Defect model of melt-grown GaAs

Publisher Website
Google Scholar
Abstract
A phenomenological model is developed to treat defect properties and defect chemistry in melt-grown GaAs. Defects are characterized by their charge and their stoichiometric signature (local deviation from stoichiometry); no specific atomic structure is assumed for them. Good fits to existing data are obtained for the room-temperature concentrations of the midgap donor (EL2), an unknown double acceptor, electrons, and holes as functions of melt composition. In obtaining these fits the role and importance of the (unknown) GaAs solidus is emphasized and it is demonstrated that many popular models of EL2 are consistent with the data analyzed. The model is extended to account for the observed decrease of EL2 concentration with increasing boron or silicon doping concentration, and here again attention is clearly focused on certain unknown information concerning the crystal formation process. Fundamental parameters in the model whose values are estimated through fits to the data include (1) the equilibrium constant in the defect reaction between EL2 and the double acceptor, (2) the equilibrium constant in the defect reaction between EL2, BAs and BGa, (3) the fractions of boron and silicon taken up on Ga sites during the crystal formation process, and (4) the proportionality factor between the melt composition and the crystal composition.