Si and Be intralayers at GaAs/AlAs heterojunctions: Doping effects

Abstract

Experimental studies that suggest the possibility to ''engineer'' band offsets in semiconductor heterojunctions by means of intralayers have been controversially discussed. Here, Si and Be insertions at GaAs-on-AlAs(100) interfaces are investigated by photoelectron spectroscopy (PES) using synchrotron radiation. Our aim is to clarify the effect that band bending imposes on the determination of interface band offsets. The $\mathrm{Ga}(3d)\ensuremath{-}\mathrm{to}\ensuremath{-}\mathrm{Al}(2p)$ core-level energy separation is found to increase upon Si insertion, and to decrease upon Be insertion. The surface Fermi level moves closer to the valence-band maximum in Si-containing samples, while it moves away in Be-containing ones. These results are consistent with the n-type and p-type doping behaviors typically exhibited by Si and Be impurities in GaAs(100). The observed core-level offset variations support an interpretation based on band-bending arguments, rather than on the commonly invoked band-offset changes. A simple ''overlayer-capacitor'' model is proposed to illustrate the physical origin of such band-bending effects.