State-to-state predissociation dynamics of OC-HF upon HF stretch excitation

Article abstract of DOI:10.1063/1.1288605

The optothermal detection method in association with two F-center lasers were employed to study the vibrational predissociation of the OC-HF complex. The fundamental H-F stretching vibration of the OC-HF complex was pumped using an F-center laser. The rotational states of the HF fragment were probed as a function of recoil angle using a second F-center laser. Photofragment angular and state distributions were measured following the vibrational predissociation of the OC-HF complex. The dissociation energy and zero point energy were determined.

Full text of DOI:10.1063/1.1288605

The state-to-state predissociation dynamics of OC–HF upon HF stretch excitation
L. Oudejans and R. E. Miller
Citation: The Journal of Chemical Physics 113, 4581 (2000); doi: 10.1063/1.1288605
View online: http://dx.doi.org/10.1063/1.1288605
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/113/11?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Potential energy surface and rovibrational energy levels of the H2-CS van der Waals complex
J. Chem. Phys. 137, 234301 (2012); 10.1063/1.4771658
The rotational dynamics of N 2 – HF and OC–HF in helium nanodroplets
J. Chem. Phys. 117, 4846 (2002); 10.1063/1.1489900
Erratum: “State-to-state vibrational predissociation dynamics of the acetylene–HF complex” [J. Chem. Phys. 110,
209 (1999)]
J. Chem. Phys. 110, 7109 (1999); 10.1063/1.478564
State-to-state vibrational predissociation dynamics of the acetylene-HF complex
J. Chem. Phys. 110, 209 (1999); 10.1063/1.478096
Vibrational predissociation dynamics of ArHF (3000) and (3110): Lifetimes and HF product state distributions
J. Chem. Phys. 109, 8836 (1998); 10.1063/1.477554
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:
131.230.73.202 On: Mon, 22 Dec 2014 15:21:46

JOURNAL OF CHEMICAL PHYSICS
VOLUME 113, NUMBER 11
15 SEPTEMBER 2000
The state-to-state predissociation dynamics of OC–HF
upon HF stretch excitation
L. Oudejans and R. E. Miller
Department of Chemistry, University of North Carolina, Chapel Hill, North Carolina 27599
͑Received 11 May 2000; accepted 19 June 2000͒
Photofragment angular and state distributions have been measured following the vibrational
predissociation of the OC–HF complex. An F-center laser is used to pump the fundamental H–F
stretching vibration of the complex and a second F-center laser is used to probe the rotational states
of the HF fragment as a function of recoil angle. The complex dissociates via two different sets of
channels, one that produces COϭ1, JHFϭ6,5,4 ͑intermolecular V–V transfer͒ and the other
v
vCO
ϭ0, JHFϭ11 ͑V–R transfer͒. Analysis of the data gives correlated final state distributions, as well
as an accurate value for the dissociation energy (D0) of the complex, namely 732Ϯ2 cmϪ1
© 2000 American Institute of Physics. ͓S0021-9606͑00͒00235-X͔
.
INTRODUCTION
longer than that of the H–F fundamental to explain the ex-
perimental results. The implication seemed to be that there is
very weak coupling between the corresponding bending ͑or
internal rotation͒ degrees of freedom and the dissociation
coordinate, namely the intermolecular stretching mode, such
that the complex can store energy in the intermolecular
modes ͑in excess of the bond energy͒ for a significant period
of time.
The OC–HF complex has also been the subject of con-
siderable theoretical work, including several that provide
predictions of the intermolecular bond dissociation
energy.24,25,27–32,34 The present state-to-state experiments
provide further insights into the nature of the energy transfer
processes that lead to dissociation and an accurate value for
the dissociation energy of the complex.
Considerable progress has recently been made in both
the spectroscopic and dynamical characterization of weakly
bound complexes. The nonstatistical nature of the vibrational
dynamics of these complexes is of particular interest since
the preferences these systems show for dissociating in spe-
cific ways provide insights into the nature of the interactions
that lead to rupture of the weak bond. For very weakly in-
teracting systems, such as Ar–HCl ͑Ref. 1͒ and Ar–HF,2,3
the spectroscopic characterization can be quite complete,
providing unique determinations of the corresponding inter-
molecular potentials.4–8 At the other extreme are the hydro-
gen bonded complexes, including the HF dimer9,10 and
HCN–HF,11 where even the extensive experimental studies
that have been reported are insufficient to completely deter-
mine the potential energy surface. In these cases, ab initio
calculations are often used to interpolate between the regions
that are probed by the experiments. The linear N2 –HF ͑Refs.
12–15͒ and OC-HF ͑Refs. 16–21͒ complexes fall between
these two limits, being characterized as having weak hydro-
gen bonds. Although there have been a large number of de-
tailed experimental12–21 and theoretical studies15,22–34 of
these complexes, a quantitative interaction potential is not
available for either.
