Photodissociation dynamics of A state ammonia molecules. II. the isotopic dependence for partially and fully deuterated isotopomers

Article abstract of DOI:10.1063/1.471368

The technique of H(D) Rydberg atom photofragment translational spectroscopy has been used to investigate the photodissociation dynamics of the mixed isotogomers NH₂D and NHD₂ following the excitation to the v′₂ = 0 and 1 levels of their lowest lying A ¹B₁ (C_2v) excited electronic states. Peaks in the resulting total kinetic energy release (TKER) spectra are assigned to levels of the NH₂, NHD, or ND₂ fragments with a wide range of quantum numbers K_a for rotation about their a inertial axes, and with N = K_a, N = K_a + 1, or N = K_a + 2 as appropriate. These data provide the first measurements of high rotational levels for the ground electronic state of the NHD radical. The least squares fitting of all these spectra, and those resulting from NH₃ and ND₃, to the best calculated NH₂, NHD, and/or ND₂ rotational term values provides accurate estimations of the respective N-H and N-D bond dissociation energies D₀⁰ across the whole series. These values are D₀⁰(H-NH₂)=37 115±20 cm^-1 (4.602±0.002 eV); D₀⁰(H-NHD)=37 240±50 cm^-1; D₀⁰(H-ND₂)=37 300±30 cm^-1; D₀⁰(D-NHD)=37 880±60 cm^-1; and D₀⁰(D-ND₂)=38 010±20 cm^-1. The differences between these values are fully consistent with differences in zero-point energies and lead to a mean value of D_e=40 510±25 cm^-1. Dissociation of NH₂D or NHD₂ through their (A-X) 20 bands to give an NHD product leads to TKER spectra with a much higher statistical character than those leading to an NH₂ or ND₂ product, and to those obtained following excitation through the 0₀⁰ bands. This is rationalized in a semiquantitative manner in terms of a varying contribution to the dissociation rate of the parent molecules from internal conversion (IC) to high levels of their respective ground states. Nuclear permutation symmetry appears to play an important role both for the IC rates and for the subsequent branching between product channels.

Full text of DOI:10.1063/1.471368

Photodissociation dynamics of à state ammonia molecules. II. The isotopic
dependence for partially and fully deuterated isotopomers
David H. Mordaunt, Richard N. Dixon, and Michael N. R. Ashfold
Citation: The Journal of Chemical Physics 104, 6472 (1996); doi: 10.1063/1.471368
View online: http://dx.doi.org/10.1063/1.471368
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/104/17?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
The correlated product state distribution of ketene photodissociation at 308 nm
J. Chem. Phys. 104, 7460 (1996); 10.1063/1.471459
Photodissociation dynamics of à state ammonia molecules. I. State dependent μv correlations in the NH2(ND2)
products
J. Chem. Phys. 104, 6460 (1996); 10.1063/1.471367
Stateselective photodissociation dynamics of NOCl: Scalar and vector properties
AIP Conf. Proc. 191, 593 (1989); 10.1063/1.38637
Linear and nonlinear resonance Raman studies of subpicosecond photodissociation dynamics
AIP Conf. Proc. 172, 709 (1988); 10.1063/1.37461
Anisotropies in laserinduced chemical processes: Ejection of OH by the HONO Ã state
AIP Conf. Proc. 146, 549 (1986); 10.1063/1.35937
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
Photodissociation dynamics of A state ammonia molecules. II.
The isotopic dependence for partially and fully deuterated isotopomers
David H. Mordaunt,a) Richard N. Dixon, and Michael N. R. Ashfold
School of Chemistry, University of Bristol, Bristol BS8 1TS, United Kingdom
͑Received 25 July 1995; accepted 12 January 1996͒
The technique of H͑D͒ Rydberg atom photofragment translational spectroscopy has been used to
investigate the photodissociation dynamics of the mixed isotopomers NH2D and NHD2 following
1
˜
the excitation to the Ј ϭ 0 and 1 levels of their lowest lying A B ͑C ͒ excited electronic states.
v2
1
2v
Peaks in the resulting total kinetic energy release ͑TKER͒ spectra are assigned to levels of the NH2,
NHD, or ND2 fragments with a wide range of quantum numbers Ka for rotation about their a inertial
axes, and with NϭKa , NϭKaϩ1, or NϭKaϩ2 as appropriate. These data provide the first
measurements of high rotational levels for the ground electronic state of the NHD radical. The least
squares fitting of all these spectra, and those resulting from NH3 and ND3, to the best calculated
NH2, NHD, and/or ND2 rotational term values provides accurate estimations of the respective N–H
and N–D bond dissociation energies D00 across the whole series. These values are
D00͑H–NH2͒ϭ37 115Ϯ20 cmϪ1 ͑4.602Ϯ0.002 eV͒; D00͑H–NHD͒ϭ37 240Ϯ50 cmϪ1
;
D00͑H–ND2͒ϭ37 300Ϯ30 cmϪ1; D00͑D–NHD͒ϭ37 880Ϯ60 cmϪ1; and D00͑D–ND2͒ϭ38 010Ϯ20
cmϪ1. The differences between these values are fully consistent with differences in zero-point
energies and lead to a mean value of Deϭ40 510Ϯ25 cmϪ1. Dissociation of NH2D or NHD2
1
˜ ˜
through their (AϪX) 20 bands to give an NHD product leads to TKER spectra with a much higher
statistical character than those leading to an NH2 or ND2 product, and to those obtained following
excitation through the 000 bands. This is rationalized in a semiquantitative manner in terms of a
varying contribution to the dissociation rate of the parent molecules from internal conversion ͑IC͒
to high levels of their respective ground states. Nuclear permutation symmetry appears to play an
important role both for the IC rates and for the subsequent branching between product channels.
© 1996 American Institute of Physics. ͓S0021-9606͑96͒01715-7͔
I. INTRODUCTION
͑Refs. 9–11͒, all give rise to structured regions of absorption
indicating longer-lived excited electronic states and illustrate
further dissociation mechanisms. In CH3 and in NH3 ͑both of
which have electronic origins near 216 nm͒, there is a barrier
on the first excited state surface between an inner well and
the dissociation channel, and predissociation from the lowest
excited quasibound levels proceeds by quantum tunneling
The photochemistry of polyatomic hydride molecules
has been shown to involve a wide variety of dissociation
mechanisms. H2O excited to its A B1 electronic state within
1
˜
a region of continuous absorption centered about 165 nm
2
dissociates directly to give HϩOH(X ⌸) ground state
products.1 The distribution of the available energy over the
rotational and ⌳-doublet states of OH is highly specific, and
is interpreted in terms of tight dynamical constraints during
the process of dissociation which takes place on a single
potential energy surface. In contrast, when CH4 is photo-
lyzed within its first absorption continuum at 121.6 nm the
HϩCH3 product state internal energy distribution appears to
be purely statistical.2 The lowest dissociation energy is to
˜
through this barrier. For higher vibrational levels of NH3 (A)
intramolecular vibrational redistribution ͑IVR͒ converts en-
ergy in a vibrational coordinate orthogonal to the dissocia-
tion path into motion along this path, so that quantum tun-
neling is no longer necessary. The consequence is a much
shortened excited state lifetime. Even so, for NH3 in its first
excited state, and for H2O molecules in their second excited
state, the dissociation dynamics are dominated by strong
forces acting in the exit channel, brought about by a conical
intersection of two potential energy surfaces.
