7967
J. Chem. Phys., Vol. 114, No. 18, 8 May 2001
Electron attachment to BrCN and CBrCl3
of the excess energy of the reaction must appear in the trans-
lation, pointing to the formation of a very-short-lived
CBrClϪ3 intermediate that dissociates before redistribution
*
of the excess energy of reaction to internal motions can oc-
cur, i.e., that has a lifetime Շ a vibrational period. In contrast
to BrCN, where the large vibrational spacing in the CNϪ
fragment requires that it be formed in its ground vibrational
state, the CBrCl2 fragment has a large number of vibrational
modes that have relatively small energy spacings, ϳ20–80
meV, and that are thus energetically accessible. As indicated
in Fig. 8, the radial distribution at large r can be reasonably
well fit by a Gaussian translational energy release distribu-
FIG. 9. The angular distribution of 35ClϪ ions formed in K(14p)/CBrCl3
collisions at Tϳ450 K. ꢀ, experimental data. The lines show model predic-
tions using the same mix of reaction channels as in Fig. 8͑b͒ and interme-
diate lifetimes for the longer-lived channel, represented by a rectangular
distribution, of 0 ͑—͒ and 10 ͑•••͒ ps.
¯
tion of 0.15 eV FWHM centered on a mean energy ⑀G
ϳ0.3 eV. ͑The width selected is comparable to the total in-
ternal vibrational energy in the target molecule at Tϳ450 K,
ϳ0.17 eV. The corresponding value at room temperature is
¯
ϳ0.08 eV.͒ If this value of ⑀G is taken as the excess energy
of reaction E, given by EA͑ClϪ͒ϪD0(CBrCl2ϪCl͒
ϩEintϩEK(eϪ), where EA͑ClϪ͒ is the electron affinity of Cl,
3.612 69͑6͒ eV,14 and the usable internal energy Eint is ϳ0.05
eV, this gives an upper limit on the bond dissociation energy
D0(CBrCl2ϪCl) of ϳ3.4 eV. This is somewhat smaller than
the value 3.6 eV calculated based on the local-spin-density
approximation of the density functional theory.1 However,
we have observed previously that density functional calcula-
tions tend to systematically overestimate dissociation ener-
gies.
that, of those capture events that lead to ClϪ production,
ϳ65% are associated with the formation of a very-short-
lived intermediate, i.e., direct dissociation, ϳ35% are asso-
ciated with formation of a longer-lived intermediate. The
same mix of reaction channels also provides a reasonable fit
to the radial distribution observed in K(14p)/CBrCl3 colli-
sions at Tϳ450 K.
The production of energetic ClϪ ions is greatly reduced
in room-temperature collisions. Indeed, analysis of the data
suggests that only ϳ20% of reactions are associated with
direct dissociation. Although more complex translational en-
ergy release distributions that provide better fits to the data
might be obtained by iteration, the simple two-component
distributions assumed here reveal the essential characteristics
of the attachment process.
As demonstrated in Fig. 9, which shows the angular dis-
tribution of ClϪ ions formed in K(14p)/CBrCl3 collisions at
Tϳ450 K, the same mix of reaction channels used in ana-
lyzing the radial distributions also provides a good fit to the
angular data. However, because direct dissociation domi-
nates, the predicted angular distributions are not very sensi-
tive to the assumed lifetime of the longer-lived intermediate
and this could not be accurately determined.
The present data suggest a reaction scenario in which
some of the ClϪ signal is associated with direct dissociation
following capture into an antibonding orbital at a Cl site.
͑Earlier calculations suggest that a sizable fraction of the
excess charge on the intermediate is located on the Cl atoms
and is antibonding with respect to the C–Cl bond.1,16͒ The
excess energy of reaction, however, is relatively small and
the presence of a potential barrier will reduce the probability
for direct dissociation. The effect of this barrier will be re-
duced as the internal energy in the target is increased, which
would account for the dramatic increase in direct dissocia-
tion observed at elevated target temperatures. If, following
capture at a Cl site, the ClϪ ion does not escape immediately,
the electron will be transiently bound, allowing energy redis-
tribution within the intermediate prior to dissociation. Fur-
thermore, dissociation might occur through breaking of the
CϪBr bond and formation of a BrϪ ion. The decrease in
overall BrϪ production observed at elevated target tempera-
The ClϪ signal at small r corresponds to the formation of
ions with relatively small kinetic energy, which requires the
creation of CBrClϪ3 intermediates whose lifetimes are suf-
*
ficient to allow the conversion of some of the excess energy
of the reaction into internal motions prior to dissociation. In
the limit that the lifetime is long enough to allow full statis-
tical redistribution of this excess energy among internal mo-
tions, unimolecular decay theory predicts a so-called two-
dimensional Boltzmann translation energy release
¯
¯
distribution of the form exp(Ϫ⑀/⑀B), where ⑀B is the mean
translational energy release.15 As illustrated in Fig. 8, model
calculations using a two-dimensional Boltzmann distribution
¯
with ⑀Bϳ0.06 eV display a pronounced peak in the radial
distribution at small r that is similar to that observed experi-
mentally. However, the presence of a signal at small r does
not require that the lifetime of the intermediate be sufficient
to allow full statistical redistribution of the excess energy
among internal motions. For example, a peak at small r is
also obtained using a rectangular translational energy release
distribution in which the probability for the release of trans-
lational energy ⑀ is maintained constant up to some cut-off
energy ⑀C . Although there is no real physical basis on which
to expect such a distribution, it does represent some reason-
able ‘‘intermediate’’ state in the evolution of a narrow high-
energy Gaussian distribution toward an exponentially decay-
ing two-dimensional Boltzmann distribution, and might
correspond to an intermediate lifetime of a few vibrational
periods. The radial distribution predicted assuming a rectan-
¯
gular distribution with mean energy ⑀Rϭ⑀C/2ϭ0.06 eV is
included in Fig. 8 and is similar to that for a Boltzmann
¯
distribution with ⑀Bϭ0.06 eV. Reasonable overall fits to the
(Tϳ450 K͒ experimental data can be obtained by assuming
141.212.109.170 On: Mon, 22 Dec 2014 13:59:39