In the present study we use the optothermal detection
method in combination with two F-center lasers to study the
vibrational predissociation of the OC–HF complex. The first
studies of this complex were carried out using microwave
spectroscopy,16,17 which revealed it to have a linear structure.
More recent infrared studies have provided detailed informa-
tion on a number of intramolecular vibrational modes of the
complex,18–21 as well as some insights into the nature of the
associated vibrational dynamics. One of the interesting as-
pects of the latter came from our studies of the H–F stretch
fundamental, which showed a very interesting perturbation
that could only be explained in terms of a ͑dark͒ state that
has more energy in the intermolecular degrees of freedom
than the dissociation energy of the complex.19 Despite this,
the vibrational predissociation lifetime of this state has to be
EXPERIMENT
The experimental approach used in this experiment in-
volves measuring the photofragment angular distribution re-
sulting from vibrational predissociation of the OC–HF com-
plex, following excitation of the H–F stretching vibration.
The translational energy release for a given final state chan-
nel, which determines the laboratory recoil angle, is anticor-
related with the total internal energy of the fragments. As a
result, data of this type can provide detailed information on
the final state distributions of the fragments, including the
interfragment correlations. In particular, one can determine
not only the state distribution of the HF fragment, but also
which of these states correlate with specific states of the CO
fragment, for a given dissociation event. To achieve this
level of characterization, it is necessary to assign all of the
structure observed in the angular distribution to specific in-
ternal photofragment states.
The molecular beam apparatus has been discussed in de-
tail elsewhere.35,36 In short, an F-center laser is used to excite
the H–F stretching vibration of OC–HF and the resulting
photofragments are detected using a bolometer detector.37
The source can be rotated so that a complete photofragment
angular distribution can be measured for a given initial state
0021-9606/2000/113(11)/4581/7/$17.00
4581
© 2000 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:
131.230.73.202 On: Mon, 22 Dec 2014 15:21:46

4582
J. Chem. Phys., Vol. 113, No. 11, 15 September 2000
L. Oudejans and R. E. Miller
of the complex. The F-center laser is equipped with an
evacuated chamber containing three confocal etalons with
free spectral ranges of 7.5 GHz, 750 MHz, and 150 MHz.
These are used in combination with a wavemeter to monitor
the tuning of the lasers. To obtain a full angular distribution,
the pump laser must be kept in resonance with an OC–HF
transition for approximately 45 min. This presented no seri-
ous problems in this case owing to the fact that the corre-
sponding transitions were quite broad.19 Angular distribu-
tions for the particular initial state of the complex were
recorded by varying the angle between the molecular beam
and the line joining the photolysis volume to the detector in
0.25° increments. The detailed kinematics that relate these
angular distributions to the associated kinetic energy release
for a given dissociation channel have been discussed
elsewhere.35,38
A second F-center laser is used to probe the internal
states of the HF fragment. In this case, the bolometer detects
the extra vibrational energy deposited in the HF fragment by
the probe laser. In the present study, we amplitude modu-
lated the pump laser, which was fixed at the OC–HF transi-
tion frequency, and scanned the cw probe laser through the
HF monomer transitions of interest. This modulation scheme
has the advantage that probe laser signals are entirely due to
HF monomer that is produced by the photolysis laser. One
disadvantage is that the probe laser signal is detected on top
of the ͑much larger͒ photofragment signal. Fortunately these
background signals were quite stable and an average of 25
probe laser scans gave reliable results at each laboratory
angle.
FIG. 1. A pendular state spectrum of the OC–HF complex, recorded at a
field strength of 19 kV/cm. Note that the field was switched off while scan-
ning the high frequency portion of the spectrum in order to record the field
free R(0) transition. The vertical arrow at
indicates the position of the
v0
field free vibrational origin of the complex. The solid line through the pen-
dular spectrum was calculated based upon the molecular constants obtained
previously for this complex ͑Ref. 19͒. The inset shows the orientation dis-
tribution for the complex at this applied field.
pear near the vibrational band origin of the spectrum, as
shown in Fig. 1. This spectrum was recorded by first scan-
ning through the R(0) transition with the dc field off ͑pro-
viding an absolute frequency calibration͒ and then turning
the field on to record the pendular spectrum. The field-
induced transitions are spaced by an amount that depends on
the difference in the rotational constant and dipole moment
between the vibrational ground and vibrational excited
states. The smooth solid line through the data is a calculation
based upon the molecular constants for OC–HF reported
elsewhere.19 The good agreement ensures that we have prop-
erly assigned the corresponding states and that the orienta-
tion distribution used to analyze the photodissociation ex-
periments is correct. The inset in the figure shows the
orientation distribution at the electric field corresponding to
the one used in the spectrum, namely, 19 kV/cm.
