distribution originates from the competition
between two distinct mechanisms for dissociation
that sample different geometries during the bond
cleavage w1x. Mordaunt et al. performed polarization
rotational energy of the ND Ž A˜ . product compared to
2
the NH Ž A˜ . fragment.
2
The qualitative trend in the rotational energy par-
titioning about the major a-axis of ND Ž A˜ . is similar
studies on the photodissociation dynamics of ND to
2
3
to that of NH Ž A˜ . from the photodissociation of
the ground electronic state ND fragment and showed
2
2
NH . The distributions are bimodal with peaks at
that planar dissociations led to low rotational excita-
tion and non-planar dissociations to high rotational
excitation w20x. We concluded in our earlier paper
3
both high Ka ŽK s5 for NH and K s7 for
a
2
a
ND ., near the energetic limit and at the lowest K .
2
a
It must be reiterated that the population distributions
that non-planar geometries accessed in the Franck–

J.P. Reid et al.rChemical Physics Letters 324 (2000) 240–248
247
X
2
Condon region at an excitation wavelength of 193.3
nm can produce an NH2 fragment with a high
degree of rotational excitation about the a-axis w1x.
intensities of transitions from both y s0 and 1. In
the current study of the photofragmentation of ND
only very weak signatures of vibrationally excited
ND were observed and these have an insufficient
signal-to-noise ratio for a spectroscopic analysis. The
relative ratio of the experimental integrated band
intensities over all rotational transitions was esti-
3
Dissociation from the Franck–Condon region re-
quires bond cleavage by tunneling through the bar-
rier to dissociation. In contrast, dissociation from
planar geometries will yield a product fragment with
very little a-axis rotation and can proceed by direct
dissociation over the barrier. Although a direct com-
parison of the ND and NH rotational distributions
2
X
mated to be 20:1 for the partitioning ND Ž A˜ , y s
2
2
X
.rND Ž A˜ , y s1.. Although the available energy
0
2
2
y1
w11x to the ND fragment is 895 cm less than that
2
2
2
is precluded in this study because the ND popula-
available to NH2 for photodissociation at 193.3 nm
2
tions remain uncorrected for the vibronic intensity
factor, the inclusion of these factors will likely lead
to an enhancement of the population in the low
rotational states relative to the high states. This
suggests that the ND2 fragments are born with a
lower degree of rotational excitation about the a-axis
of ND3 and NH , respectively, this apparent dra-
matic reduction in the degree of vibrational excita-
tion upon deuteration of the parent molecule cannot
3
be due to an energetic factor. The ND fragment has
2
y1
3325 cm available for partitioning into the internal
energy of the excited electronic state fragment. One
possible explanation might simply be that the inte-
grated band intensities cannot be compared in this
way, since the vibronic intensity factors have not
been accounted for. This seems an improbable expla-
nation, as this would require that the oscillator
strength for the entire set of rotational transitions
from the excited vibrational level must be dimin-
ished by close resonances to dark states in the ground
electronic state w1x. While such degradation of oscil-
X
X
than NH , with ND Ž A˜ , N sK s0, 1 and 2,
2
2
a
X
2
y s0. dominating in the ND dissociation, but with
3
X
X
a
X
2
NH Ž A˜ , N sK s4, 5 and 6, y s0. dominating
2
the NH dissociation. This could reflect the lower
3
tunneling efficiency of D-atoms when compared with
H-atoms during the bond cleavage, and less dissocia-
tion from the FC region at non-planar geometries
that leads to rotational excitation. In addition, the
lower zero-point energy of ND than NH in the
3
3
ground electronic state suggests that the excitation
step leads to an ND3 molecule that lies lower in
energy on the excited state potential energy surface
lator strength is observed for certain states in the
spectroscopy of NH Ž A˜ ., only a very small subset of
2
transitions is affected. Thus, the large preference for
X
2
X
2
a ND Ž A˜ , y s0. over a ND Ž A˜ , y s1. product
than NH . The barrier is, thus, larger to bond cleav-
3
2
2
age at the large out-of-plane bend angles sampled in
the FC region and the molecule is channeled more
effectively to planar geometries where the barrier
can be surmounted directly without tunneling in the
cleavage of the N–D bond. This reduces the extent
of rotational excitation about the a-axis in the frag-
ment. Previous studies have suggested a disparity in
dissociation rates for NH3 and ND3 and it was
concluded that this difference was due to the differ-
ing tunneling efficiencies of H and D atoms w3x.
fragment must have a dynamic origin.
