In recent experiments carried out by
Watanabe and Matsumoto[1] xenon atoms adsorbed on
modified and unmodified silicon surfaces were irradiated with photons of variable energy
(1.16-6.43 eV). The
resulting time-of-flight spectra of the desorbed atoms show two velocity
components: a slow one centered around 0.25 eV and a fast one around 0.85
eV. The slow component appears over the complete excitation energy range and
has been assigned to the energy
transfer between local phonon excitations and the adsorbed atoms.[2] The fast component appears only for modified (oxidized) surfaces when the
energy is >3.5 eV. Its origin, however, is less well understood. Watanabe
and Matsumoto have proposed a mechanism in which an electron is transferred
from a Xenon atom to the surface, placing the system in the first (charge-transfer)
state.
The system subsequently evolves in this state and the positively charged
atom is attracted to the negative surface until the system returns to the
ground state. If the evolution time is sufficiently long and since the excited
state potential
is expected to be much more attractive than the ground state, the system
returns to the ground state in a highly repulsive region, providing the atom
with sufficient energy to desorbe. This mechanism is summarized in the following
diagram:
The
limit energy for the charge-transfer state corresponds to the difference
between the ionization energy of the Xe atom (12.12 eV) and the electron affinity
of the surface. Since the oxidation state of the surface is not completely
know, this energy can not determined and the authors equate it to the work
function of silicon (4.6 eV). With this assumption, the charge-transfer state
is lowered sufficiently bringing it within the range of the experimental
excitation energies.
Our group has been recently studying the formation and localization of Self-Trapped Excitons (STE) in quartz, and the energy transfer between them and adsorbed atoms. The energy required to form a STE in SiO2 is about 3.0 eV, in good agreement with the energy threshold observed by Watanabe and Matsumoto. Moreover, the exciton is observed to evolve towards the surface where a dangling bond pushes the adsorbed atoms away. In light of these results we have decided to carry out calculations to assess the feasibility of the charge-transfer mechanism in the case of xenon atoms adsorbed on quartz.
We have used the CASSCF and CASPT2 methods to model the ground and excited (charge-transfer) state interaction between a xenon atom and the quartz surface. This surface was represented by the cluster model shown in Fig. 1. The active space chosen was the minimal required to accurately describe the states involved and, to ensure the accuracy of the CASSCF results, two basis set schemes were used. Both basis sets were augmented with diffuse functions to accurately represent the anionic state of the surface.
Our group has been recently studying the formation and localization of Self-Trapped Excitons (STE) in quartz, and the energy transfer between them and adsorbed atoms. The energy required to form a STE in SiO2 is about 3.0 eV, in good agreement with the energy threshold observed by Watanabe and Matsumoto. Moreover, the exciton is observed to evolve towards the surface where a dangling bond pushes the adsorbed atoms away. In light of these results we have decided to carry out calculations to assess the feasibility of the charge-transfer mechanism in the case of xenon atoms adsorbed on quartz.
We have used the CASSCF and CASPT2 methods to model the ground and excited (charge-transfer) state interaction between a xenon atom and the quartz surface. This surface was represented by the cluster model shown in Fig. 1. The active space chosen was the minimal required to accurately describe the states involved and, to ensure the accuracy of the CASSCF results, two basis set schemes were used. Both basis sets were augmented with diffuse functions to accurately represent the anionic state of the surface.
Figure
1: Cluster model used to represent the quartz surface. The unsaturated oxygen
atoms are capped with hydrogen atoms (not shown).
Results
Fig.
2 presents the CASSCF and CASPT2 potential energy curves (PEC's) for the
interaction between a xenon atom and the quartz
surface. As expected,
the minima in the PEC's are shifted to shorter interaction distances when
electron correlation is included. This results from the better description
of the dispersion energy proivided by the CASPT2 method. The energy required
to access the excited state PEC is 8.51 eV at CASPT2 level (8.97 eV at CASSCF
level), far greater than the experimentally observed threshold.
Figure 2: CASSCF and CASPT2 potential energy curves for the ground and charge-transfer state of Xe...SiO2. The arrows indicate the presence of a local minimum.
This
is a consequence of two different factors: First, the electron affinity of
the surface is negligible, providing no stabilization upon charge-transfer.
Second, the very diffuse nature of the transferred electron reduces the electrostatic
interaction energy, raising the position of the minimum in the excited state.
These results clearly show that the mechanism proposed by Watanabe and Matsumoto
can not be applied to a quartz surface.
References
References
- K. Watanabe and Y. Matsumoto, Faraday Discuss. 117 (2000) 203.
- K. Watanabe, H. Kato and Y. Matsumoto, Surface Science 446 (2000) L134.