We began the photodissociation studies by recording an
angular distribution in the absence of the electric field, cor-
responding to the R(0) transition shown in Fig. 1. This tran-
sition provides the best resolution in the photofragment dis-
tributions ͑assuming that dissociation occurs via axial recoil͒
when the laser polarization is aligned perpendicular to the
molecular beam axis.38 In this case, the excited state align-
ment imposed by the excitation laser ensures that the most
probable recoil direction in the laboratory frame is orthogo-
nal to the molecular beam, giving rise to the maximum pos-
sible laboratory frame recoil angle for a given translational
energy release. As a result, the fragment channels with dif-
ferent translational energies scatter to quite different labora-
tory angles. In contrast, if the laser polarization is aligned
parallel to the molecular beam direction, the photofragments
are preferentially ejected along the beam and the various
final state channels all scatter near zero degrees in the labo-
As discussed in detail previously,39–43 extensive use has
been made in our laboratory of the pendular state orientation
method to enhance the information content in these experi-
mental angular distributions. This involves orienting the par-
ent molecule in the laboratory frame, such that the two frag-
ments ͑in this case CO and HF͒ recoil in opposite directions
in the laboratory frame and can be detected separately. Ori-
entation of the parent complex is achieved by applying a
moderate dc electric field ͑19 kV/cm͒ to the photolysis re-
gion and taking advantage of the fact that the dipole moment
of the OC–HF complex will be oriented in the direction of
the electric field. Since the bolometer is an energy detector,
the relative signal levels depend upon the energy content of
the two fragments. This information is extremely helpful in
determining the internal states of the two cofragments.
In the present study the molecular beam was formed by
expanding a mixture of 0.6% HF and 30% CO in helium
through a 40 ␮m diam nozzle, from a source pressure of 550
kPa. In order to relate the recoil angle of a fragment to the
associated recoil energy, an accurate measurement of the
stream velocity of the molecular beam is needed. This was
accomplished by Doppler spectroscopy, yielding a stream
velocity of 1300 m/s for the present expansion conditions.
RESULTS
In all of the experiments reported here the laser polar-
ization was aligned parallel to the dc field, such that the
relevant selection rule is ⌬Mϭ0.39 Under these conditions,
the electric field induced transitions of a linear molecule ap-
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:
131.230.73.202 On: Mon, 22 Dec 2014 15:21:46

J. Chem. Phys., Vol. 113, No. 11, 15 September 2000
Predissociation of OC–HF
4583
FIG. 4. A set of pendular state angular distributions, corresponding to ex-
citation of the most intense pendular transition in Fig. 1. This transition
pumps the Mϭ0 excited state, which is the most highly oriented by the
electric field. The HF and CO fragments appear at positive and negative
angles, respectively. The solid line through the data points is the result of a
fit to the data, as described in the text.
FIG. 2. An angular distribution obtained by pumping the R(0) transition of
the OC–HF complex, with the laser polarization direction aligned perpen-
dicular to the molecular beam axis. The dashed line through the data is a
calculation based upon the probabilities determined from the pendular state
results. The good agreement indicates that the applied electric field used to
obtain the oriented angular distributions has little or no influence on the
photodissociation dynamics of this system.
are clearly many open channels, so many in fact that if all
were populated statistically, we would not observe the dis-
tinct peaks evident in Fig. 2. From this we can immediately
conclude that the dissociation is nonstatistical for this sys-
tem, a situation that is generally encountered in the dissocia-
tion of these weakly bound systems. The second problem is
that we do not have an accurate value for the dissociation
energy of this complex, so that the location of the horizontal
line in Fig. 3, which represents the available energy upon
excitation of the H–F stretch, is not known. The final prob-
lem is that the bolometer used to measure the angular distri-
bution is sensitive to both the CO and HF fragments. Thus
the angular distribution shown in Fig. 2 has contributions
from two fragments of different masses. Since conservation
of momentum tells us that these two fragments have different
recoil velocities, this angular distribution is really the super-
position of two distributions. Thus each final state channel
will give rise to two peaks in the angular distribution, mak-
ing its assignment even more difficult. In what follows we
will address these problems in turn.
ratory frame. A detailed discussion of these zero electric field
cases, in relation to the photodissociation of the HF dimer,
can be found elsewhere.38
The angular distribution obtained by pumping the R(0)
transition is shown in Fig. 2. From a first inspection of the
data it becomes clear that there are a number of important
final state channels having different recoil energies. This is
fortunate since our goal here is to assign the individual peaks
in the angular distribution to specific final state channels and
the more structure that is present in the angular distributions,
the better our chances of making a unique assignment. There
are several problems that must be addressed if we are to
explain the structure observed in this angular distribution.
