Photodissociation dynamics of the reaction CF2Br2 + hv → CF2 + 2Br. Energetics, threshold and nascent CF2 energy distributions for λ = 223-260 nm

Article abstract of DOI:10.1039/b000203h

The dissociation dynamics of the reaction CF₂Br₂ + hv → CF₂ + 2 Br have been studied for a variety of dissociation energies, E(diss) = 460-535 kJ mol^-1 (corresponding to λ = 260-223 nm). The laser induced fluorescence spectrum of nascent CF₂ products was measured for various dissociation energies within this range. Analysis of the spectra yielded the CF₂ vibrational distribution and average rotational energy. The translational energy of CF₂ was measured via the Doppler broadening of various fully resolved rovibronic transitions. The most detailed analysis of energy disposal in the CF₂ fragments was carried out at E(diss) = 486 kJ mol^-1 (or λ = 246 nm). At this energy each degree of freedom of CF₂ had an average energy of E(vib) = 0.4 ± 0.2 kJ mol^-1, E(rot) = 2.5 ± 0.5 kJ mol^-1, and E(trans) = 24 ± 3 kJ mol^-1. These CF₂ energies, coupled with the available thermochemical data, allow us to determine unambiguously that CF₂ production must be accompanied by the production of two atomic Br fragments. A photofragment excitation spectrum of CF₂Br₂, probing for the production of CF₂ fragments, provided a reaction threshold of 460 ± 3 kJ mol^-1 (corresponding to 260 ± 1.5 nm). The range of previously published reaction enthalpies varies from 392 to 438 kJ mol^-1, all of which are substantially below the observed threshold. Additionally, at E(diss) = 486 kJ mol^-1, the energy of the CF₂ fragment was 27 kJ mol^-1 on average, already in excess of the available 26 kJ mol^-1, and without considering the kinetic energy of the recoiling Br atoms. We rationalise these data by proposing that the reaction might have a small barrier in the exit channel. The observed threshold corresponds to the top of the barrier (460 kJ mol^{- 1}), while the final energy in the fragments is determined by the asymptotic reaction energy (~ 424 kJ mol^-1). Simple dynamical models are presented to show that the proposed mechanism is reasonable. Key future experiments and calculations are identified that would enable a clearer picture of the dynamics of this reaction.

Full text of DOI:10.1039/b000203h

View Article Online / Journal Homepage / Table of Contents for this issue
Photodissociation dynamics of the reaction CF Br + hm Ç CF +
2
2
2
2
Br. Energetics, threshold and nascent CF energy distributions for
2
k = 223–260 nm
Melanie R. Cameron,¤ Stephen A. Johns, Gregory F. Metha” and Scott H. Kable*
School of Chemistry, University of Sydney, NSW 2006, Australia.
E-mail: s.kable=chem.usyd.edu.au
Received 11th January 2000, Accepted 3rd April 2000
Published on the Web 12th May 2000
The dissociation dynamics of the reaction CF Br ] hl ] CF ] 2 Br have been studied for a variety of
2
2
2
dissociation energies, E \ 460È535 kJ mol~1 (corresponding to j \ 260È223 nm). The laser induced
diss
Ñuorescence spectrum of nascent CF products was measured for various dissociation energies within this
2
range. Analysis of the spectra yielded the CF vibrational distribution and average rotational energy.
2
The translational energy of CF was measured via the Doppler broadening of various fully resolved
2
rovibronic transitions. The most detailed analysis of energy disposal in the CF fragments was carried out at
2
E
E
\ 486 kJ mol~1 (or j \ 246 nm). At this energy each degree of freedom of CF had an average energy of
diss
vib
2
\ 0.4 ^ 0.2 kJ mol~1, E \ 2.5 ^ 0.5 kJ mol~1, and E
\ 24 ^ 3 kJ mol~1. These CF energies,
rot
trans
2
coupled with the available thermochemical data, allow us to determine unambiguously that CF production
2
must be accompanied by the production of two atomic Br fragments. A photofragment excitation spectrum of
CF Br , probing for the production of CF fragments, provided a reaction threshold of 460 ^ 3 kJ mol~1
2
2
2
(
corresponding to 260 ^ 1.5 nm). The range of previously published reaction enthalpies varies from 392 to 438
kJ mol~1, all of which are substantially below the observed threshold. Additionally, at E \ 486 kJ mol~1,
diss
the energy of the CF fragment was 27 kJ mol~1 on average, already in excess of the available 26 kJ mol~1,
2
and without considering the kinetic energy of the recoiling Br atoms. We rationalise these data by proposing
that the reaction might have a small barrier in the exit channel. The observed threshold corresponds to the top
of the barrier (460 kJ mol~1), while the Ðnal energy in the fragments is determined by the asymptotic reaction
energy (D424 kJ mol~1). Simple dynamical models are presented to show that the proposed mechanism is
reasonable. Key future experiments and calculations are identiÐed that would enable a clearer picture of the
dynamics of this reaction.
tion products have been the subject of some debate over this
time. Several pathways have been postulated to be dominant,
including:
I. Introduction
ScientiÐc interest in halomethanes has been sustained over
many decades. Originally, this interest arose due to their
many industrial and commercial applications as refrigerants,
blowing agents, Ñame retardants and aerosols. More recently,
this interest has continued from the e†ect these molecules
have produced in the atmosphere (particularly chloroÑuoro-
carbons and halons). The well-known ozone depleting cycle is
initiated by their photolysis in the stratosphere.1,2 A consider-
able body of work has therefore accumulated on the photo-
lytic reactions of halons, including the molecule investigated
here, CF Br . DibromodiÑuoromethane, or Halon 1202 has
hl
CF Br ÈÈ
Õ
CF ] Br * H \ 231 kJ mol~1
(1)
(2)
2
2
2
2
r
hl
CF Br ÈÈ
Õ
CF Br ] Br * H \ 274 kJ mol~1
2
2
2
r
hl
CF Br ÈÈ
Õ
CF ] 2Br * H \ 424 kJ mol~1 (3a)
2
2
2
r
2
2
hl
CF Br ÈÈ
been a†orded some attention in the popular media of late3
because it is not covered by the Montreal Protocol and its
concentration in the atmosphere has been observed to be
increasing rapidly over the past few years.4
Õ
CF Brs ] Br ÈÈ
Õ
CF ] 2Br
2
2
2
2
*
H \ 424 kJ mol~1 (3b)
hl
r
The ultraviolet photodissociation dynamics of CF Br have
been investigated using a wide variety of methods for almost
2
2
hl
CF Br ÈÈ
Õ
CF Br ] Br ÈÈ
Õ
CF ] 2Br
2
2
2
2
40 years. For this halon species, the primary photolysis reac-
*
H \ 424 kJ mol~1 (3c)
r
¤
Current address: Max-Planck-Institut fuer Chemie, Abteilung Luft-
The reaction enthalpies above are taken from a large com-
pendium of previously published data, and are discussed more
fully and referenced in Section IV. Reactions (1)È(3) produce
chemie, Postfach 3060, D-55128 Mainz, Germany.
