Metal-InP(110) interface properties: Temperature, dopant-concentration, and cluster-deposition dependencies.

Abstract

Synchrotron-radiation photoemission has been used to investigate the movement of the surface Fermi level, ${\mathit{E}}_{\mathit{F}}$, in the gap as a function of temperature, bulk dopant concentration, and the technique of interface formation for Ag/InP(110) and Ti/InP(110). Studies involving atom deposition at 300 and 60 K reveal temperature-independent substrate disruption with substrate retreat that is estimated to be \ensuremath{\sim}4.4 monolayers (ML) for Ti and \ensuremath{\sim}1 ML for Ag deposition. Atom distributions differ, however, because In atoms released by substrate disruption are kinetically trapped near the interface at 60 K but they are distributed in and atop the metal overlayer at 300 K. Moreover, Ag deposition at 60 K produces uniform overlayers because of restricted surface mobility whereas Ag clustering occurs for atom deposition at 300 K. Despite these significant differences in interface chemistry and morphology, both overlayers induce very different amounts of band bending at 60 K than at 300 K for n-type InP(110) doped at 4\ifmmode\times\else\texttimes\fi{}${10}^{17}$ ${\mathrm{cm}}^{\mathrm{\ensuremath{-}}3}$. This temperature-dependent difference in band bending is significantly smaller for n- and p-type InP(110) doped at 2.5\ifmmode\times\else\texttimes\fi{}${10}^{18}$ ${\mathrm{cm}}^{\mathrm{\ensuremath{-}}3}$. At higher coverage above the metallization threshold, the Schottky-barrier height is largely independent of substrate dopant concentration or measurement temperature but exhibits metal-specific values for Ag and Ti overlayers. The relationship between ${\mathit{E}}_{\mathit{F}}$ movement and the details of interfacial bonding and morphology has been further examined by depositing preformed Ag clusters (rather than atoms) on pristine InP(110) surfaces. For cluster deposition, there was no noticeable substrate disruption and the Fermi-level position was not dependent on the size or number of clusters. This indicates that changes in surface relaxation under and around the clusters introduce states that determine band bending.