The first is simply to choose from the many open channels,
the few that contribute to the distribution. Figure 3 shows an
energy level diagram that illustrates the difficulty here. There
We begin with the problem of separating the angular
distributions associated with the CO and HF fragments. As
indicated above, this can be done using the pendular state
orientation method. As shown by the inset in Fig. 1, the
OC–HF complex is oriented by the strong field, such that the
HF and CO fragments recoil in opposite directions, towards
the positive and negative electrodes, respectively. As a re-
sult, angular distributions recorded on the positive and nega-
tive electrode sides of the apparatus, while pumping the most
intense pendular state transition in Fig. 1, will correspond to
the HF and CO fragments, respectively. Figure 4 shows a set
of angular distributions obtained in this way. The structure in
the angular distributions is now better resolved and each
peak must be due to a different photofragment channel. It is
interesting to note that the relative intensities of the two main
peaks observed on either side are different, the preliminary
indication that they come from channels with very different
partitioning of the available energy between the two frag-
FIG. 3. An energy level diagram for the OC–HF system. The horizontal
dashed line indicates the available energy ͑3112 cmϪ1͒, based upon the laser
excitation energy ͑3844 cmϪ1͒ minus the dissociation energy ͑732 cmϪ1
͒
determined from the experimental results. All open rotational states of the
HF fragment are depicted, upon which are built the CO rotational levels.
Two main sets of channels can be identified corresponding to COϭ0 and
v
ϭ1. HF rotational states observed in the present study are indicated ͑ ͒.
vCO
*
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:
131.230.73.202 On: Mon, 22 Dec 2014 15:21:46

4584
J. Chem. Phys., Vol. 113, No. 11, 15 September 2000
L. Oudejans and R. E. Miller
FIG. 6. A comparison between the HF angular distribution and that obtained
by probing the HF P(6) transition. A typical probe laser scan through this
transition is shown as an inset. The dashed line through the probe data
FIG. 5. A comparison between the HF angular distribution and that obtained
by probing the HF P(11) transition. A typical probe laser scan through this
transition is shown as an inset. The dashed line through the probe data
corresponds to the COϭ1, JHFϭ6 contribution to the fit in Fig. 4.
v
shows the COϭ0, JHFϭ11 contribution to the fit in Fig. 4.
v
ments. The fact that the peak at smaller angles is relatively
larger on the CO side of the distribution, suggests that the
corresponding photofragment channel has a significant frac-
tion of the available energy in the CO fragment. This will be
confirmed in what follows. The fact that we do not known
the dissociation energy of this complex still makes the as-
signment of these peaks to specific final state channels some-
what challenging. Indeed, by varying the dissociation energy
we can bring any number of channels to the correct energy to
account for the peaks in the angular distribution.
contributes most significantly in this region. To test this, we
recorded a state specific angular distribution for JHFϭ6,
which is shown in Fig. 6. Once again, the inset shows a
typical probe laser scan of the P(6) transition. Note that the
relative intensities of the state specific angular distributions
in Figs. 5 and 6 have been scaled so that their sum gives the
best fit to the overall distribution, which is in qualitative
agreement with the relative transition intensities. It is clear
from Fig. 6 that the JHFϭ6 state only contributes at smaller
angles and peaks in the correct place to account for the dif-
ferences between the two distributions in Fig. 5. The fact that
the JHFϭ6 channel appears at smaller angles than JHFϭ11
can be appreciated from the energy level diagram in Fig. 3.
Given the available energy, the JHFϭ11 channel can only be
The next step is to use the second, probe laser to mea-
sure angular distributions corresponding to individual HF ro-
tational states. It is important to note that only the ϭ0 state
v
of the HF fragment is energetically accessible in these ex-
periments. To perform such a pump–probe experiment, we
first positioned the apparatus near an intensity maximum of
the HF fragment distribution and searched for signals asso-
ciated with the various HF rotational transitions ͓R(0),P(1)
through P(13)͔. At 10° we found that only the transitions
originating from the JHFϭ11 and JHFϭ6 states of HF had
significant intensity. We should note here that the P(5) tran-
sition was inaccessible due to the fact that it overlaps a
strong water vapor transition. The present result is quite
similar to that observed for N2–HF, where only the JHF
ϭ12 and JHFϭ7 states were populated, corresponding to
production of the N2 fragment in the ground and vibra-
tionally excited states, respectively.40
produced in coincidence with the COϭ0 state. On the other
v
hand, if JHFϭ6 where produced in coincidence with
vCO
ϭ0 and the fragments were detected at 10°, the CO fragment
would have to carry away approximately two thousand wave
numbers of rotational energy. Given the small rotational con-
stant for CO ͓1.9225 cmϪ1 ͑Ref. 44͔͒, this would necessitate
populating the JCOϭ39 state. Such a large change in rota-
tional quantum number upon dissociation seems unlikely and
we have never seen such in the systems we have studied. As
a result, we conclude that the JHFϭ6 channel is produced in
coincidence with the COϭ1 state. As Fig. 3 illustrates, this
v
places the two channels at just the right relative energies to
explain the experimental data. Clearly the internal energy of
Figure 5 shows a comparison between the total HF an-
gular distribution obtained from the pendular state excitation
and that obtained by probing only the JHFϭ11 state. The
inset in the figure shows a typical probe spectrum corre-
sponding to the P(11) transition. Although the signal to
noise ratio of the state specific angular distribution is rather
poor, it clearly shows that most of the signal at larger angles
comes from the JHFϭ11 channel. The most prominent dif-
ference between the total angular distribution and that corre-
sponding to JHFϭ11 is that the latter does not show the
pronounced peak near 10°, suggesting that the JHFϭ6 state
the JHFϭ6, COϭ1 channel is larger than that of the JHF
v
ϭ11, COϭ0 channel, explaining why the latter scatters to
v
larger angles.