Current address: Department of Chemistry, University of Adelaide,
”
Adelaide, South Australia, 5005.
DOI: 10.1039/b000203h
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547
This journal is ( The Owner Societies 2000
2539

View Article Online
di†erent chemical products, while reactions (3a), (3b) and (3c)
Dopplerimetry. Finally, the photofragment excitation spec-
trum of CF Br
production. These data lead to unambiguous determination of
the reaction products, the energetic threshold for the reaction,
and a discussion of possible reaction mechanisms.
each produce CF and two Br atoms, but via di†erent mecha-
was measured near the threshold for CF2
2
2
2
nisms. Reaction (3a) is the concerted breaking of both CÈBr
bonds, without the formation of an intermediate. Reaction
(
3b) is a sequential reaction, where the CF Br intermediate is
2
produced. Reaction (3c) likewise involves a CF Br interme-
diate, but the breaking of the second CÈBr bond requires a
2
II. Experimental
second photon.
The issue of the reaction products of CF Br photolysis and
DibromodiÑuoromethane (CF Br , 98%, Aldrich) was used
2
2
2
2
the mechanism for this reaction was Ðrst discussed in 1960 by
Mann and Thrush5 who utilised the, then, new technique of
Ñash photolysis to dissociate CF Br . Their observation of
throughout with no further puriÐcation. A slush bath of
ethanol and water maintained the sample at a temperature of
approximately [40 ¡C. Helium at a pressure of 2 atm was
passed over this sample, and the resultant 2È3% CF Br ÈHe
2
2
the CF product (which was supported the following year by
2
2
2
Simons and Yarwood6) led them to propose reaction (1) as the
mixture was expanded into a vacuum chamber through a
pulsed nozzle (Precision Instruments, PV-M3) with a 0.5 mm
diameter exit oriÐce. The free-jet chamber has been fully
described previously.14 The expansion was then crossed
approximately 7 mm downstream by either one or two lasers.
For one-laser experiments, the output from a frequency
doubled dye laser (Lambda Physik LPD 3001e) pumped by a
XeCl excimer laser (Lambda Physik Lextra 200) was used to
both dissociate the CF Br and probe the nascent CF frag-
appropriate pathway.
By 1972, a second reaction pathway had been postulated by
Walton.7 The dissociation of CF Br at 265 nm yielded the
2
2
detectable products (CF Br) and Br . From these results,
2
2
2
they concluded that the primary dissociation channel at this
wavelength was reaction (2).
In 1978, Sam and Yardley8 Ðrst used laser photolysis at 248
nm as the means of initiating the photodissociation of
CF Br . The reaction channel producing CF and Br [eqn.
2
2
2
ments. This was possible as a result of the coincidental nature
of the CF Br and CF absorption spectra (see Section III).
2
2
2
2
(
1)] was supported, although no Br was actually detected.
2
2
2
2
Later, in a similar experiment, Wampler et al.9 observed the
The laser was scanned continuously from 246 to 262 nm,
using C503 dye and a BBO doubling crystal.
Br molecule in addition to the CF fragments and Br atoms
2
2
seen by Sam and Yardley. These results led to the conclusion
that both reactions 1 and 2 were occurring.
Experiments were also conducted in which the pump and
probe lasers were independently tunable. The lasers entered
the chamber mutually perpendicular to the free jet expansion.
Perhaps the most deÐnitive experiments were carried out by
Krajnovich et al.10 in 1984. They also used an excimer laser at
48 nm with a time-of-Ñight mass spectrometer to detect frag-
Thus, the CF fragments that were probed could only have
2
2
come from parent molecules in the coldest part of the expan-
ments. They reported CF Br` and Br` ions and concluded
that the primary channel at this wavelength is reaction (2). In
sion. The probe laser was the excimer pumped dye laser
described above. The photolysis laser was either a Raman
shifted Nd : YAG laser (Continuum Surelight I-20) or a fre-
quency doubled (BBO) Nd : YAG pumped dye laser
(Continuum Surelight II-10, Lambda Physik Scanmate, C503
dye). The Raman shifted Nd : YAG provided Ðxed wave-
lengths within the absorption proÐle of CF Br by selecting
2
addition, low concentrations of CF` and CF ` were detected.
2
From the Ñight times of these species it was postulated that
they were produced as a result of the secondary photo-
dissociation of the CF Br radical [eqn. (3c)], rather than as a
2
result of the spontaneous dissociation of vibrationally hot
2
2
CF Br [eqn. (3b)].
various anti-Stokes Raman lines from either the 355 or 266
nm Nd : YAG harmonic, including 266, 246, 238, 223 and 218
nm. The doubled dye laser was used to tune continuously
from 246 to 264 nm (the tuning range of C503 dye). The line-
width of the Raman shifted laser was about 2 cm~1. Both
doubled dye lasers were approximately 0.3 cm~1 linewidth,
although this was narrowed by an etalon to D0.1 cm~1 when
higher resolution was necessary. The probe laser was timed to
arrive approximately 100 ns after the pump beam in order to
detect only nascent CF . Nascent CF spectra were measured
2
Ultrafast (picosecond) experiments were performed in 1991
by Gosnell et al.,11 again dissociating CF Br at 248 nm. The
2
2
major products were found to be CF Br and Br, but unlike
2
Krajnovich et al., they concluded that the CF Br sponta-
neously dissociated to produce CF and Br [eqn. (3b)]. Evi-
dence for the direct spectroscopic detection of CF Br was
presented, as well as an interpretation of the spectra based on
2
2
2
a slow intramolecular vibrational relaxation (IVR) process.
The quantum yield for reaction (3) was estimated to be 20%
2
2
[the remaining 80% being reaction (2)].
by Ðxing the frequency of the pump laser while scanning over
Felder et al.12 used photofragment translational spectros-
various CF transitions. Alternatively, a photofragment exci-
tation (phofex) spectrum could be measured by Ðxing the
2
copy to examine CF Br photodissociation dynamics at 193
2
2
nm. They concluded that two channels were manifest, the
probe laser frequency corresponding to a particular CF tran-
2
major one being reaction (2), with the CF Br fragment formed
sition and scanning the pump laser.
2
in unstable states. In addition, reaction (1) was observed
In all experiments performed, Ñuorescence from the nascent
(
although the Br may have been unstable), which was inter-
CF fragments was imaged using a quartz lens ( f \ 50 mm)
2
2
preted as the opening of the molecular channel which could
not be accessed in the 248 nm data of Krajnovich.