2
˜
HϩCH3(X A1), although at higher internal energy of the
fragments there is competition between a number of open
dissociation channels. In this case the potential energy sur-
The accompanying paper12 ͑hereafter referred to as pa-
per I͒ summarizes current knowledge of the photochemistry
of ammonia, and describes improved experiments concern-
ing the dissociation dynamics of NH3 and ND3 excited to
1
˜
face for the CH4(A T2) excited state does not correlate di-
rectly with the HϩCH3 ground state products, and dissocia-
tion is thought to be initiated through internal conversion to
high vibrational levels of the CH4 ground state. Both of these
cases involve excitation through absorption continua.
The first excited states of NH3 ͑Refs. 3–6͒ and the me-
thyl radical ͑Refs. 7 and 8͒, and higher excited states of H2O
either the origin level, or the 21 vibrational level, of their
˜
A
1A2Љ excited states. These experiments reveal a dramatic
correlation between the fragment recoil velocity vectors and
the a-axis rotational angular momentum of the NH2͑ND2͒
fragment.
a͒
Present address: Department of Chemistry, University of California, Ber-
keley, CA 94720, USA.
The tunneling rate for breaking an N–D bond is an order
6472
J. Chem. Phys. 104 (17), 1 May 1996
0021-9606/96/104(17)/6472/10/$10.00
© 1996 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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
6473
of magnitude slower than that for an N–H bond. One might
therefore predict that in partially deuterated ammonias an
NH bond would break in preference to an ND bond, and that
the dynamical properties would be intermediate between
those of NH3 and of ND3.
This paper presents new high resolution experimental
measurements of the internal energy distribution of the frag-
ments of dissociation of NH2D and NHD2 following excita-
tion to excited levels with Ј ϭ 0 or 1. We find that in all four
v2
cases the branching ratio does indeed favor the breaking of
an NH bond, but also that this ratio varies widely across the
series. Comparison of these branching ratios and absorption
linewidth measurements suggests that, in addition to the dy-
namical dissociation controlled by quantum tunneling, there
is a competing statistical dissociation mechanism. Further-
more, the spectra of the total kinetic energy release ͑TKER͒
of the mixed isotopomers are not intermediate in character
between those of NH3 and of ND3. In the most extreme case,
that of ͓NHD2͑210͒→DϩNHD͔, the internal energy distribu-
tion has a dominant statistical character, whereas the com-
peting and more favored ͓NHD2͑210͒→HϩND2͔ channel has
a much higher dynamical character. These various observa-
tions are rationalised in terms of a varying contribution to the
overall dissociation rate from an initial internal conversion to
high levels of the parent molecular ground state. Nuclear
permutation symmetry appears to play a significant role in
controlling both this rate and the subsequent dissociation,
giving the highest statistical character for a mixed isoto-
pomer dissociating to give the unsymmetrical NHD product.
II. EXPERIMENT
FIG. 1. Total kinetic energy release ͑TKER͒ spectra, all plotted on a com-
mon energy scale, for the H/DϩNHxD2Ϫx fragments resulting from photoly-
sis of jet-cooled samples of the various isotopomers of ammonia excited
The photofragment spectrometer has been described in
detail in previous publications, and is summarized in paper
I.12 Briefly, a skimmed pulsed supersonic beam of NH3 and
ND3 ͑in equimolar proportions͒—typically seeded as a 1%
mixture in argon—is intercepted at right angles by three
pulsed laser beams to initiate photolysis and ‘‘tag’’ the na-
scent H or D atomic fragments. The photolysis photons are
generated by third harmonic generation from a tuneable dye
laser operating on the dye DCM. The electric vector of this
3␻ radiation could be rotated relative to the atomic flight
axis. The times-of-flight to a detector of these tagged ‘‘Ry-
dberg atoms’’ are recorded and transformed to a recoil ki-
netic energy distribution. As in paper I, the energy resolution
of these total kinetic energy release spectra was improved
over that of earlier studies of NH3 and ND3 photolysis4–6 by
the use of a longer flight path of 825.6 mm from the inter-
action region to a grounded mesh immediately before the
front face of the detector.
Both H and D atoms were probed, and for the mixed
isotopomers it was important to measure the branching ratios
between these two dissociation channels. The shift between
the excitation wavelengths for H and D atoms is sufficiently
small that there was no difficulty in switching between these.
Even so, in order to normalize the two spectra thus gener-
ated, it was not only necessary to hold conditions stable
across both experiments, but also to take into account the
within the 000 bands of their AϪX transitions, and with ⑀phot aligned first
˜ ˜
parallel, and then perpendicular, to the TOF axis.
differing sampling of the Doppler widths of the H/D atomic
transitions. This difference arises from the mass dependence
of the momentum balance for the kinetic energy release,
since the Doppler widths were greater than the laser band-
width.
Under conditions of jet cooling, the 000 and 201 bands of
˜ ˜
the AϪX electronic band systems of all of the
isotopomers—NH3, NH2D, NHD2, and ND3—are essentially
free from overlapping, as is discussed in connection with
band simulations described later. Spectra were recorded for
excitation at chosen frequencies within all eight of these
bands. Both H atoms and D atoms were sought whenever
expected, but the D atom flux from NH2D was sufficiently
weak that a tolerable signal-to-noise ratio could only be ob-
tained in perpendicular polarisation. Other than this, at least
four spectra were recorded for each target molecule; for two
polarizations and two frequencies, and an additional four for
a second dissociation channel where appropriate.
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
6474
III. BOND DISSOCIATION ENERGIES
It is clear from Fig. 1 that for excitation of NHD2 via its
000 band the maximum total kinetic energy release for
NHD2→DϩNHD is lower than that for NHD2→HϩND2.
Thus the dissociation energy for the ND bond is higher than
that for the NH bond. Similar comparisons can be made
across the whole range of isotopomers after taking into ac-
count the differences in excitation energy. The potential en-
ergy surfaces for all these processes must be the same at the
level of the Born–Oppenheimer approximation. These differ-
ences should therefore reflect the differences in zero-point
energies between the parents and between the fragments.
