Surface and interface states on GaAs(110): Effects of atomic and electronic rearrangements

Abstract

Research during the last year has led to a better understanding of the electronic and atomic structure of the (110) surfaces of III–V semiconductors. In this paper we will briefly review these new developments as well as point out areas where agreement has been found between various experimental results presented in the literature. It is now generally agreed that there are no intrinsic surface states in the band gap on GaAs and the smaller band-gap materials (e.g., GaSb, InAs, and GaSb) and that Schottky barrier pinning must be due to states produced when the metal adlayer is applied. Particular attention is focused in this paper on the large surface rearrangement which takes place on the (110) GaAs surface and effects of the strain which may be produced in joining this rearranged surface layer to the rest of GaAs crystal. It is pointed out that this may lead to variations in the surface rearrangement which can produce variations in the valence electronic structure at the surface. Such variations are shown in experimental energy distribution curves obtained by the photoemission technique which samples principally the last two molecular layers. It is further shown that surprisingly small amounts of chemisorbed oxygen can produce first-order effects in the valence-band electronic structure. On all GaAs (110) surfaces studied, a phaselike transformation was observed with a few hundredths of a monolayer coverage of chemisorbed oxygen. Near this coverage, the Ga 3d exciton structure disappears and the oxygen uptake increases significantly. On certain samples, first-order changes in the valence-band electronic structure were observed at a coverage of a hundredth of a monolayer or lower. These transformations are discussed in terms of the electronic and atomic configurations at the surface. Experimental data showing As and Ga 3d chemical shifts for oxidation as well as chemisorption are also presented and used to point out difficulties to be expected in passivating practical surfaces. In particular, the effect of mixed As and Ga oxides, the desirability of bonding passivating layers to the GaAs through As bonds, and the effect of strain-induced interface states are discussed.