Multiphoton dissociation at 248 nm was used by van Hoey-
missen et al.13 in 1994 in a study that derived the absorption
cross section of CF Br. In addition, it was shown that
onto the entrance slits of a SPEX Minimate monochromator
used to Ðlter scattered laser light. The monochromator
bandpass was set at 20 nm. A photomultiplier (EMI 9789QB)
detected Ñuorescence signal through the monochromator,
which was subsequently processed by a boxcar integrator
(SRS-250), displayed on a Tektronix TDS-320 oscilloscope,
and recorded on a personal computer. The timing of the
excimer and YAG lasers, nozzle valve and electronics was
controlled by a digital delay generator (SRS DG-535).
During the course of each spectrum the power of the scan-
ning laser was monitored and recorded by directing a back
reÑection from one of the steering prisms onto a cuvette con-
taining Rhodamine 6G dye. Fluorescence from the dye was
detected by a fast silicon photodiode, processed by a PAR
162/165 boxcar integrator and sent to the same computer. The
2
(
30 ^ 10)% of the primary CF Br dissociated spontaneously
2
into CF ] Br. This result is in agreement with the work of
Gosnell et al. described above.
2
The present work is concerned with a more complete study
of the photodissociation dynamics of CF Br at a variety of
wavelengths between 223 and 260 nm by examination of the
nascent energy deposited into the CF fragment. The CF
vibrational state distribution and the rotational contour was
measured as a function of photolysis wavelength. The trans-
lational recoil energy of the CF fragment was estimated by
2
2
2
2
2
2540
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547

View Article Online
CF Ñuorescence signal was then corrected for any change in
the scanning laser power. The power of the non-scanning laser
in a two-laser experiment, was not measured continuously,
however, it was monitored before and after scans.
probe-alone signal. Details of how these problems were dealt
with and turned to advantage will be described as appropriate
in the following sections.
2
B. Estimating the threshold for CF production
2
III. Results
Establishment of the energy threshold for the production of
CF from CF Br is an essential ingredient in determining the
2
2
2
A. Absorption spectra of CF Br and CF
chemical pathway by which CF is produced, and to under-
2
2
2
2
stand the dynamics of its production. Three di†erent types of
The upper panel in Fig. 1 shows the ultraviolet absorption
spectrum of room temperature CF Br vapour taken in a
experiments were carried out in an attempt to measure this
threshold. In the Ðrst set of experiments, a single laser was
used to both dissociate CF Br and probe CF . Fig. 2 shows
2
2
Cary-4 UV/VIS spectrometer. The spectrum is featureless with
a maximum near 227 nm and absorption that continues to the
red beyond 300 nm. The di†useness and Gaussian band shape
are typical of a transition to a directly dissociative electronic
state.
The electronic spectrum of the CF molecule, unlike that of
its parent, is highly structured. The lowest panel in Fig. 1
shows a low resolution scan of the A
transition of jet-cooled CF , formed after pyrolysis of CF Br
2
2
2
a series of low resolution, nascent CF spectra, probed via
2
di†erent vibronic transitions. In all cases the transitions are
A
3
(0,t ,0) ^ X
3
(0,0,0), probing the vibrationless level of the CF2
2
product. The dissociation energy varies from approximately
38 200 cm~1 (lower panel) to 40 200 cm~1 (upper panel)
throughout these spectra. Since it is a one-laser experiment,
2
3
1B ^ X
3
1A electronic
1
1
the CF wavenumber on the abscissa is also the dissociation
2
2
2
2
wavenumber. The solid symbols in the upper panel of Fig. 1
in a heated nozzle apparatus as described previously.15 It
show the position of the dissociation wavelength in reference
to the CF Br absorption band. Within any one vibronic
clearly shows the strong progression in the l bending mode
2
originating from the (0,0,0) level in the ground electronic state,
2
2
band the di†erence in dissociation energy is only slight
along with hot band sequences arising from the (0,1,0) and
(
approximately 100 cm~1) and is unlikely to have a signiÐcant
(
0,2,0) levels. Elevated vibrational temperatures precipitated
e†ect on the qualitative results.
by the hot nozzle conditions caused the hot bands to be more
intense in this spectrum than would be detected in a normal
room temperature scan. The vibronic origin is located at 268.7
nm16 with the l progression extending to higher energy. The
maximum in the FranckÈCondon intensity distribution lies
Three observations are particularly noteworthy in the
spectra in Fig. 2. Firstly, when the laser is tuned near 38 200
cm~1 (261.5 nm) no CF signal is observed. The (0,2,0)
2
2
near 250 nm, corresponding to the (0,5,0) ^ (0,0,0) transition.
The absorption spectra of these two species, the parent and
fragment in the reaction, are therefore strongly overlapping.
This caused some experimental problems in distinguishing
two-laser pumpÈprobe signal from one-laser pump-alone or
Fig. 1 (a) Room temperature absorption spectrum of CF Br
2
2
vapour. The points indicated on the spectrum represent dissociation
wavelengths used in this work, either as one laser (solid circles) or two
laser (open circles) experiments. (b) Photofragment excitation spec-
trum of CF Br . The signal in the phofex spectrum has fallen to
2
2
essentially zero by 260 nm, whereas the absorption spectrum con-
Fig. 2 Nascent CF LIF spectra following dissociation of CF Br .
2
2
2
tinues to almost 300 nm; (c) jet-cooled CF LIF spectrum taken using
The same laser is used as both pump and probe. The spectra broaden
2
a pyrolysis nozzle (see text); the CF vibrational assignments in panel
with higher pump energy, and no CF signal was observed for pump
2
2
(
a) refer to the strong peaks in this spectrum.
energies below 38 600 cm~1.
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547
2541

View Article Online
spectra. The spectrum observed after pumping/probing the
(
0,3,0) ^ (0,0,0) vibrational band exhibits S/N B 13. The
absorption strength of CF Br drops by a factor of 1.4
between 258.4 and 261.7 nm (using the room temperature
2
2
spectrum as an estimate). The CF FranckÈCondon factor for
2
the A
3
(0,2,0) ^ X
3
(0,0,0) transition (at 261.7 nm) drops by 2.4 in
comparison with the A
3
(0,3,0) ^ X3 (0,0,0) transition.17 Therefore
a CF spectrum with S/N B 4 would be expected, (before
2
accounting for the increase in signal that would be created by
the smaller rotational partition function for CF at this lower
2
energy). Consequently, it seems that if CF Br is absorbing
2
2
light and dissociating to produce CF then the spectrum
2
should have been observed.
Additional evidence is provided by the spectra in Fig. 2,
which show that the rotational envelopes become narrower as
the dissociating/probing energy is decreased. The decreasing
width of the rotational contours indicate that the rotational
energy in the CF fragment is also decreasing, due to the
2
lower amount of energy imparted in the original photo-
dissociation event. The fact that this contraction is quite
marked from 255 to 260 nm suggests that the reaction thresh-
old for the formation of CF is near.