It is necessary to know the ground state rotational ener-
gies for ND2 and NHD before we can establish accurate
values for all these dissociation energies. Biesner et al.5 re-
ported experimentally fitted eigenvalues for the important
FIG. 2. Total kinetic energy release ͑TKER͒ spectrum for the DϩNHD
fragments resulting from the photolysis of NHD2 at 46 529 cmϪ1 with ⑀phot
aligned perpendicular to the TOF axis ͑see also Fig. 1͒. The combs above
the spectrum indicate the assignments of peaks to levels of NHD with
2ϭ0, and with NϭKa , NϭKaϩ1, and NϭKaϩ2. The low Ka head for
v
v
2ϭ1 underlies these series near 7300 cmϪ1
.
˜
˜
range of levels of NH2(X) and ND2(X) from the TKER
spectra from photolysis of NH3 and ND3, respectively. In
striving for higher accuracy we use the term values for
basis it has proved possible to make unambiguous assign-
ments of the TKER spectra for an NHD partner ͑Fig. 2͒ and,
hence, to make a refined fit to the observations. An important
feature of these assignments is the detection for NHD of far
stronger peaks with NϾKa than are observed for NH2 or
ND2—a point to which we return later. The final calculated
rotational term values for the ground vibrational level of
NHD are given in the Appendix.
NH ͑ 2 , NϭKa͒ which have been derived by optical
v
2
spectroscopy.13 Those experimental measurements had been
analyzed by fitting them to the eigenvalues of a ‘‘semirigid
bender Hamiltonian,’’14 which includes Renner–Teller cou-
2
2
˜
˜
pling between the A A1 and X B1 electronic states which
correlate with a single 2⌸u state of linear NH2. The potential
energy curves thus derived for the two component states in-
corporate the zero-point energies of the two stretching vibra-
tions and compensate to some extent for the effects of
stretch-bend interaction while the molecule rotates. This
model calculation provides a good representation of the
Given these accurate values for the term values for NH2,
NHD, and ND , least squares fits to the 2ϭ0 and low Ka
v
2
portions of the various TKER spectra yield the values for the
TKER appropriate to the NϭKaϭ0 levels of each ground
state product. At the time that these experiments were carried
out there were no detailed experimental assignments for the
bands of the mixed NHxD3Ϫx parent isotopomers, and the
TKER spectra were obtained by choosing excitation wave
numbers within the 000 and 210 bands that maximized the H/D
atom fluxes from our cold molecular beams. Very recently,
Henck et al. have used microwave detected, microwave-
ground state levels of 2ϭ0 and 1 for all Ka and NϷKa .
v
Substitution of the mass of D for that of H in these calcula-
tions then gave good initial estimates for the eigenvalues for
2
˜
the X B1 state of ND2. From this starting point it proved
possible to assign the individual peaks in the TKER spectra
for H or D atoms with an ND2 partner. The resulting term
values for ND2 are reported in paper I.12
The use of this model is less straightforward for NHD
because of the loss of C2v symmetry. This causes a rotation
of the principal inertial axes to allow for the imbalance of H
and D. A simplification was therefore introduced which is
reasonably valid for the rotational levels for 2ϭ0. The
v
bending zero-point energy and other higher order terms were
also incorporated into a suitably modified ground state bend-
ing potential energy curve. For the high rotational levels of
interest Ka is a good quantum number. Consequently, for
each required level this curve was augmented at every angu-
lar coordinate for the N- and Ka-dependent centrifugal en-
ergy in a symmetric-top principal axis coordinate system,
and the eigenvalue estimated as the minimum of this effec-
tive potential. In this manner the experimental term values
for NH ( 2ϭ0) could be reproduced to within Ϯ1 cmϪ1. A
good fit to the less precise data for ND2 was obtained with
slightly modified potential parameters. A starting point for
the NHD calculations was then made by interpolating be-
tween these two sets of potential parameters and making the
appropriate changes of mass and rotation of axes. On this
v
2
0
1
˜ ˜
FIG. 3. Calculated absorption profiles for the 00 and 20 bands of the AϪX
transitions of NH2D and NHD2 using a rotational temperature of 15 K, but
a 300 K nuclear spin temperature. The dominant rovibronic transitions are
indicated in each case. For consistency these are all given in an oblate top
nomenclature ⌬JKc(J), with ⌬Kcϭ0 for all transitions. Spectroscopic con-
stants and linewidth parameters are from MODR experiments ͑Ref. 15͒.
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
6475
TABLE I. Excitation transitions and bond dissociation energies where measured.
Principal
Ϫ1
Isotopomer
␯
͒
Vibronic band
transition͑s͒
D00 ͑N–H͒/͑cmϪ1
͒
D00 ͑N–D͒/͑cmϪ1
͒
¯
͑cm
NH3
46 200
47 069
46 359
47 155
46 529
47 264
46 708
46 724
47 362
000
201
000
201
000
201
000
000
201
R0(1) and R1(1)
R0(0)
37 115Ϯ20
NH2D
NHD2
ND3
R1(1)
R0(0)
R0(0)
R0(0)
Q1(1)
R0(0)
Q1(1)
37 240Ϯ50
37 300Ϯ30
37 880Ϯ60
38 010Ϯ20
D H–NH ͒ϭD0 H–NH ͒ϩE ͑NH3)ϪEZP͑NH2).
optical double resonance ͑MODR͒ to analyze excitation
spectra for all the isotopomers.15 Figure 3 presents absorp-
tion spectra for the 000 and 210 bands of the mixed isotopomers
calculated using their molecular parameters for our beam
temperature of 15 K. Figure 3, together with Fig. 3 of Ref.
12 which presents equivalent calculated spectra for NH3 and
ND3, demonstrates that at this beam temperature there is no
significant overlap between the bands of the various isoto-
pomers present in our mixed sample. We are thus able to
deduce the precise parent molecule and excitation transitions
for each of our TKER spectra, although lifetime broadening
denies the possibility of excitation via a single rovibronic
transition in most cases. The bond dissociation energies are
then derived using
͑
͑
e
2
2
ZP
0
͑2͒
The vibrations of NH3, NH2, and their isotopomers are no-
ticeably anharmonic, and the zero-point energies are most
accurately represented by including quadratic terms in the
vibrational quantum numbers. For NH3 and ND3,
EZP͑NX )ϭ1/2
␻ d ϩ1/4
x d d ,
rs r s
͑3͒
͚
͚͚
3
r
r
r
rуs
in which the degeneracy factors dr are equal to 1 for rϭ1
and 2, and 2 for rϭ3 and 4. For the less symmetric mol-
ecules,
EZPϭ1/2
␻ ϩ1/4
r
x
,
rs
͑4͒
͚
͚͚
D00 H–NH ͒ϭE ͑NH )ϩh␯
r
rуs
͑
2
int
3
where rϭ1–3 for NH2, NHD, and ND2; and rϭ1–6 for
NH2D and NHD2. Unfortunately, NH3 is the only one of
these molecules for which all the necessary parameters have
been derived from the analysis of spectroscopic data.17 How-
ever, high quality ab initio potential energy surfaces for
ammonia18 and the amino radical19 have been used to calcu-
late spectroscopic parameters for all their isotopomers. Table
ϪTKER͑HϩNH2)–Eint͑NH2),
͑1͒
and its corresponding isotopic variants for the other isoto-
pomers. Fragment levels with 2ϭ0 and low values of N
v
and Ka are substantially populated for excitation within the
000 bands, but much less so for the higher vibronic bands. The
extrapolation from the observed peaks to zero internal en-
ergy is therefore minimized for the 000 bands, and these have
been preferred for the measurement of the bond dissociation
energies. Table I presents the dissociation energies thus de-
rived for all the processes studied. Note that these span a
II lists all the harmonic ␻ values from these calculations,
r
and the empirical values for NH3. The empirical zero-point
energy for NH3 calculated from Eq. ͑3͒ with the complete set
of anharmonic parameters is 2.8% lower than that from the
range of almost 1000 cmϪ1
.