The sharp cut off of the JHFϭ6 channel is helpful in that
it provides an accurate determination of the dissociation en-
ergy of the complex. The highest recoil energy occurs when
the CO is produced in the JCOϭ0 state. The best overall fit
to the data was obtained with a dissociation energy of 732
cmϪ1, which is the value used in Fig. 3. The solid line
through the experimental data in Fig. 4 is the result of a least
squares fit that uses a Monte Carlo procedure to include 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:
131.230.73.202 On: Mon, 22 Dec 2014 15:21:46

J. Chem. Phys., Vol. 113, No. 11, 15 September 2000
Predissociation of OC–HF
4585
instrumental factors that control the resolution of the appa-
ratus. The only fitting parameters were the individual prob-
abilities for the various open channels. The dashed lines
through the state specific angular distributions in Figs. 5 and
6 are the contributions from the corresponding states, ob-
tained from the fit. This excellent fit to all of the data could
only be obtained for a dissociation energy of 732Ϯ2 cmϪ1
,
making the assignment unique.
It is important to point out that the relative intensities on
the CO and HF sides of the angular distribution provide fur-
ther confirmation that we have properly assigned the final
state channels. Indeed, other combinations were tried but did
not give the correct relative intensity. This is due to the fact
that the bolometer is an energy detector, so that the relative
intensities depend upon which fragment carries away the
larger fraction of the available energy. We should also point
out that the final fit included contributions from the JHFϭ5,
FIG. 7. Comparisons between the experimental angular distributions ͑data
points͒ and those of a full statistical calculation ͑solid line͒ and a restricted
phase space calculation ͑dashed line͒, as described in the text.
COϭ1 and JHFϭ4, COϭ1 channels ͑even though we did
v
v
not probe these͒ to account for some of the signals at larger
angles. Fits that did not include these channels showed too
much structure in the angular distribution at large angles and
did not give the correct relative intensities for the CO and HF
angular distributions. Therefore, although we do not have
probe laser results for JHFϭ4 and 5, the data strongly sug-
gests that these are populated. Further support for this comes
from the fact that the probe laser induced signals for the
case, the agreement is still rather poor, due to the fact that the
statistical calculation tends to overestimate the contributions
from the higher CO rotational states.
It is important to consider whether or not the electric
field, used to orient the complex, has any effect on the state-
to-state dynamics. To test this, we used the state-to-state
probabilities obtained from the pendular state angular distri-
butions to calculate the angular distribution for the field free
R(0) transition. The result is shown as the solid line in Fig.
2, which is in reasonably good agreement with experiment.
This is in agreement with the previous systems we have
studied42,45,50 that also show that the electric field has no
significant influence on the state-to-state dynamics.
Finally, we consider the states that showed interesting
perturbations in the spectroscopy at zero electric field.19 The
pendular spectrum showed no evidence of perturbations so
that studies of these had to be carried out at zero electric
field. In view of the fact that the character of the perturbing
state is known to be quite different from the H–F stretch
fundamental, we might expect that the corresponding final
state distributions would also be quite different. Thus experi-
ments designed to probe these final states could provide fur-
ther insights into the nature of the corresponding dark states.
Angular distributions were recorded for the two components
of the perturbed R(6) transition. However, within the experi-
mental uncertainties the two angular distributions were the
same. In retrospect, this is not unexpected, given that the
perturbing state has a significantly longer predissociation
lifetime than the H–F fundamental.19 Thus, even though the
final state distribution from the dark state is likely to be quite
different from the bright state ͑the H–F stretch͒, the disso-
ciation rate out of the former is too slow to contribute sig-
nificantly to the experimental results.