2
A series of two-laser experiments were performed to try and
isolate the e†ects of diminishing CF Br and CF absorption
2
2
2
strength with decreasing energy. In these experiments the
probe laser was kept Ðxed at the CF origin transition at
2
2
68.7 nm (well below the one laser detection threshold) while
the photolysis laser was scanned from 246 to 265 nm (the
range of C503 dye). The ensuing photofragment excitation
(phofex) spectrum, corrected for the changing laser power over
this range is shown in Fig. 1(b). The ““structureÏÏ in the phofex
spectrum is not reproducible but is an indicator of the under-
lying noise in the spectrum. A Gaussian Ðt has been overlaid
as a guide for comparison of this spectrum with the absorp-
tion spectrum above it. Near 260 nm, where the room tem-
perature absorption spectrum is still registering a signiÐcant
level of absorption, the signal in the phofex spectrum has
reached essentially zero.
Fig. 3 Low resolution, nascent CF LIF spectra obtained following
2
CF Br dissociation at (a) 223 and (b) 246 nm. The spectrum in panel
2
2
(
c) is the same as in Fig. 1(c), using the properties of the pyrolysis
nozzle and jet-cooing to enhance vibrational hot-bands. Population in
CF (t \ 1 and 2) is much more signiÐcant for the higher energy
2
2
photolysis energy, though not as much as in the pyrolysis experiment.
Finally, a two-laser experiment was performed, attempting
to dissociate CF Br at 266 nm, where more than two orders
2
2
^
(0,0,0) band should lie in this region,16 [see Fig. 1] centred
of magnitude additional laser power was available to circum-
vent the weaker absorption strength. The probe laser was set
at 38 211.7 cm~1. At slightly higher energy (38 700 cm~1 or
258.4 nm) the CF signal is obvious and corresponds to the
at 261.7 nm, i.e. for detection via the A3 (0,2,0) ^ X3 (0,0,0) tran-
2
(
0,3,0) ^ (0,0,0) transition. Our detection threshold for CF
sition. However, the large increase in photolysis power was
ine†ective as no CF
2
production from CF Br photolysis (which may or may not
be the energetic threshold) is therefore located between 38 200
2 signal could be observed at this wave-
2
2
length.
and 38 700 cm~1 (or 258.4 and 261.8 nm).
The conclusion from the evidence above is that for energies
below 38 450 ^ 250 cm~1 (j [ 260 ^ 1.5 nm) the excited
CF Br molecules dissociate to form CF Br and a Br atom.
The second observation from Fig. 2 is that the CF rota-
2
tional contour broadens as the dissociation energy is
increased. The width of the contour is an indication of the
2
2
2
At higher photolysis energies the molecules are sufficiently
amount of CF rotational excitation produced in the reaction,
which increases with increasing dissociation energy (see also
excited to form CF and two atomic Br fragments.
2
2
section III. C).
C. CF product state distributions
The third observation concerns the intensity of each spec-
trum, indicated by the scale on the ordinate. The intensity of
the CF2 LIF spectrum increases markedly from lower to
higher energy. The reasons are twofold: the absorption coeffi-
cient of the CF Br parent is increasing rapidly over this
range [see Fig 1(a)], and also the CF FranckÈCondon factors
increase monotonically from the lowest spectrum to a peak at
the (0,5,0) ^ (0,0,0) transition17 [see Fig. 1(c)].
The Ðrst and third observations above are linked because
2
The one-laser experiments sketch out an overview of the
CF Br dissociation mechanism that includes a threshold
2
2
near 38 450 cm~1 and increasing rotational excitation of the
CF fragment as the energy above this threshold is increased.
2
2
2
However, these experiments are not ideal for measuring the
2
details of the energy deposited in the CF fragment because
the available energy is changing throughout the spectrum, as
2
is the absorption strength of the parent. Therefore, the experi-
ments from which energy distributions are obtained are all
two-laser experiments with a Ðxed photolysis frequency. All
two-laser experiments were designed so that any “pump aloneÏ
or “probe aloneÏ signal had been minimised or eliminated.
This included temporal separation, the use of monochro-
mators and reduction in laser intensity. Vibrational distribu-
tions were obtained from a variety of photolysis wavelengths
in the range 223È259 nm. Rotational temperatures were esti-
the intensity of the CF Ñuorescence signal drops o† with
decreasing energy due to both the CF Br and CF absorp-
2
2
2
2
tion strengths. It is reasonable to question whether the lack of
signal near 261.5 nm is due to an energetic constraint or
simply to the combined reduction in CF Br and CF absorp-
tion coefficients and our subsequent ability to detect the
fragments. To examine this question information may
be extracted from the signal-to-noise (S/N) ratios in the
2
2
2
2542
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547

View Article Online
mated from spectral contour Ðtting over the range j
\
diss
46È259 nm. Sub-Doppler line shapes were measured for
2
several single rovibronic transitions for photolysis at 246 nm.
Vibrational distribution. Vibrational distributions were mea-
sured for a variety of wavelengths between 258 and 250 nm,
and then two further Ðxed wavelengths, 246 and 223 nm,
which are the 3rd and 4th anti-Stokes Raman shifted wave-
lengths from the 355 nm Nd : YAG harmonic. An overview of
the CF spectra produced in the 246 and 223 nm dissociation
2
of CF Br , showing the ““hotÏÏ vibrational structure, is shown
2
2
in Fig. 3. Part of the strong progression in the l bending
2
mode is shown, namely (0,t ,0) ^ (0,0,0) for t \ 2, and 3.
2
2
Reproduced below these spectra is the CF spectrum produc-
2
ed in a pyrolysis nozzle followed by free jet expansion. The
combination of the properties of the pyrolysis nozzle and dif-
fering rates of cooling in the free jet for the di†erent degrees of
freedom accentuates the vibrational hot bands. The most
prominent of these progressions originates from the (0,1,0)
vibrational state, and is labelled in the Ðgure. Comparison of
the spectra in Fig. 3(a) and 3(b) with that in Fig. 3(c) imme-
diately highlights the important feature of the former, viz.,
there is very little vibrational hot band activity in the nascent
spectrum.
Fig. 5 Medium resolution scan of the CF (0,5,0) ^ (0,0,0) transition.
2
(
a) Nascent CF following dissociation of CF Br at 246 nm; (b)
2
2
2
thermal (300 K) CF . The nascent spectrum has a narrower contour,
2
indicating less rotational energy than thermal CF .
