The unified defect model has been successful in explaining a wide variety of phenomena as oxygen or a metal is added to the III–V surface. These phenomena cover a range from a small fraction of a monolayer of adatoms to practical III–V structures with very thick overlayers. The tenets of the unified defect model are outlined, and the experimental results leading to its formulation are briefly reviewed. InP levels 0.4 and 0.1 eV and GaAs levels 0.7 and 0.9 eV below the conduction-band minimum (CBM) are associated with either missing column III or V elements. In InP, it has been found possible by a number of workers to ’’switch’’ between the two defect levels by variations in surface processing, temperature, and/or selection of the deposited atom. The need to apply the proper concepts for surface and interface chemistry and metallurgy is recognized, and the danger of using solely bulk concepts is emphasized. The reason for this is examined for certain cases on an atomic level. The need for new fundamental attacks on interface interaction is shown. The importance of semiconductor–oxide chemical stability is also recognized and, drawing on a large body of work from several laboratories, it is suggested that there will be more difficulties with ’’native’’ oxides on GaAs than on InP. It is concluded that ’’scientific engineering’’ of interfaces to give optimum performance should be a goal and test of the fundamental work described here. Specific possibilities are discussed for Schottky barriers on III–V’s.
Read full abstract