purely theoretical parameters. Comparisons of other ␻ val-
r
The photolysis laser wave numbers were calibrated to a
precision of Ϯ0.5 cmϪ1 in these experiments and in each
case the mean internal energy of the parent molecule is less
than 20 cmϪ1. Some imprecision stems from the laboratory
to center-of-mass energy Jacobian to allow for the parent
beam velocity, estimated as 450Ϯ75 m sϪ1. This shifts the
observed TKER spectrum by about Ϫ9 cmϪ1 in the case of
NH3. The final best estimate of D00͑H–NH2͒ is 37 115Ϯ20
cmϪ1 ͑4.602Ϯ0.002 eV͒, as reported in paper I. This value
improves on the previous determination of 4.64͑5͒ eV from
the Hϩ ion time-of-flight ͑TOF͒ technique,4 and the currently
recommended literature value of 4.635Ϯ0.003 eV.16
ues where available indicate similar errors in the ab initio
values. Zero-point energies for the whole set of isotopomers
have therefore been derived from the ab initio spectroscopic
parameters by applying a consistent reduction by 2.8%. The
values of De thus obtained from the five experimental values
of D00 are in excellent agreement with one another ͑Table III͒
and lead to a mean value of De͑H–NH2͒ϭ40 510Ϯ25 cmϪ1
.
IV. PRODUCT STATE POPULATION DISTRIBUTIONS
TKER spectra have been recorded for dissociation of all
the NHxD3Ϫx isotopomers through selected excitations
within their 000 and 201 bands ͑Table I͒. In each case measure-
ments have been made with the time-of-flight path both par-
allel and perpendicular to the electric vector of the dissocia-
tion laser and, additionally, at intermediate angles in some
These various D00 dissociation energies are related to the
equilibrium dissociation energies De , which should be es-
sentially isotopomer independent, through the differences in
zero-point energy ͑EZP͒. Thus, for NH3,
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
6476
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
˜
TABLE II. Harmonic vibration frequencies for the X states of the H/D isotopomers of ammonia and the amino
radical.
Mode
Molecule
␻ ͑cmϪ1
͒
␻ ͑cmϪ1
͒
␻ ͑cmϪ1
͒
␻ ͑cmϪ1
͒
␻ ͑cmϪ1
͒
␻ ͑cmϪ1
͒
1
2
3
4
5
6
a
NH3
NH3
3503
3546
3598
3644
2534
3374
2498
2440
1030
1141
2649
2590
867
1524
1337
1119
3592
3687
1699
1318
2712
3481
3430
2561
1690
1733
1052
952
b
NH2Db
3687
2712
1478
1550
b
NHD2
b
ND3
NH2
1258
c
NHDc
c
ND2
aEmpirical fit to spectroscopic data ͑Ref. 17͒.
bFrom ab initio calculations ͑Ref. 18͒.
cFrom ab initio calculations ͑Ref. 19͒.
cases. Many of these spectra exhibit a significant anisotropy:
this is discussed in detail in paper I. We merely note here
that the integral product state distributions can be obtained as
a weighted mean of the measurements at 0° and 90°.
different from the equilibrium value for the ground state of
NH2 ͑105°͒; hence the predominance of levels with
͑NH2͒ϭ0.
v2
The dissociation of ND3 through its 000 band has previ-
ously been studied using Dϩ ion TOF spectroscopy, but the
resolution was inadequate for all but the coarsest rotational
structure.5 Here we report much higher resolution spectra
͑paper I, and also Fig. 1͒. The energy disposal is very similar
to that for NH3. The predominant peaks are again those for
A. Dissociation through the 000 bands
The dissociation dynamics of NH3 excited via its 000 band
has been amply discussed in previous publications.4–6,20
There is an almost uniform population of levels of NH2 with
2ϭ0 and NϭKa , with N ranging from 0 to 20. A weaker
v
population of the 2ϭ0, NϭKa levels of ND2, with weaker
v
series for 2ϭ1, NϭKa , and Nϭ0–19 is also discernible,
v
series for 2ϭ0, NϭKaϩ1 and 2ϭ1, NϭKa . This has
v
v
the low N head of which is responsible for an underlying
broad peak in each spectrum of Fig. 1, as are two short series
allowed us to derive term values ͑and confirm our calculated
values͒ for these previously unobserved levels of
for 2ϭ0 and 1, with NϭKaϩ1. The source of this rota-
v
2
˜
ND2(X B1) ͑see paper I͒.
tional excitation derives from the zero-point energies of the
We now turn our attention to the mixed isotopic parent
molecules NH2D and NHD2. For NH2D→DϩNH2, and for
NHD2→HϩND2, the fragmentation patterns are very similar
to those for NH3 and ND3, respectively, with the same domi-
nant series of peaks. These are the processes for which the
breaking bond is parallel to the C2 axis of the molecular
fragment. In contrast, although the peaks in the spectra for
NH2D→HϩNHD and NHD2→DϩNHD may be assigned in
two disappearing bending vibrations of NH3: ␯Ј correlates
with rotation of NH2 about its a rotational axis; whereas the
2
b2(C2v) component of ␯Ј correlates with rotation about its c
4
˜
axis. A conical intersection in the exit channel between the A
and X potential energy surfaces of NH3 has a dramatic influ-
˜
ence on the first of these, providing strong acceleration of the
out-of-plane bending motion which results in excitation of a
wide range of Ka states. In contrast, there are no strong
forces to cause such an acceleration for the c axis rotation.
The observation that levels with NϾKa are weak and mainly
at low Ka can thus be understood by noting that for levels
with NϭKaϩ1 the c-axis component of the rotational en-
ergy increases proportionally to (Kaϩ1). Furthermore, the
ЄHNH angle at the transition state between the quasibound
inner well of NH3 and the exit channel ͑ϳ110°͒ is not greatly
the same way, principally to levels with 2ϭ0, their relative
v
intensities differ substantially from those described above
but are similar to each other ͑Figs. 1 and 2͒. For low values
of Ka the strongest peaks are for NϭKa , but above Kaϭ6
the strongest peaks become those with NϭKaϩ1, and those
with NϭKaϩ2 are almost as strong. The interpretation of
this observation lies in the lack of C2 symmetry during the
TABLE III. Zero-point energies and equilibrium bond-dissociation energies De for the isotopomers of ammo-
nia.
Parent
Molecule
Daughter
product
EZP ͑parent͒
EZP ͑daughter͒
D00 ͑cmϪ1
͒
͑cmϪ1
͒
͑cmϪ1
͒
De ͑cmϪ1
͒
NH3
ND3
NHD2
NHD2
NH2D
NH2
ND2
ND2
NHD
NHD
37 115
38 010
37 300
37 880
37 240
7415
5467
6117
6117
6785
4005
2938
2938
3506
3506
40 525
40 539
40 479
40 491
40 519
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
6477
Nevertheless, although there is little population of the low
rotational levels for excitation through the 210 bands ͑Fig. 4͒,
the important series of levels are again those with NϷKa .4,5
This was rationalized by considering the form of the respec-
tive out-of-plane vibrational wave functions of the parent
molecule: that for Ј ϭ 0 favors dissociation from near-
v2
planar geometries, whereas that for Ј ϭ 1 has a node which
v2
discriminates against near-planar dissociation and low result-
ant angular momentum.
The TKER spectrum for NHD2→HϩND2 has a very
similar intensity distribution to that for ND3→DϩND2, with
the same prominent rotational series ͑Fig. 4͒: there is an
equivalent
comparison
for
NH2D→DϩNH2
and
NH3→HϩNH2, although all these spectra show a more
prominent background continuum than for dissociation via
the 000 bands. These processes all preserve C2v symmetry.
The situation is strikingly different where NHD is the mo-
lecular fragment. It is still possible to assign peaks to the
series of levels with NϷKa for NH2D→HϩNHD, although
there are many other unassigned peaks and a substantial con-
tinuous background. However, in the case of
NHD2→DϩNHD the continuum is dominant and few peaks
can be assigned with any certainty. This continuum is stron-
gest at low kinetic energy release, corresponding to high
NHD internal energy. This is qualitatively what would be
expected for a statistical distribution over the vibrational and
rotational degrees of freedom of NHD, rather than the dy-
namical population of a specific sub set of levels as in the
more symmetrical molecules. A clue to this variable behav-
ior is to be found in the pattern of H/D branching ratios for
the mixed isotopomers.
FIG. 4. Total kinetic energy release ͑TKER͒ spectra for the H/DϩNHxD2Ϫx
fragments resulting from photolysis of jet-cooled samples of the various
bands of their ˜AϪ˜X transi-
isotopomers of ammonia excited within the 201
V. ISOTOPIC BRANCHING RATIOS
tions, and with ⑀phot aligned first parallel, and then perpendicular, to the TOF
axis. The sharp spikes for all these spectra are genuine and repeatable peaks,
not noise.
The branching ratios for the production of H and of D
atoms from the mixed isotopomers should provide further
detailed evidence concerning the dissociation mechanism.
Several experiments were therefore carried out under care-
fully controlled conditions to establish absolute branching
ratios.
bond breaking process. The forces acting at the conical in-
tersection remain in the plane bisecting the ЄHND angle,
but the a-inertial axis of the NHD fragment now makes an
angle of about 21° to the perpendicular to this plane. Conse-
quently, the angular momentum generated by this force not
only has a major component along a axis of NHD, but also
has a significant component along the b axis.
The aforementioned discussion is concerned with the re-
solved rotational structure in each of the TKER spectra. It
may be noted ͑Fig. 1͒ that many of these spectra also exhibit
a weak underlying continuum, particularly for low kinetic
energy release. We shall see that this is a more prominent
feature for dissociation via the 210 bands.
For each TOF spectrum the total atom flux was obtained
by summation over the complete spectrum. The flux is, in
general, anisotropic as is evident from Figs. 1 and 4. This
anisotropy, and its variation with the excitation transition and
the associated product states, is discussed in detail in paper I.
For the purposes of measuring branching ratios the conse-
quences of this anisotropy were eliminated by taking the
weighted mean intensity as (Iʈϩ2IЌ)/3. Furthermore, since
it had not proved feasible to measure the ratio of Iʈ to IЌ in
all cases, the reasonable assumption was made where neces-
sary that this polarization ratio was the same for the corre-
sponding H and D channels. The measurements are summa-
rized in Table IV which also includes the homogeneous
linewidths ͓full width at half maximum ͑FWHM͔͒ recently
measured by MODR.15
B. Dissociation through the 210 bands
We have previously observed that when NH3 or ND3 is
˜
dissociated through excitation to higher members of the A–
X
n
˜
20 progression the rotational population distribution of
The dynamics of dissociation through excitation within
the 000 and 201 bands of NH3 and ND3 has been attributed to
quantum tunneling through any one of three equivalent exit-
the fragment becomes strongly inverted, and there is an in-
creasing excitation of the bending vibration of the fragment.
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
6478
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
TABLE IV. Total H and D atom flux ratios from photolysis of isotopomers of ammonia. nm represents not
measured.
H channel
0°:90°
D channel
0°:90°
H:D ratio
90°
⌫ ͑cmϪ1
FWHM
Ref. 15
͒
Weighted
mean
Isotopomer
0°
000 band
NH3
NH2D
NHD2
ND3
0.62Ϯ0.03
0.65Ϯ0.05
0.62Ϯ0.06
ϱ
nm
10Ϯ2
0
ϱ
ϱ
34.0Ϯ0.8
34.0Ϯ1.8
20 Ϯ2
nm
0.70Ϯ0.09
0.54Ϯ0.04
33Ϯ9
12Ϯ3
0
32Ϯ9
11Ϯ2
0
4.2Ϯ0.2
210 band
NH3
NH2D
NHD2
ND3
1.16Ϯ0.07
1.18Ϯ0.07
1.08Ϯ0.09
ϱ
nm
3.0 Ϯ 0.6
0
ϱ
21 Ϯ 4
3.6 Ϯ 0.7
0
ϱ
20 Ϯ 4
3.3 Ϯ 0.7
0
29.2Ϯ0.6
23.0Ϯ0.6
12.8Ϯ0.6
1.5Ϯ0.1
nm
1.19 Ϯ 0.10
0.99 Ϯ 0.02
channel barriers. The various linewidths for these excited
states have also been shown to be in reasonable quantitative
accord with this assumption, albeit with a barrier height
rather lower than the currently available ab initio value.21,22
To a first approximation it might be expected that there
would be a characteristic tunneling rate for an N–H bond,
and a lower value for an N–D bond. From the data of Table
for some of the product states. This definitely suggests the
involvement of a large number of degrees of freedom. It is
also instructive to note that for NHD2 the continuum is much
more in evidence for the channel to ͓DϩNHD͔ in direct
competition with that to ͓HϩND2͔, particularly for dissocia-
tion through the 210 band ͑Fig. 4͒. Furthermore, the presence
of assignable structure on top of a smooth continuum also
suggests that there are contributions from two independent
IV these elementary widths per bond for Ј ϭ 0 are 11.3
v2
cmϪ1 for N–H and 1.4 cmϪ1 for N–D, ͑a ratio of 8.1:1͒; and
0
˜
mechanisms, even for the 00 excitation. The levels of the A
for Ј ϭ 1 are 9.7 cmϪ1 for N–H, and 0.5 cmϪ1 for N–D ͑a
˜
v2
state are embedded in the continua of both the X state and
3
˜
ratio of 19.5:1͒. The lower widths and higher ratio for
Ј
v2
the a AЉ state, coupling to either of which could provide
ϭ 1 have been attributed to an increase in the barrier height
additional routes to dissociation.