2
The vibrational energy distributions were calculated from
these spectra using the integrated intensities of the transitions,
are identiÐable even for K \ 9. This indicates that the rota-
tional energy of the nascent fragments is less than that avail-
able at room temperature. These contours were simulated at
various temperatures using the ASYROT asymmetric rotor
program of Birss and Ramsay18 and rotational constants of
Mathews.16,19 The thermal spectrum was modelled well, but
the nascent spectrum less so. Although it is possible to model
(
0,2,0) ^ (0,0,0), (0,3,0) ^ (0,0,0), (0,3,0) ^ (0,1,0), (0,4,0) ^ (0,1,0)
and (0,4,0) ^ (0,2,0). The intensities were scaled by their
respective FranckÈCondon factors.17 The sum of the popu-
lation in the states was then normalised. These vibrational dis-
tributions are shown in Fig. 4. The circles represent the
normalised populations. The line Ðtted to the data represents
the decrease in rR band intensity, to do so would require a
K
the expected vibrational distribution in CF for temperatures
rotational temperature substantially colder than 100 K. Such
2
of 150, 325 and 650 K. The average vibrational energy of the
low temperatures, however, have very little surrounding rota-
tional structure, whereas the nascent spectrum is rich in rota-
tional structure. It is therefore impossible to Ðt exactly a single
temperature to the nascent spectrum. A simulation at 200 K
represented the best compromise. At this temperature the
band head intensity decreases fairly rapidly, yet the rotational
complexity is preserved. A temperature of 200 K corresponds
to a rotational energy (3/2kT ) of 210 cm~1 or 2.5 kJ mol~1.
CF fragments was calculated from this distribution by
2
directly summing the energy of each of the states scaled by its
normalised population. The average vibrational energies, Evib ,
calculated in this manner are shown in Table 1.
2
. Rotational energy. The rotational contour of (0,5,0)
^
(0,0,0) band of CF following 246 nm dissociation is shown
2
in Fig. 5. In comparison, the same spectrum, using the same
laser, was measured for thermalised (300 K) CF and is shown
beneath. The contour of the nascent spectrum is clearly nar-
rower than the 300 K spectrum. In the nascent spectrum the
3. Translational energy. Repeated high resolution etalon
scans were taken of nine non-overlapping rotational tran-
sitions for photolysis at 246 nm. The transitions chosen were,
in the rR branch; (JA \ 9, 10 and 11), in the rR branch;
2
rR band heads (labelled in the Ðgure) decrease sharply with
K
3
4
increasing K, so much so that by K \ 5 they are no longer
(JA \ 8, 10, 11, 16, 22) and rQ (14). Fig. 6 shows the Doppler
4
clearly distinguishable from the surrounding rotational struc-
ture. By comparison, the band heads in the thermal spectrum
proÐles of the rR (10) and rR (11) peaks, with a small unas-
4
4
signed transition between them. The line width of the laser,
estimated to have a full width at half maximum (fwhm) of 0.10
Fig. 6 High resolution scan of a small region of the nascent spec-
trum in Fig. 5. The laser line width is 0.10 cm~1 and is indicated by a
dotted line. Solid lines through the three peaks are Gaussian functions
with a fwhm of 0.20 cm~1.
Fig. 4 CF vibrational state population distributions for three pho-
2
tolysis wavelengths. Boltzmann distributions corresponding to tem-
peratures of 150, 325 and 650 K are indicated for comparison.
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547
2543

View Article Online
cm~1 is shown as a dotted line. It is apparent from this Ðgure
sient absorption spectrum of CF Br has been reported by
2
that the Doppler width of the recoiling CF fragment is only
several groups.11,13,22,23 The spectrum appears to evolve with
2
slightly broader than the laser line width indicating that the
time in the Ðrst few picoseconds of formation of CF Br. This
2
recoiling fragments do not possess a large velocity.
suggests that the fragment is born with a non-statistical dis-
To estimate the average translational energy, we Ðt each
peak to a Gaussian function (shown in the Ðgure) and
extracted the fwhm. The Ðtting routine allowed the peak posi-
tion, height and width to vary without constraint for the two
larger peaks. In both cases, a fwhm of 0.204 ^ 0.010 cm~1
resulted. For the Gaussian Ðt to the central peak the width
was held at this value and only the position and height
allowed to vary. To compensate for the broadening of the
transition due to the laser line width, the fwhm of the peak
was deconvolved with the fwhm of the laser. The deconvolved
peak width of the K \ 4 data is 0.173 ^ 0.010 cm~1. This cor-
responds to a translational temperature of 1950 ^ 200 K, or
tribution of vibrational energy, which then randomises over
this timescale.11
The formation of CF has been somewhat more controver-
2
sial. Observations have been plentiful, especially in infrared
multiphoton dissociation of CF Br ,24h27 the mechanism for
2
2
its formation not so clear. Early workers postulated Br for-
2
mation, which has been subsequently discounted. Krajnovich
et al. suggested that it is only formed when primary CF Br
2
fragments absorb a second photon. The weight of most recent
opinion is that CF may be produced as a primary product at
2
248 nm with quantum yields varying from 1È30%.11,13,21,23
The quantum yield of Br atom production has been mea-
an average translational energy of E
or 24.3 ^ 3.0 kJ mol~1.
\ 2030 ^ 200 cm~1
sured as 1.01 ^ 0.15 at 248 nm and 1.63 ^ 0.19 at 222 nm.
trans
These values are consistent with 1% of hot CF Br having suf-
2
At lower photolysis energies the signal strength dropped o†
signiÐcantly, and line widths approached the laser line width.
At much higher photolysis energies rotational congestion and
broader Doppler widths prevented measurement of clean,
single rovibronic transitions. Consequently, we have only been
Ðcient energy to spontaneously break the CÈBr bond at 248
nm, rising to 63% at 222 nm. The electronic state of the Br
fragment is not well-known. Talukdar et al. report the obser-
vation of both ground (2P ) and excited (2P ) spin-orbit
3
@2
1@2
states, however, the relative abundance of each was not
reported.21
able to estimate the nascent CF kinetic energy for the single
2
photolysis wavelength reported above, i.e. j \ 246 nm.
In the experiment described above we have measured the
nascent internal and translational energy of CF . The thresh-
2
old to CF production has also been measured. Below we
2
D. Summary of data
discuss how the observed energetics relate to the picture of
In the experiments above we have determined the threshold
CF Br chemistry that has been sketched out above.
2
2
for production of CF fragments from the photolysis of
2
CF Br to be at 260 ^ 1.5 nm, corresponding to a threshold
2
2
A. The appropriate reaction for CF formation
energy of 460 ^ 3 kJ mol~1. The nascent energy deposited
into the vibrational, rotational and translation degrees of
2
There are two di†erent reaction pathways that could lead to
the formation of CF from CF Br at the wavelengths used in
freedom of the CF fragment has been measured for a variety
2
2
2
2
of photolysis wavelengths, most completely for j \ 246 nm.
this work: (i) formation of CF and molecular bromine
diss
Most of the energy released to the CF fragment in this reac-
2
2
[reaction (1)]; (ii) formation of CF and two bromine atoms
[reactions (3a, b, c)]. Heats of formation and hence reaction
2
tion is dissipated into translational motion, with a smaller
amount into rotation and an almost negligible amount into
enthalpies are not particularly well-known for the species and
reactions involved. Table 2 summarises the range of values in
the recent literature, as well as the values that we have used
here. The heats of formation of CF , Br and Br are reason-
vibration. The nascent vibrational energy released into CF
following dissociation at 252.5 and 223 nm was also measured.
