Photodissociation of chlorine adsorbed on a LiF (001) surface at 25-70 K
has been investigated by means of angularly-resolved resonantly-enhanced
multiphoton ionization spectroscopy (REMPI).
The translational-energy distributions
and angular distributions for forming Cl(g) photofragments were determined.
Photolysis was performed employing 351 nm radiation, with laser pulse energies
of 0.3-1.2 mJ/cm2. A peak in the translational energy of Cl(g)
at about 0.4 eV was identified as being due to the direct photodissociation of
the Cl2(ad) molecule by 3.5 eV photons. Particular interest attached
to the observation of a further channel (termed "A") for photodissociation
leading to Cl(g) with translational energy peaking at ~1 eV and extending
to 1.5 eV. The available photon energy renders it highly unlikely that
this "high-energy" Cl(g) originates in Cl2(ad). Channel A had
the same linear dependence of Cl-atom flux on laser pulse-energy as did
the lower energy (0.4 eV) channel, termed "B", but differed from it in
exhibiting a slow approach to steady state. It appears that channel A requires
the prior build-up of Cl(ad) concentration due to the photodissociation
of Cl2. It is proposed that this leads to the formation of a
steady-state concentration of Cl2···Cl
which when photolysed yields high-energy Cl(g) via channel A. Channel A
exhibits a distinctive angular distribution at low coverage and a characteristic
Cl*/Cl ratio, as compared with channel B. The suggested mechanism
for channel A is
Cl2···Cl + h
Cl3*
Cl2 # Cl
Cl2 + Cl'
(where * is an electronically excited state and # represents repulsion
in the lower electronic state to which Cl3* reverts)
and Cl' is the fast chlorine atom (channel A). This mechanism
is discussed in terms of an extensive DIM model for the trichlorine radical,
shown to be in agreement with high-level ab initio MRCI calculations
and to be consistent with the observations.