with departure from planarity.23
A. Dissociation initiated through internal conversion
Two conclusions would follow from this assumption of
independent bond dynamics. Firstly, for the mixed isoto-
pomers the linewidths should show a linear dependence on
the number of D atoms. However, the linewidths for NH2D
and NHD2 are substantially greater than predicted by inter-
˜
The a-state surface has been calculated to lie less than
˜
0.5 eV below that for the A state, with a very similar poten-
tial function apart from a much lower barrier in the exit
˜
channel, and no strong coupling where it crosses the X-state
polation between NH and ND for both Ј ϭ 0 and 1. Sec-
surface.24 It is therefore doubtful that it would have a suffi-
cient density of states to permit efficient intersystem crossing
͑ISC͒. Furthermore, any such ISC would be unlikely to lead
to a fully statistical product state energy disposal. On the
v2
3
3
ondly their H/D branching ratios should be related to the
ratios of these line-widths. On this basis, and given the rela-
tive numbers of H and D atoms, the following values are
predicted for the H/D ratios. Tunneling from 000 band excita-
tion:
˜
other hand, the densities of vibrational levels of the X states
˜
at the energy of the A-state origins, calculated in the har-
monic approximation ͑which surely gives an underestimate͒,
are about 40/cmϪ1 for NH3 and 200/cmϪ1 for ND3 which are
high enough to facilitate internal conversion ͑IC͒.
NH2D H/Dϭ16.2; experimentϭ 32Ϯ9͒,
͑5͒
͑6͒
͑
NHD2 H/Dϭ4.0; experimentϭ 11Ϯ2͒.
͑
The most probable additional mechanism is therefore in-
ternal conversion to high-lying levels of the molecular
ground states, but with a rate relative to tunnelling which is
quite state specific. The IC rates in each molecule should
increase with increasing Ј since this mode transforms as the
v2
direct product of ˜Aϫ˜X, but the variation across the isoto-
Tunneling from 210 band excitation:
NH2D H/Dϭ39.0; experimentϭ 20Ϯ4͒,
͑7͒
͑8͒
͑
NHD2 H/Dϭ9.7; experimentϭ 3.3Ϯ0.7͒.
͑
Even though these predictions rest heavily on an assumption
of pure tunneling for NH3 and ND3, this simplistic model is
clearly not adequate. Either the remaining vibrational coor-
dinates make a cooperative contribution to the effective po-
tential for tunneling, or there is a further contribution to the
dissociation mechanism.
pomer series will depend on the interplay between an in-
creasing density of states and a decreasing coupling matrix
element. There is, however, an additional factor that must be
taken into account; the subdivision of the densities of states
to allow for nuclear permutation symmetry, which is prob-
˜
We now return to the observation of an apparent statis-
tical contribution to the rovibrational population distribution
ably the only relevant symmetry for very high levels of the X
states. For NH3 and ND3, the permutation symmetry group
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
6479
␲ is isomorphous with D3 , and the density can be subdi-
3
vided into [ 16ϫA1ϩ 61ϫA2ϩ 32ϫE]. For NH2D and NHD2,
the group is ␲2, isomorphous with C2 , and the division is
[ 12ϫAϩ 12ϫB]. Since, in all cases, the parent rotational an-
gular momentum is very low ͑JЈϭ1 or 2, Table I͒, and the
coupling is almost certainly vibronic, we consider only the
vibronic symmetry of the initially excited levels. These are
either A ( Ј ϭ 0) or A ( Ј ϭ 1) for NH and for ND , and A
v2 v2
2
1
3
3
for both vibrational states of NH2D and NHD2. Since per-
mutation symmetry must be conserved in IC, the accessible
fraction of the density of states is only one-sixth of the total
for NH3 and ND3 and one-half for NH2D and NHD2. This
provides a plausible explanation for the increased back-
ground continuum for the mixed isotopomers, particularly
for Ј ϭ 1.
v2
We now consider the branching between product chan-
nels from a highly excited ground state level for the mixed
isotopomers. There is no barrier to hinder such a process, so
the dissociation will take place through a loose transition
state. The rates can therefore be calculated by unimolecular
decay theory and will depend on the total densities of states
for the highly rotating and vibrating bent triatomic molecular
product, together with the three-dimensional partition func-
tion for the translational recoil of the H/D atom. For a given
parent molecule and energy the ratio of these branching rates
can be calculated from purely geometrical and mass factors.
The vibrational density is proportional to ͓det G͔Ϫ1/2, the ro-
tational density to ͓det I͔1/2, and the translational density to
FIG. 5. The total kinetic energy release ͑TKER͒ spectrum of the DϩNHD
fragments resulting from photolysis of NHD2 at 47 264 cmϪ1 ͑see also Fig.
4͒. The smooth curve is calculated for a statistical distribution over all the
vibrational, rotational and translational degrees of freedom of the fragments,
representing dissociation following internal conversion to high levels of the
parent molecular ground state. This has been scaled to indicate how the total
flux might be made up from this smoothly varying internal conversion con-
tribution and a structured contribution from tunneling.
Let us now use the H/D branching ratios to assess the
relative rates from tunneling ͑kTUN͒ and internal conversion
͑kIC͒ following 210 band excitation. Note that the experimen-
tal ratios for both pairs of isotopomer dissociation processes
are intermediate between the statistical values of Eqs. ͑9͒ and
͑10͒ and the purely tunneling values of Eqs. ͑7͒ and ͑8͒. For
NHD2, the experimental H/D ratio can be reproduced using a
ratio of rates ͑kIC͒/͑kTUN͒ϭ0.3. This predicts that the TKER
spectrum for the ͓DϩNHD͔ channel should include an IC
continuum which is 70% of the total signal, which is in
qualitative agreement with the conclusions presented in Fig.