2
The CF2 vibrational energy was shown to increase with
increasing photolysis energy, although remaining fairly small
throughout. The average energy in each degree of freedom is
reported in Table 1 as a function of photolysis wavelength.
2
2
ably secure. However CF Br is known with only moderate
2
2
accuracy, and CF Br is poorly known. Even so, the values in
Table 2, coupled with the energy deposited in the CF frag-
ment enable an unambiguous determination of the appropri-
2
2
ate chemical pathway in this instance.
IV. Discussion
Firstly, we can conclude that CF is produced after absorp-
2
Many workers have examined the photodissociation of
CF Br at wavelengths similar to those employed here. Exci-
tion of only one photon. We measured the dependence of CF
2
signal on photolysis laser power and observed a linear depen-
2
2
tation into the Ðrst band system of CF Br corresponds to a
dence (not shown).20 However, as previous workers have dis-
cussed,10,11,13 this alone is not a deÐnitive result. If two
photons were absorbed then there would be an extra 486 kJ
mol~1 (the photon energy) to be disposed of in the fragment
2
2
B ^ A transition.10,20 Although early workers observed or
1
1
inferred Br as a product,5h8 later work appears to support
2
Br not being a primary product.10,21 The dominant channel
2
at 248 nm is the simple CÈBr bond cleavage yielding the
degrees of freedom. The total CF energy is only 27 kJ mol~1,
2
CF Br radical and atomic Br (either 2P or 2P
1@2). The tran-
leaving about 480 kJ mol~1 to be deposited in the trans-
2
3@2
Table 1 Summary of average energy in each degree of freedom of CF for three di†erent photolysis wavelengths
2
CF2
average energy/kJ mol~1 (% E
excess)
jdiss/nm
E
/kJ mol~1a
Vibration
Rotation
Translation
Total
excess
2
2
2
52.5
14 ^ 3
26 ^ 3
76 ^ 3
0
0%)
1.3 ^ 0.5
(9 ^ 4%)
2.5 ^ 0.5
(10 ^ 2%)
È
È
È
È
È
(
(
(
46
0.4 ^ 0.1
24.3 ^ 3.0b
(94 ^ 8%)
È
27.2 ^ 2.5
1.5%)
(106 ^ 10%)
23
3.7 ^ 0.6
5%)
È
a Relative to CF detection threshold at 460 ^ 3 kJ mol~1. b Determined speciÐcally for K \ 3 or 4, J \ 8È22. The average translational energy
2
a
for all CF states is likely to be slightly lower than this.
2
2544
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547

View Article Online
Table 2 Enthalpies of formation and reaction used in this work
H0
*
* H0
f
298
f
298
or * H0
or * H0
r
298
r
298
Species
range quoted/
preferred value/
or reaction
kJ mol~1
Refs.
kJ mol~1
Ref.
CF Br (g)
È362 to È382
È219 to È225
È172 to È194
112 to 118
31
392 to 438
213 to 225
248 to 276
12, 20
10, 12, 20
10, 12, 20, 28
10, 12, 29
29
12, 21, 23, 24
9, 12, 21, 22, 23
7, 10, 12, 13, 21,
È382 ^ 11
È220 ^ 8
È182 ^ 5
111.87 ^ 0.12
30.91 ^ 0.11
424a
20
20
28
29
29
2
2
CF Br (g)
2
CF (g)
2
Br (g)
Br (g)
2
CF Br ] CF ] 2 Br
2
2
2
CF Br ] CF ] Br
231a
274a
2
2
2
2
CF Br ] CF Br ] Br
2
2
2
2
2, 23, 24
CF Br ] CF ] Br
146 to 160
193
10, 12, 13
150a
193a
2
2
Br ] 2 Br
2
a Calculated from preferred heats of reaction. In the case of CF Br ] CF ] Br
2 this results in the preferred value lying slightly outside the
2
2
2
range of previously quoted values.
lational energy of the two recoiling bromine atoms. The trans-
nied by two Br atoms because the energy di†erences between
the chemical channels were large. However, the uncertainty in
the published heats of formation in Table 2 reveal an uncer-
tainty of up to 46 kJ mol~1 in the energy threshold. For dis-
lational energy of the CF is only 24 kJ mol~1 and therefore it
2
is difficult to construct a plausible mechanism where the two
heavier atoms receive vastly more translational energy than
the lighter CF species. Only a contrived mechanism in which
cussion of the mechanism of CF production a more accurate
2
2
the two bonds break in a concerted manner with the bromine
threshold is desirable.
atoms recoiling directly away from each other could produce
The most reliable values in Table 2 seem to indicate a value
such a result. Such a result also makes the CF appearance
of * H0 B 424 kJ mol~1 for breaking both CÈBr bonds in
2
r 298
threshold of 460 kJ mol~1 difficult to rationalise in any way.
Therefore, we reject this scenario as highly improbable.
The formation of molecular bromine can also be discounted
CF Br . We presented evidence above for a threshold of 460
2
2
kJ mol~1. The evidence included two di†erent types of thresh-
old measurements, and an unsuccessful attempt to measure
at these wavelengths. At j \ 246 nm, almost half of the 486
CF signal with a high intensity 266 nm source. Thermoche-
mical values for bromine-containing compounds are not very
diss
2
kJ mol~1 photon energy is required to overcome the reaction
endothermicity (231 kJ mol~1). The remainder (255 kJ mol~1)
is available to be distributed amongst the fragment degrees of
reliable, yet this discrepancy seems signiÐcant. An appearance
threshold, of course, need not be associated with a thermoche-
mical threshold. There are several scenarios that that would
give rise to an appearance threshold lying above the thermo-
chemical threshold: (i) If one or both of the Br atoms were
produced in an excited electronic state, then the appearance
freedom. Table 1 shows that the total energy of the CF has
2
been estimated to be 27 kJ mol~1, 24 of which is translational
energy. Since linear momentum must be conserved, the sibling
Br fragment must possess 7.5 kJ mol~1 of translational
2
energy. By di†erence, therefore, the average internal energy of
threshold for CF would be higher by this amount. There has
2
the Br2 molecule formed must be about 220 kJ mol~1. This
been one previous measurement of the Br spin-orbit state
amount is in excess of the energy required to dissociate Br
populations, which indicated that both ground state Br(2%
3@2
and excited state Br*(2% ) are produced, although the rela-
)
2
(
192 kJ mol~1), in which case bromine atoms and not mol-
1
@2
ecules would be formed. (Note that we are not discounting a
potential energy surface that leads to Br , but where the Br
is too energetic to be stable. At this stage we make no claim
about the potential energy surfaces involved in the reaction
mechanism, but simply that the ultimate chemical products
associated with CF
tive population of each was not reported.21 The higher spin-
orbit component lies 44.10 kJ mol~1 above the ground state
which, using the values in Table 2, would place the threshold
for production of CF ] Br(2% ) ] Br(2% ) at 468 kJ
2
2
2
3@2
1@2
mol~1. While this places the threshold in closer agreement
with the observed threshold, it does not solve the problem of
the excess energy in the fragments. At a dissociation energy of
486 kJ mol~1, there would only be B18 kJ mol~1 available
for CF , Br and Br*. Yet 27 kJ mol~1 is found in CF alone,
2 formation must be two atomic Br frag-
ments.)