J
HFϭ11 state of HF ͑relative to the total signals͒ are smaller
in this system than in others we have studied, where we were
certain that only JHFϭ11 contributed.45 Indeed, experiments
on OC–HF and acetylene–HF, carried out back-to-back,
consistently gave probe laser signals for the former that were
25% smaller, relative to the total signal, than the correspond-
ing signals for the latter. The implication is that in OC–HF
͑at larger angles͒ there are contributions to the total signal
from states other than JHFϭ11, in qualitative agreement with
the results from the fit. From the fit shown in Fig. 4, a com-
plete set of individual state-to-state probabilities was ob-
tained for the various state channels and is available through
EPAPS.46 Nevertheless, some of these probabilities are inter-
dependent. Therefore, the best way to make comparisons
with theoretical calculations of the state-to-state probabilities
is to calculate the full angular distribution. For the present
purposes, we note that the probabilities obtained from these
channels indicate that the V–V and V–R channels are
equally probable to within the experimental uncertainty.
The state distribution obtained from fitting the experi-
mental results is highly nonstatistical. This is further illus-
trated in Fig. 7, which shows a comparison between the ex-
perimental angular distributions and the results of a purely
statistical calculation,47,48 the latter shown as the solid line.
In the past41,42,45,49 we have found much better agreement
with a restricted phase space calculation, in which the two
high frequency modes of the system, namely, the HF rotation
and the CO vibration in this case, were constrained to agree
with the experimental results. In this case, only the rotation
of the heavy fragment and the translational degrees of free-
dom are treated statistically. The dashed line in Fig. 7 shows
the result of such a calculation. Although the overall width of
the angular distribution is more faithfully reproduced in this
DISCUSSION
In the previous section we presented detailed results re-
lated to the state-to-state photodissociation dynamics of 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:
131.230.73.202 On: Mon, 22 Dec 2014 15:21:46

4586
J. Chem. Phys., Vol. 113, No. 11, 15 September 2000
L. Oudejans and R. E. Miller
OC–HF complex, corresponding to excitation of the H–F
stretch. Broadly speaking, dissociation proceeds via two dis-
tinctly different channels, namely a V–V channel, where
most of the energy is transferred from the HF molecule to the
CO fragment and a V–R channel, corresponding to the HF
fragment carrying away most of the energy in rotation. The
state-to-state probabilities can now be used to determine the
partitioning of the available energy among the various de-
grees of freedom of the fragments. The total available energy
to the fragments is simply the difference between the photon
excitation energy ͑3844 cmϪ1͒ and the dissociation energy
determined above ͑732 cmϪ1͒, namely, 3112 cmϪ1. Consider
first the V–R channel, for which the CO fragment is pro-
duced in the ground vibrational state, so that ECv Oϭ0. The
initio potential surface is probably much better than this lack
of agreement would imply.
SUMMARY
In the previous sections we have presented data that
gives a rather complete picture of the state-to-state photodis-
sociation dynamics of the OC–HF complex, resulting from
excitation of the H–F stretching vibration. The complex is
observed to dissociate via two distinctly different channels,
namely, a V–R process which produces the HF fragment in
Jϭ11 and an intermolecular V–V channel, for which vibra-
tionally excited CO ( ϭ1) is produced in combination with
v
the HF fragment in Jϭ6, 5, and 4. This assignment of the
final state channels is based upon simultaneous fits to the
angular distributions of the HF and CO fragments and the
direct measurements of the state-specific angular distribu-
tions for JHFϭ11 and 6. The correlated final state distribu-
tions show that the vibrational predissociation of this com-
plex is highly nonstatistical. What is surprising about these
results is that two such different channels have similar rates.
Indeed, the V–R channel relies on coupling provided by the
anisotropy of the intermolecular potential in order to access
the JHFϭ11 state. On the other hand, the V–V channel re-
sults from a coupling between the stretching modes of the
two diatomic molecules provided by the intermolecular in-
teraction. The fact that these two are of comparable magni-
tude is surprising. What is even more interesting is the fact
that the same is observed in N2–HF, where we again observe
both channels.40 Full 6D potentials, along with the corre-
sponding dynamical calculations, are needed to provide us
with further insights into the reason for this interesting coin-
cidence. The dissociation energy of the complex is deter-
energy of the JHFϭ11 state is 2677 cmϪ1 51
thus accounting
,
for most of the available energy ͑86%͒. If we average over
the rotational states of the CO fragment, the average transla-
tional energy release for this channel is Etϭ254 cmϪ1 ͑8%͒,
while ECROϭ181 cmϪ1 ͑6%͒. Next, we consider the V–V
channel, where the majority of the available energy now ap-
pears as vibration of the CO fragment, namely, EvCO
ϭ2143 cmϪ1 ͑69%͒. In this case, the average energy in HF
rotation is EHJ Fϭ555 cmϪ1 ͑18%͒, while ECJ Oϭ97 cmϪ1
͑3%͒. The translational recoil energy, averaged over the
J
HFϭ6,5,4 states, is Etϭ317 cmϪ1 ͑10%͒. In both of these
channels, therefore, translation and heavy molecule rotation
account for only a small fraction of the excess energy.