5, but that for the ͓HϩND2͔ channel IC would contribute
only 10%. For NH2D the equivalent required ratio of rates is
also ͑kIC͒/͑kTUN͒ϭ0.3. The corresponding predicted propor-
tions of the IC continuum for the ͓HϩNHD͔ channel is 25%,
͓and for ͑DϩNH2, not measured͒ is 60%͔. These various IC
contributions to the TKER spectra are also qualitatively cor-
rect ͑see Fig. 4͒, but tend to underestimate the continuum
contribution. However, note from Fig. 4 that there is also a
minor continuum contribution to the ͓DϩND2͔ TKER spec-
trum for ND3 dissociation ͑which was assumed to occur by
pure tunneling͒, the neglect of which in the calculations
would lead to an under estimation of IC across the whole
series. For 000 band excitation the TKER spectra ͑Fig. 1͒
show that internal conversion is of minor importance, even
for ͓NHD2→DϩNHD͔, and a similar analysis is not pos-
sible.
␮
1/2. This leads to total statistical ratios of ͓1.88:1͔ for
͓NH2D→HϩNHD͔:͓NH2D→DϩNH2͔, and ͓1.89:1͔ for
͓NHD2→HϩND2͔:͓NHD2→DϩNHD͔.
For both parent molecules the NHD product can be
formed by the breaking of either of two bonds, and there is
no nuclear permutation restriction. However, there is only
one route to yield either NH2 or ND2, and only one-half of
the product states have the same nuclear permutation sym-
metry as any one parent state. The overall calculated branch-
ing ratios between the two pairs of channels are as follows.
Internal conversion:
NH D→HϩNHD͔:͓NH D→DϩNH ϭ 7.5:1 ,
͑9͒
͓
͓
͔
͓
͔
2
2
2
NHD →HϩND ͔:͓NHD →DϩNHD ϭ 0.47:1 . ͑10͒
͔
͓
͔
2
2
2
On this basis the greatest contributions of internal conversion
to the overall dissociation rate across the whole series of
isotopomers should be for ͓NH2D→HϩNHD͔ and for
͓NHD2→DϩNHD͔. These are precisely the processes for
which the TKER spectra have the highest proportion of
background continuum for Ј ϭ 1, and to a lesser extent for
v2
Ј ϭ 0. Figure 5 compares the TKER spectrum for the
v2
͓NHD2͑210͒→DϩNHD͔ dissociation with a statistical distri-
bution based on a harmonic oscillatorϩrigid rotor density of
states for the three vibrational and three rotational degrees of
freedom of NHD and the relative translation of D. This gives
a faithful reproduction of the continuum underlying the re-
solved structure and it is thus clear that IC provides the
dominant decay route for this process.
VI. CONCLUSIONS
The calculations presented here should only be taken as
qualitative, but they do provide a convincing rationalization
of all the observations. We therefore conclude that both
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
6480
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
TABLE V. Observed and calculated term values/cmϪ1 for ϭ0 levels of NHD(X).
˜
v
NϭKa
NϭKaϩ1
NϭKaϩ2
Obs. Calc.
41
NϭKaϩ3
Obs. Calc.
82
Ka
Obs.
Calc.
Obs.
Calc.
0
0
27
93
14
54
1
2
34
94
95
189
323
150
257
404
127
259
135
254
3
193
200
4
347
345
414
413
495
590
5
6
7
8
526
750
528
749
610
845
610
844
708
957
705
953
814
1076
1374
1708
2076
2478
2913
3378
3875
4402
4958
5541
6154
6793
7459
8151
8869
9614
10 381
1009
1303
1632
1991
2391
2821
3277
3763
4281
4822
5393
5989
6607
1006
1299
1627
1988
2382
2808
3265
3752
4268
4813
5386
5986
6614
7269
7949
8656
9387
1119
1427
1768
2144
2552
2992
3463
3959
4485
5040
5622
6226
6860
1115
1422
1763
2138
2546
2985
3455
3955
4484
5042
5629
6243
6883
7550
8243
8962
9706
1242
1559
1915
2307
2728
3180
3660
4170
4708
5272
5862
1238
1558
1913
2301
2722
3175
3658
4172
4714
5286
5885
6511
7164
7844
8550
9282
10 037
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
4943
quantum tunneling and internal conversion to high levels of
the ground state contribute to the overall dissociation mecha-
proportions which vary over a wide range according to the
parent isotopomer and the specific initial excitation. Qualita-
tive models now exist to rationalize all these various data.
Finally, we comment that ammonia shows some aspects
of its dissociation dynamics which are typical of the small
molecule limit, and others typical of the large molecule limit.
The density of states of the ground state at the energy of the
first excited state is just sufficient to permit internal conver-
sion from some of the longest-lived quasibound levels. For
methane, with one more atom and thus three more internal
degrees of freedom, internal conversion has become facile
and there is little dynamical character to its dissociation. The
ammonia molecule must exhibit more variety to its dissocia-
tion dynamics than any other molecule, and is surely the text
book example of state-selected molecular photodissociation.
nism from excited levels with Ј ϭ 0 or 1, with a relative
v2
importance that varies in a complex way with both the iso-
topomer and the quantum state excited. The statistical route
will add most to the total decay rate for the dissociation of
the least symmetrical isotopomers to give the least symmetri-
cal product ͑NHD͒, most particularly from excited levels
with Ј ϭ 1 which are the levels with the slowest tunneling
v2
rate. The most striking consequence of the effects of symme-
try is that for a single excited quantum state of NHD2 the
TKER spectrum for the ͓HϩND2͔ channel has a dominant
structured and dynamical character, whereas that for the
competitive ͓DϩNHD͔ channel is largely statistical. Perhaps
it was fortunate that we initiated our study of ammonia dis-
sociation dynamics by studying the lowest excited levels of
the most symmetrical isotopomers!
We have assumed above that the rates of IC are indepen-
dent of those for tunneling, consistent with IC directly from
the inner well of the excited state. It is possible that part of
the outgoing flux on the excited surface could be reflected
back onto the ground state surface at the conical intersec-
tions, providing an indirect route to IC. Were this to be the
dominant IC route the IC rates would be proportional to the
tunneling rates, and our model calculations of the branching
ratios would be far less successful.
In conclusion, accurate dissociation energies are now
available for all the members of the NHxD3Ϫx series of iso-
topomers of ammonia. The dissociation dynamics of their
first excited states has been shown to involve the mecha-
nisms of quantum tunneling, of IVR through vibrational an-
harmonicity and/or Coriolis coupling, and through IC, in
ACKNOWLEDGMENTS
Once again it is a pleasure to acknowledge Professor K.
H. Welge and Dr. L. Schnieder ͑Universitat Bielefeld͒ for the
¨
help, advice, and encouragement they have given us with all
of our H atom photofragment translational spectroscopy
work. We are also grateful to colleagues in Bristol for their
contributions to this work, to professor K. Lehman for help-
ful discussion, and to SERC ͑now EPSRC͒ for equipment
grants and a studentship ͑to D.H.M.͒. Finally we acknowl-
edge the award of a European Commission individual Re-
search Fellowship administered under the Human Capital
and Mobility Program ͑to D.H.M.͒ and a Royal Society Le-
verhulme Trust Senior Research Fellowship ͑to M.N.R.A.͒
during the tenure of which this work was written.