The only remaining pathway is production of CF and two
2
bromine atoms. Using a single 246 nm photon, this leaves
2
2
about 62 kJ mol~1 to be divided between the three fragments
and another D23 kJ mol~1 is likely to be found in trans-
lational energy of the Br fragments (based on conservation of
momentum). Therefore, although a pathway in which one of
the Br fragments is electronically excited helps to rationalize
the appearance threshold, it fails to account for the energy
according to the values in Table 2. About 27 kJ mol~1 are
accounted for in the CF fragment leaving 35 kJ mol~1 in
2
translational energy of the two bromine atoms. This amount
of translational energy is easily justiÐed. A simple kinematic
calculation, starting with a tetrahedral CF Br structure,
disposal. Therefore, we believe that two ground state Br(2%
)
2
2
3@2
allowing the two (ground state) bromine atoms to be released
atoms are the most likely siblings to the CF fragment. We
2
simultaneously to create a CF fragment with 24 kJ mol~1 of
note that this is not in disagreement with the observation by
2
kinetic energy and conserving linear momentum, requires that
Talukdar et al., who observed both spin-orbit states at a
similar energy to that used here. Several groups10,11,13,21,23
have reported that, at wavelengths around 248 nm, the largest
fraction (probably D80%) of CF Br molecules decompose to
the energy of the bromine atoms be 11.5 kJ mol~1 each. This
is sufficiently close to the 35 kJ mol~1 required (fortuitously
close for the assumptions made), and provides conÐdence that
2 2
form CF Br and Br, where the CF Br is stable with respect to
the energy balance for the CF ] 2 Br channel is reasonable.
2
2
2
further spontaneous dissociation. Therefore any experiment
that probes for Br production at this wavelength must observe
Br from both the CF Br ] Br channel, and the CF ] 2Br
B. Energetic threshold for the production of CF
2
2
2
The heats of formation used above proved to be sufficiently
channel. At these energies, the CF Br ] Br channel has ample
excess energy (212 kJ mol~1 according to Table 2) to produce
2
accurate to determine that CF production must be accompa-
2
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547
2545

View Article Online
Br*. (ii) When the reaction energy lies very close to the thresh-
old the rate of reaction slows considerably (an e†ect modelled
well by transition state theory30). In these experiments, we
probe for products 100 ns after the photolysis event. This time
is sufficiently short to prevent collisions in the beam, however,
it is a long time scale for intramolecular vibrational redistri-
bution (IVR). Indeed, in CF Br the time scale for complete
2
IVR at these energies has been measured to be in the order of
6
ps.11 Also, the appearance threshold is D35 kJ mol~1 above
the thermochemical threshold. In molecules of this size, this is
too far above the threshold for kinetic constraints to be
observed on a nanosecond time scale. Therefore we reject a
kinetic constraint as being responsible for the observed
threshold in this case. (iii) A barrier in the exit channel would
also provide an appearance threshold that lies above the ther-
mochemical value. Such a barrier is shown schematically in
Fig. 7. In the Ðgure the asymptotic product energy is set at the
thermochemical value, which is consistent with the average
Fig. 8 Picture of the electronic states and electron orbital
energy found in the CF fragment. The top of the barrier is
occupancies of CF Br, CF and Br showing how a barrier might
2
2
2
placed at the appearance threshold, which provides an exit
eventuate from the interaction of two diabatic surfaces.
channel barrier in the order of 35 kJ mol~1. In the next
section we explore possible origins of such a barrier.
system leads to an exit channel barrier of D20 kJ mol~1 en
route to production of ground state CF and Br. Of course,
whether a barrier is actually formed depends on how strong
the interaction is. A very strong interaction may yield a
surface that remains below the asymptotic energy of the pro-
ducts and hence a ““barrierÏÏ is not formed.
Another source of a barrier in the present reaction might
arise from the two spin-orbit states of the Br atom, which are
separated by 44 kJ mol~1. A similar argument to that present-
ed above would apply for the two states of Br, rather than two
C. The barrier?
There has been no study, either experimental or theoretical,
on the reaction of CF Br ] CF ] Br, nor the reverse reac-
2
2
tion. There is also no experimental data, to our knowledge, on
the dissociation dynamics of any halogenated methyl radical
to compare with CF Br. Therefore, we consider the plausi-
2
bility of a barrier, and situations where one might arise only
in the broadest terms.
There are at least two ways by which a barrier might arise
CF states. This time the spin-orbit interaction would couple
2
in this system. The ground state of the CF Br radical is a
the two surfaces, probably at quite long range. A barrier
2
doublet with the unpaired electron occupying a r-like orbital.
might again result, depending on the range and strength of the
coupling.
The message from the general discussion above is that there
are reasonable mechanisms by which a barrier might arise in
If the CÈBr bond is stretched diabatically, then the resultant
products will be a Br atom and a CF molecule with two
2
unpaired electrons (a biradical) as shown in Fig. 8 by follow-
ing the dashed curve. The ground state of CF , however, is a
the dissociation of CF Br. Good, high level calculations on
2
2
carbene conÐguration in which the two non-bonding electrons
this reaction might assist in supporting or refuting the pres-
of the carbon occupy the same, in-plane, sp -like orbital.
ence of the hypothesized barrier. Such calculations will not be
straightforward as they involve open-shell species, heavy
atoms, and must include spin-orbit coupling.
More experimental evidence might also elucidate the pres-
ence of a barrier. In particular, the detailed rotational dis-
2
Therefore, the adiabatic surface must involve a crossing of the
two diabatic surfaces which, when they interact, will form two
adiabatic surfaces such as sketched in Fig. 8. Such an avoided
crossing commonly leads to the formation of a barrier in the
reaction potential energy surface. For example, a similar
system is the dissociation of CFBr in the excited electronic
state.31,32 The diabatic surface for breaking the CÈBr bond
leads to excited state CF. A similar curve crossing in this
tribution of the CF fragment could be quite sensitive to the
2
presence of a barrier, depending on the transition state struc-
ture, and the height of the barrier. If the reaction turns out not
to have a barrier, then the CF rotational distribution will
2
probably be characterised reasonably by phase space theory
or another statistical model. If a barrier exists, then the rota-
tional distribution might be non-statistical and reÑect the
nature of the barrier. Experimentally, this requires resolving
and measuring the hundreds of CF rovibronic transitions.
