Group III metals Al, Ga, In —silicon

Work on these combinations is in large part attributable to Rowe et al. [287—292] with some theoretical input from Chalikowsky [293]. Again, the interface is not easy to define, but to quote Rowe et al. [291] results clearly show that the electronic states at the metal semiconductor interfaces are different from those of the clean silicon surface, or from those one would expect at an abrupt junction . 

The experimental results have been obtained almost entirely from LEED, UPS and LEELS measurements for coverages ranging from a fraction of a monolayer to 20 monolayers, but an important point has been the use of molecular beam techniques for metal deposition to minimze contamination effects. (The metal was effused from clean Knudsen sources.) If we consider first the LEED observations, the initial state is a thermally produced 111 7 x 7 structure. The 7x7 periodicity is retained up to 1 monolayer of metal deposit, but as an extrinsic 7x7 pattern induced by the metal rather than a simple decrease in intensity of the Si 7 x 7 reflections. This was quoted as evidence of an essentially two-dimensional morphology for the metal deposit, since the formation of three-dimensional nuclei with clean silicon between them would have only reduced the intensity of the intrinsic 7x7 pattern. Beyond one monlayer, growth could follow a Stranski-Krastanov mode, however. 

There are two important effects of metal deposition on the electronic structure of the silicon surface. The first is the saturation of the dangling bonds and removal of the associated back bond states by a very small metal coverage so that the Fermi level is not pinned by dangling bond 

In the proposed model, initial metal atoms are chemisorbed either as substitutional impurities in the silicon lattice, with strongly localized bonds, or to fill free surface vacancies. With further deposition, the bonds formed are still much more covalent than could occur for a pure metallic region, but nevertheless the interface region formed at this stage has considerable metallic character, as shown by the high electron density and the tailing of states into the gap. Its effect is to introduce a high density of new states in the gap and pin the Fermi level in a new position. 

This system has been studied by Goldstein [294] and Levine [295], and seems to be an example of very site-specific adsorption in which the Cs atoms occupy four-fold coordination sites above the uppermost Si atoms. There is some similarity with other systems in that the Si dangling bond states are removed by Cs deposition to be replaced by Cs-induced gap states. There is, however, no evidence for interface instability effects. 

---