The dissociation energy (D0) of the OC–HF complex
has been determined in this study to be 732Ϯ2 cmϪ1. We
can compare this value with previous estimates from both
experiment16 and theory.24,25,27–32,34 The only previous ex-
perimental estimate of D0 was obtained by Legon et al.16 by
fitting the microwave results to a pseudodiatomic model,
yielding 987 cmϪ1. This approach is well known to give
unreliable results, so that the disagreement with the present
results is not surprising. Many ab initio calculations have
been carried out on the OC–HF complex,24,25,27–29,31,32,34 as
well as molecular modeling studies.30 In the present discus-
sion we focus only on the most recent results, which we
assume are the most accurate. The main problem with mak-
ing comparisons between experimental D0 values and those
obtained from ab initio theory is that the latter suffer from
two problems. The first is just the level of the calculations
͑basis sets, etc.͒ and the second is the estimate of the zero
point energy, which is often done using harmonic frequen-
cies. For example, Curtiss et al.28 have reported calculations
at the MP3 level, yielding Deϭ1172 cmϪ1 and D0
ϭ539 cmϪ1. This is to be compared with the most recent
work of Handy and co-workers,31 which yielded De
ϭ1199 cmϪ1 and D0ϭ583 cmϪ1. In both cases, the zero
point energy was estimated from the harmonic frequencies.
Since the latter are larger than the anharmonic frequencies,
we can assume that the zero point energy correction is too
large and that the true D0 will be somewhat larger. Thus,
although these calculated dissociation energies are expect-
edly smaller than the present experimental value, the com-
parisons are unsatisfactory at the quantitative level. Never-
theless, the results do suggest that the problem is mainly with
the estimate of the zero point energy, suggesting that the ab
mined in the present study to be D0ϭ732Ϯ2 cmϪ1
.
ACKNOWLEDGMENTS
This work was supported by the National Science Foun-
dation ͑CHE-97-10026͒. We also acknowledge the donors of
The Petroleum Research Fund, administered by the ACS, for
partial support of this research.
1 R. C. Cohen and R. J. Saykally, J. Phys. Chem. 96, 1024 ͑1992͒.
2 G. T. Fraser and A. S. Pine, J. Chem. Phys. 85, 2502 ͑1986͒.
3 C. M. Lovejoy, M. D. Schuder, and D. J. Nesbitt, J. Chem. Phys. 85, 4890
͑1986͒.
4 J. M. Hutson, J. Phys. Chem. 96, 4237 ͑1992͒.
5 D. E. Woon, K. A. Peterson, and T. H. Dunning, Jr., J. Chem. Phys. 109,
2233 ͑1998͒.
6 J. M. Hutson, J. Chem. Phys. 96, 6752 ͑1992͒.
7 H.-C. Chang, F. M. Tao, W. Klemperer, C. Healey, and J. M. Hutson, J.
Chem. Phys. 99, 9337 ͑1993͒.
8 T. van Mourik and T. H. Dunning, Jr., J. Chem. Phys. 107, 2451 ͑1997͒.
9 M. Quack and M. A. Suhm, in Conceptual Perspectives in Quantum
Chemistry, edited by J.-L. Calais and E. Kryachko ͑Kluwer Academic,
Dordrecht, 1997͒, pp. 415–463.
10 M. Quack and M. A. Suhm, Advances in Molecular Vibrations and Col-
lision Dynamics ͑JAI, Greenwich, CT, 1998͒, Vol. 3, pp. 205–248.
11 A. McIntosh, A. M. Gallegos, R. R. Lucchese, and J. W. Bevan, J. Chem.
Phys. 107, 8327 ͑1997͒.
12 P. D. Soper, A. C. Legon, W. G. Read, and W. H. Flygare, J. Chem. Phys.
76, 292 ͑1982͒.
13 K. D. Kolenbrander and J. M. Lisy, J. Chem. Phys. 85, 2463 ͑1986͒.
14 K. W. Jucks, Z. S. Huang, and R. E. Miller, J. Chem. Phys. 86, 1098
͑1987͒.
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:
131.230.73.202 On: Mon, 22 Dec 2014 15:21:46

J. Chem. Phys., Vol. 113, No. 11, 15 September 2000
Predissociation of OC–HF
4587
15 D. J. Nesbitt, T. G. Lindeman, J. T. Farrell, Jr., and C. M. Lovejoy, J.
Chem. Phys. 100, 775 ͑1994͒.
35 E. J. Bohac, M. D. Marshall, and R. E. Miller, J. Chem. Phys. 96, 6681
͑1992͒.
36 E. J. Bohac and R. E. Miller, Phys. Rev. Lett. 71, 54 ͑1993͒.
37 T. E. Gough, R. E. Miller, and G. Scoles, Appl. Phys. Lett. 30, 338
͑1977͒.