2
Fig. 5 indicates the complexity of the spectrum, but the
resolution shown in Fig. 6 demonstrates that it is feasible.
V. Conclusions and future work
In the present work we have explored the photolysis of
CF Br at wavelengths between 260 and 223 nm, concentrat-
2
2
ing mostly at 246 nm. The CF fragment was probed by LIF
2
and the nascent energy in vibration, rotation and translation
measured. CF is born vibrationally and rotationally cold (0.4
2
and 2.5 kJ mol~1 respectively at 246 nm), but with substan-
tially more translational energy (24 kJ mol~1). A threshold for
CF production was found at 260 ^ 1.5 nm, which corre-
Fig. 7 Schematic of the potential energy surface leading to CF Br
2
2
sponds to an energy threshold of 460 ^ 3 kJ mol~1. This
]
Br and CF ] 2 Br. Best estimates of the thermochemical enth-
2
alpies, based on Table 2 are shown. The hypothesized barrier in the
threshold seems inconsistent with the available, but insecure,
thermochemical data, which indicate that the energetic thresh-
CF Br ] CF ] Br channel is indicated.
2
2
2546
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547

View Article Online
8
9
C. L. Sam and J. T. Yardley, Chem. Phys. L ett., 1979, 61, 509.
old should be closer to 424 ^ 20 kJ mol~1. Analysis of the
F. B. Wampler, J. J. Tiee, W. W. Rice and R. C. Oldenborg, J.
Chem. Phys., 1979, 71, 3926.
energy distribution in the CF fragment leads to the unam-
2
biguous conclusion that the CF fragment must be formed in
2
10 D. Krajnovich, Z. Zhang, L. Butler and Y. T. Lee, J. Phys. Chem.,
concert with two Br atoms. The analysis also suggests that
1
984, 88, 4561.
both Br atoms are in their ground 2%
state. Comparison
with previous experiments, coupled with the analysis of the
energy balance suggests that the reaction is a two step process,
11 T. R. Gosnell, A. J. Taylor and J. L. Lyman, J. Chem. Phys., 1991,
3
@2
94, 5949.
1
2 P. Felder, X. Yang, G. Baum and J. R. Huber, Israel J. Chem.,
1
993, 43, 33.
forming CF Br ] Br in the Ðrst instance, with at least some of
2
13 J. van Hoeymissen, W. Uten and J. Peeters, Chem. Phys. L ett.,
the CF Br having sufficient energy to produce CF ] Br
2
2
1994, 226, 159.
spontaneously. We hypothesize that this second step proceeds
over a barrier, perhaps about 35 kJ mol~1 from the exit
channel, to account for the threshold discrepancy.
1
4
A. C. Terentis, M. Stone and S. H. Kable, J. Phys. Chem., 1994,
98, 10802.
15 M. R. Cameron, S. H. Kable and G. B. Bacskay, J. Chem. Phys.,
1
995, 103, 4476.
These conclusions indicate a need for further work. Firstly,
high quality ab initio calculations should reveal whether the
hypothesized barrier is reasonable, although with 2 Br atoms
in the system, open shell species and spin-orbit coupling, this
will not be a straightforward exercise. Secondly, more detailed
experiments, resolving the full J, K rotational distribution
might indicate whether a barrier is present. These experiments
are likewise non-trivial.
1
1
6
7
C. W. Mathews, Can. J. Phys., 1967, 45, 2355.
D. S. King, P. K. Schenck and J. C. Stephenson, J. Mol. Spectro-
sc., 1979, 78, 1.
1
8
F. W. Birss and D. A. Ramsey, Comput. Phys. Commun., 1984, 38,
83.
19 C. W. Mathews, National Research Council, Depository of Unpub-
lished Data, National Science Library, Ottawa, Canada K1A 0S2,
1
967.
2
2
0
1
M. R. Cameron, PhD Thesis, University of Sydney, 1998.
R. K. Talukdar, G. L. Vaghiani and A. R. Ravishankara, J.
Chem. Phys., 1992, 96, 8194.
R. K. Vatsa, A. Kumar, P. D. Naik, K. V. S. Rama Rao and J. P.
Mittal, Chem. Phys. L ett., 1993, 207, 75.
Acknowledgements
2
2
The authors thank the Joint Laser Facility (funded by the
Australian Research Council) for the loan of a laser used in
these experiments. We also thank Dr George Bacskay for con-
tinuous discussion about this system, in particular regarding
the presence of a barrier. MRC is grateful for an APA post-
graduate scholarship.
23 R. K. Vatsa, A. Kumar, P. D. Naik, K. V. S. Rama Rao and J. P.
Mittal, Bull. Chem. Soc. Jpn., 1995, 68, 2817.
2
2
2
4
5
6
B. Abel, H. Hippler, N. Lange, J. Schuppe and J. Troe, J. Chem.
Phys., 1994, 101, 9681.
R. J. S. Morrison, R. F. Loring, R. L. Farley and E. R. Grant, J.
Chem. Phys., 1981, 75, 148.
M. E. Jacox, Chem. Phys. L ett., 1978, 53, 192.
References
27 D. S. King and J. C. Stephenson, Chem. Phys. L ett., 1977, 51, 48.
8
2
M. W. Chase, Jr., NIST -JANAF T hemochemical T ables, Fourth
Edition, J. Phys. Chem. Ref. Data, Monograph 9, 1998, pp.
1È1951.
1
2
M. J. Molina and F. S. Rowland, Nature, 1974, 249, 810.
B. J. Finlayson-Pitts and J. N. Pitts, Jr., Atmospheric Chemistry:
Fundamentals and Experimental T echniques, Wiley, NY, 1986.
New Scientist, 12-September-1998, p. 12; Sydney Morning Herald,
29 J. D. Cox, D. D. Wagman and V. A. Medvedev, CODAT A Key
V alues for T hermodynamics, Hemisphere Publishing Corp., New
York, 1984, p. 1.
30 See, for example, R. G. Gilbert and S. C. Smith, T heory of
Unimolecular and Recombination Reactions, Blackwell, Oxford,
1990.
3
4
3
1-August-1998, p. 4.
P. J. Fraser, D. E. Oram, C. E. Reeves, S. A. Penkett and A.
McCulloch, J. Geophys. Res., 1999, 104, 15985.
5
6
D. E. Mann and B. A. Thrush, J. Chem. Phys., 1960, 33, 1732.
J. P. Simons and A. J. Yarwood, T rans. Faraday Soc., 1961, 57,
31 P. T. Knepp, C. K. Scalley, G. B. Bacskay and S. H. Kable, J.
2
167.
Chem. Phys., 1998, 109, 2220.
7
J. C. Walton, J. Chem. Soc., Faraday T rans., 1972, 68, 1559.
32 P. T. Knepp and S. H. Kable, J. Chem. Phys., 1999, 110, 11789.
Phys. Chem. Chem. Phys., 2000, 2, 2539È2547
2547