Without theoretical calculations to test hypothe-
ses, only significant factors that could influence the
dynamics may be tentatively suggested and it is
hoped that these will stimulate further theoretical
work on these systems. The primary differences in
the photodissociation of NH and ND occur in the
3
3
excitation step. NH undergoes excitation to NH3
3
1
XX
2
Ž A˜ A , y s6. followed by rapid predissociation
2
that has been suggested to involve a Fermi-resonance
between the bending vibration and the symmetric
stretch w3x. The ND undergoes excitation to ND
The extent of vibrational energy partitioned into
X
bending excitation of the fragment, NH Ž A˜ , y s
2
2
3
3
X
2
1 XX
.rNH Ž A˜ , y s1., in the photodissociation of NH
was found to be 3:1 under the room-temperature
Ž A˜ A , y s7. followed by slower predissociation
0
2
3
2
2
w3x. We suggest that understanding the character of
the vibrational mode excited in the parent molecule
is essential before any substantive conclusions can
be drawn as to the reason for the differences in the
vibrational energy partitioning in the products from
conditions w1x. Since the vibronic intensity factors are
X
similar for transitions from both y s0 and 1, this
2
population ratio derives directly from the relative
ratio of the experimentally observed integrated band

2
48
J.P. Reid et al.rChemical Physics Letters 324 (2000) 240–248
w4x J. Biesner, L. Schnieder, J. Schmeer, G. Ahlers, X. Xie, K.H.
the photodissociation of NH and ND . In addition,
3
3
Welge, M.N.R. Ashfold, R.N. Dixon, J. Chem. Phys. 88
Ž1988. 3607.
the character of the vibrational mode may signifi-
cantly influence the regions of the potential energy
surface sampled during the dissociation and the pos-
sible competition between adiabatic and non-adia-
batic dynamics.
w5x J. Biesner, L. Schnieder, G. Ahlers, X. Xie, K.H. Welge,
M.N.R. Ashfold, R.N. Dixon, J. Chem. Phys. 91 Ž1989.
2901.
w6x D.H. Mordaunt, M.N.R. Ashfold, R.N. Dixon, J. Chem.
Phys. 104 Ž1996. 6460.
w7x R.N. Dixon, Mol. Phys. 88 Ž1996. 949.
w8x M.I. McCarthy, P. Rosmus, H.-J. Werner, P. Botschwina, V.
Vaida, J. Chem. Phys. 86 Ž1987. 6693.
Acknowledgements
w9x P. Rosmus, P. Botschwina, H.-J. Werner, V. Vaida, P.C.
Engelking, M.I. McCarthy, J. Chem. Phys. 86 Ž1987. 6677.
The authors would like to thank the Department
of Energy for support of this research and the Na-
tional Science Foundation for additional equipment.
R.A.L. gratefully acknowledges the support of the
National Research Council for a Postdoctoral Fel-
lowship with the National Institute of Standards and
Technology. The authors would also like to thank G.
Duxbury for invaluable discussions concerning the
electronic spectroscopy of NH and ND .
w10x
S.A. Henck, M.A. Mason, W.-B. Yan, K.K. Lehmann, S.L.
Coy, J. Chem. Phys. 102 Ž1995. 4783.
w11x D.H. Mordaunt, R.N. Dixon, M.N.R. Ashfold, J. Chem.
Phys. 104 Ž1996. 6472.
w12x A. Nakajima, K. Fuke, K. Tsukamoto, Y. Yoshida, K. Kaya,
J. Phys. Chem. 95 Ž1991. 571.
w13
x
T. Seelemann, P. Andresen, J. Schleipen, B. Beyer, J.J. Ter
Meulen, Chem. Phys. 126 Ž1988. 27.
w14x G. Ebel, R. Krohne, H. Meyer, U. Buck, R. Schinke, T.
Seelemann, P. Andresen, J. Schleipen, J.J. ter Meulen, G.H.F.
Diercksen, J. Chem. Phys. 93 Ž1990. 6419.
2
2
w15x S.C. Ross, F.W. Birss, M. Vervloet, D.A. Ramsay, J. Mol.
Spec. 129 Ž1988. 436.
w16x C. Jungen, K.-E.J. Hallin, A.J. Merer, Mol. Phys. 40 Ž1980.
References
2
5.
w17x G. Duxbury, R.N. Dixon, Mol. Phys. 43 Ž1981. 255.
w1x R.A. Loomis, J.P. Reid, S.R. Leone, J. Chem. Phys. 112
Ž2000. 658.
w
18x G. Herzberg, Molecular Spectra and Molecular Structure III.
Electronic Spectra and Electronic Structure of Polyatomic
Molecules, Van Nostrand-Reinhold, New York, 1966.
w2x E.L. Woodbridge, M.N.R. Ashfold, S.R. Leone, J. Chem.
Phys. 94 Ž1991. 4195.
w19x G. Duxbury, private communication.
w20x D.H. Mordaunt, M.N.R. Ashfold, R.N. Dixon, J. Chem.
Phys. 109 Ž1998. 7659.
w3x V. Vaida, M.I. McCarthy, P.C. Engelking, P. Rosmus, H.-J.
Werner, P. Botschwina, J. Chem. Phys. 86 Ž1987. 6669.