GaAs(110)–In: The black sheep in a well-behaved interface family

Abstract

Photoemission results on the interface formation process for In on cleaved GaAs(110) are completely at odds with respect to those on other III–V/metal interfaces. We do not observe the transition between an ultralow coverage (<0.1 monolayer) ‘‘chemisorption’’ stage and an intermediate coverage (0.1–2 monolayer) ‘‘clustering’’ stage, observed for Al overlayers. Also for Al overlayers, the transition corresponds to the onset of a cation exchange reaction and to the pinning of the Fermi level. The exchange reaction is not observed for GaAs(110)–In, and the Fermi-level pinning occurs at much larger coverages. Furthermore, the pinning position for n-type GaAs is not the canonical one observed for a wide variety of metal, semiconductor, and oxygen overlayers, but 0.2–0.25 eV above it. All of these ‘‘odd’’ results demonstrate that the kinetics of the interface formation is a fundamental factor in the defect model, and clarify its role.

Absence of Fermi level pinning at metal-InxGa1−xAs(100) interfaces

“…Figure 3 illustrates these trends for Au, AI, and In across the lnx Ga l _ x As (l00) alloy series. Here, the valence-band energies are referred to a common vacuum level as determined by photoemission threshold measure- 21 for cleaved GaAs (110) (e.g., 0.4 vs 0.9 eV 2 ).…”