38 M. D. Marshall, E. J. Bohac, and R. E. Miller, J. Chem. Phys. 97, 3307
͑1992͒.
39 M. Wu, R. J. Bemish, and R. E. Miller, J. Chem. Phys. 101, 9447 ͑1994͒.
40 R. J. Bemish, E. J. Bohac, M. Wu, and R. E. Miller, J. Chem. Phys. 101,
9457 ͑1994͒.
41 L. Oudejans and R. E. Miller, J. Phys. Chem. 99, 13670 ͑1995͒.
42 L. Oudejans and R. E. Miller, J. Chem. Phys. 109, 3474 ͑1998͒.
43 L. Oudejans and R. E. Miller, Chem. Phys. 239, 345 ͑1998͒.
44 G. Guelachvili, J. Mol. Spectrosc. 75, 251 ͑1979͒.
45 L. Oudejans, D. T. Moore, and R. E. Miller, J. Chem. Phys. 110, 209
͑1999͒; 110, 7109 ͑1999͒.
16 A. C. Legon, P. D. Soper, and W. H. Flygare, J. Chem. Phys. 74, 4944
͑1981͒.
17 W. G. Read and E. J. Campbell, J. Chem. Phys. 78, 6515 ͑1983͒.
18
¨
E. K. Kyro, P. Shoja-Chaghervand, K. McMillan, M. Eliades, D. A. Dan-
zeiser, and J. W. Bevan, J. Chem. Phys. 79, 78 ͑1983͒.
19 K. W. Jucks and R. E. Miller, J. Chem. Phys. 86, 6637 ͑1987͒.
20 G. T. Fraser and A. S. Pine, J. Chem. Phys. 88, 4147 ͑1988͒.
21 Z. Wang and J. W. Bevan, J. Chem. Phys. 91, 3335 ͑1989͒.
22 M. A. Benzel and C. E. Dykstra, J. Chem. Phys. 77, 1602 ͑1982͒.
23 D. E. Woon, T. H. Dunning, Jr., and K. A. Peterson, J. Chem. Phys. 104,
5883 ͑1996͒.
24 M. A. Benzel and C. E. Dykstra, J. Chem. Phys. 78, 4052 ͑1983͒; 80, 3510
͑1984͒.
25 M. A. Benzel and C. E. Dykstra, Chem. Phys. 80, 273 ͑1983͒.
26 S.-Y. Liu and C. E. Dykstra, J. Phys. Chem. 90, 3097 ͑1986͒.
27 C. A. Parish, J. D. Augspurger, and C. E. Dykstra, J. Phys. Chem. 96,
2069 ͑1992͒.
46 See EPAPS Document No. E-JCPSA6-113-002035 for the table with in-
dividual state-to-state probabilities. This document may be retrieved via
the EPAPS homepage ͑http://www.aip.org.pubservs/epaps.html͒ or from
ftp.aip.org in the directory /epaps/. See the EPAPS homepage for more
information.
28 L. A. Curtiss, D. J. Pochatko, A. E. Reed, and F. Weinhold, J. Chem.
Phys. 82, 2679 ͑1985͒.
29 A. E. Reed, F. Weinhold, L. A. Curtiss, and D. J. Pochatko, J. Chem.
Phys. 84, 5687 ͑1986͒.
30 M. A. Spackman, J. Chem. Phys. 85, 6587 ͑1986͒.
31 I. L. Alberts, N. C. Handy, and E. D. Simandiras, Theor. Chim. Acta 74,
415 ͑1988͒.
47 J. C. Light, J. Chem. Phys. 40, 3221 ͑1964͒.
48 J. C. Light, Faraday Discuss. Chem. Soc. 44, 14 ͑1967͒.
49 L. Oudejans, R. E. Miller, and W. L. Hase, Faraday Discuss. Chem. Soc.
102, 323 ͑1995͒.
32 C. Tuma, A. D. Boese, and N. C. Handy, Phys. Chem. Chem. Phys. 1,
3939 ͑1999͒.
50 L. Oudejans and R. E. Miller, J. Phys. Chem. 103, 4791 ͑1999͒.
51 R. S. Ram, Z. Morbi, B. Guo, K.-Q. Zhang, P. F. Bernath, J. Vander
Auwera, J. W. C. Johns, and S. P. Davis, Astrophys. J., Suppl. Ser. 103,
247 ͑1996͒.
33 S. A. C. McDowell and A. D. Buckingham, J. Chem. Soc., Faraday Trans.
89, 4253 ͑1993͒.
34 B. Civalleri, E. Garrone, and P. Ugliengo, J. Mol. Struct.: THEOCHEM
419, 227 ͑1997͒.
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:
131.230.73.202 On: Mon, 22 Dec 2014 15:21:46