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44

˜
Mordaunt, Dixon, and Ashfold: Photodissociation of A state ammonia molecules. II
6481
TABLE VI. Potential energy parameters for NHD(X 2A ).
˜
Љ
The optimum parameters, most of which were taken over
from NH2,13 are given in Table VI.
fsϭ3.0ϫ105 ͑cmϪ1 ÅϪ1
͒
͒
fbϭ33 271.2 ͑cmϪ1 radϪ2
1 P. Andresen, G. S. Ondrey, B. Titze, and E. W. Rothe. J. Chem. Phys. 80,
2548 ͑1984͒.
hϭ12 023.6 ͑cmϪ1
␪0ϭ76.25 ͑deg͒
r0ϭ1.0 ͑Å͒
͒
2 D. H. Mordaunt, I. R. Lambert, G. P. Morley, M. N. R. Ashfold, R. N.
Dixon, C. M. Western, L. Schnieder, and K. H. Welge. J. Chem. Phys. 98,
2054 ͑1993͒.
c1ϭϪ167.8 ͑cmϪ1
͒
3 A. E. Douglas, Farad. Disc. Chem. Soc. 35, 158 ͑1963͒.
4 J. Biesner, L. Schnieder, J. Schmeer, G. Ahlers, X. Xie, K. H. Welge, M.
N. R. Ashfold, and R. N. Dixon, J. Chem. Phys. 88, 3607 ͑1988͒.
5 J. Biesner, L. Schnieder, G. Ahlers, X. Xie, K. H. Welge, M. N. R. Ash-
fold, and R. N. Dixon, J. Chem. Phys. 91, 2901 ͑1989͒.
6 M. N. R. Ashfold, R. N. Dixon, S. J. Irving, H.-M. Koeppe, W. Meier, J.
R. Nightingale, L. Schnieder, and K. H. Welge, Philos. Trans. R. Soc.
London Ser. A 332, 375 ͑1990͒.
APPENDIX: TERM VALUES FOR ROTATIONAL
LEVELS OF THE GROUND STATE OF NHD (X 2AЉ)
˜
The experimental rotational term values for NHD ͑Table
V͒ are derived from the TKER spectrum for
͓NH2D͑000͒→DϩNHD͔ ͑Fig. 2͒, with an extrapolated maxi-
mum kinetic energy release of 8663Ϯ5 cmϪ1. These were
fitted by least squares to model calculations in which, for
each ͉N,Ka level, the term value was estimated as the mini-
͘
mum value with respect to ␪ and r of the effective potential
function:
7 G. Herzberg and J. Shoosmith, Canad. J. Phys. 34, 523 ͑1956͒.
8 S. H. S. Wilson, J. D. Howe, K. N. Rosser, M. N. R. Ashfold, and R. N.
Dixon. Chem. Phys. Lett. 227, 456 ͑1994͒.
9 A. Hodgson, J. P. Simons, M. N. R. Ashfold, J. M. Bayley, and R. N.
Dixon. Molec. Phys. 54, 351 ͑1985͒.
10 H. J. Krautwald, L. Schnieder, K. H. Welge and M. N. R. Ashfold, Farad.
Discuss. Chem. Soc. 82, 99 ͑1986͒.
11 D. H. Mordaunt, M. N. R. Ashfold, and R. N. Dixon, J. Chem. Phys. 100,
7360 ͑1994͒.
2
Erot N,K ͒ϭV ␪,r͒ϩA ␪,r͒ K ͒
͑
͑
͑
͑
a
a
12 D. H. Mordaunt, M. N. R. Ashfold and R. N. Dixon, preceding paper, J.
Chem. Phys. 104, 6460 ͑1996͒.
ϩ 12 B ␪,r͒ϩC ␪,r͒ N Nϩ1͒Ϫ K ͒
,
2
͓ ͑ ͔͓ ͑ ͔
͑
͑
a
13 R. N. Dixon, S. J. Irving, J. R. Nightingale, and M. Vervloet, J. Chem.
Soc. Faraday Trans. 87, 2121 ͑1991͒.
͑A1͒
14 G. Duxbury and R. N. Dixon, Mol. Phys. 43, 255 ͑1981͒.
15 S. A. Henck, M. A. Mason, W.-B. Yan, K. K. Lehmann, and S. L. Coy, J.
Chem. Phys. 102, 4772, 4783 ͑1995͒.
where ␪ is the external interbond angle, r is the NH and ND
bondlength, A(␪,r), B(␪,r), and C(␪,r) are ‘‘rotational
constants’’ in a principal axis system, and V(␪,r) is a poten-
tial function to be optimized. This model takes explicit ac-
count of both centrifugal stretching and centrifugal bending,
although it does impose equality of the two bondlengths. The
chosen potential function is14,25
16 J. Berkowitz, G. B. Ellison, and D. Gutman, J. Phys. Chem. 98, 2744
͑1994͒.
17 A. R. Hoy, I. M. Mills, and G. Strey, Mol. Phys. 24, 1265 ͑1972͒.
18 L. O. Hargiss and W. C. Ermler, J. Phys. Chem. 92, 300 ͑1988͒.
19 W. Gabriel, G. Chambaud, P. Rosmus, S. Carter, and N. C. Handy, Mol.
Phys. 81, 1445 ͑1994͒.
20 R. N. Dixon, Mol. Phys. 68, 263 ͑1989͒.
21 P. Rosmus, P. Botschwina, H.-J. Werner, V. Vaida, P. C. Engelking, and
M. I. McCarthy, J. Chem. Phys. 86, 6677 ͑1987͒.
2
hfb ␪2Ϫ␪ ͒
͑
0
V ␪,r͒ϭf rϪr ͒2ϩ
͑
͑
r
0
2
0
2
0
2
22 M. I. McCarthy, P. Rosmus, H.-J. Werner, P. Botschwina, and V. Vaida,
J. Chem. Phys. 86, 6693 ͑1987͒.
f ␪ ϩ 8hϪ␪ ͒␪
͔
͓
͑
b
23 M. N. R. Ashfold, C. L. Bennett, and R. N. Dixon, Chem. Phys. 93, 293
4
c1 3␪4Ϫ␪ ͒cos 2␲␪/5␪ ͒sin2 ␲␪/␪ ͒
͑
͑
͑
0
0
0
͑1985͒.
ϩ
.
24
4
␪4ϩ␪ ͒
¨
U. Manz, P. Rosmus, H-J. Werner, and P. Botschwina, Chem. Phys. 122,
͑
0
387 ͑1988͒.
͑A2͒
25 Ch. Jungen, K.-E. J. Hallin, and A. J. Merer, Mol. Phys. 40, 25 ͑1980͒.
J. Chem. Phys., Vol. 104, No. 17, 1 May 1996
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:
155.247.166.234 On: Sat, 22 Nov 2014 15:18:44