Theory of polar semiconductor surfaces

Publisher Website
Google Scholar
Abstract
Effective atomic charges of atoms at the surface and in the bulk of a polar semiconductor are systematically calculated by modifying the nuclear charges of a homopolar semiconductor and estimating the electronic charge redistribution. Poisson’s equation is then integrated through the surface to determine possible charge accumulation or possible dipole layers, both of which are assumed energetically unfavorable. Finally, the simplest surface geometry is sought which contains neither. A planar (through reconstructed)surface is favorable on a (110) face, but we obtain a (111) As face with an overlayer, corresponding to one quarter of an atomic plane of Ga. A similar overlayer was suggested by Nosker, Mark, and Levine, but it is proposed here that the Ga atoms are in bridging positions saturating three quarters of the dangling As hybrids. A corresponding geometry with Ga and As interchanged is anticipated for the (111) Ga surface. These are consistent with the observed 1×1 on (110) and 2×2 on (111) Ga, but the observed 1×1 on (111) As must be attributed to randomization of the pattern since a true 1×1 gives charge localization. On the (100) surface the simplest favored geometry is a 1×4 pattern two atom layers deep and analogous to the Webb model of the 2×1 reconstruction on silicon (100) surfaces. The observed centered 2×8 is not consistent with a favored pattern less than four atom layers deep.

This publication has 0 references indexed in Scilit: