Rare-earth-metal–semiconductor interfacial reactions: Thermodynamic aspects

Abstract

Chemical reactions at rare-earth-metal–semiconductor interfaces are discussed based on heats of formation of bulk compounds and estimated heats of adsorption. Photoemission results for rare-earth-metal overlayers on the Si(111), Ge(111), and GaAs(110) surfaces, which have shown valence changes with chemical reactions and clustering of metal atoms, were used to estimate the heats of adsorption. The heats of adsorption of rare-earth metals on the Si surface were found to be considerably reduced as compared to bond strengths in bulk rare-earth silicides. This weak bond is attributed to an effectively zero-valent (6$s^2$) or monovalent (5$d^1$6$s^2$) valence-electron configuration of the rare-earth adatom on the Si surface. The heats of adsorption on the GaAs surface were found to be larger and somewhat closer to bond strengths in bulk rare-earth compounds than in the rare-earth–Si system, probably due to significant ionic contribution to the rare-earth—As bond. The fact that the surface disruption is observed in the initial stage of Ce adsorption on the GaAs surface but not for Sm on GaAs is explained by the larger heat of adsorption for Ce than for Sm (by ∼50 kcal/mol). The heats of formation of bulk rare-earth compounds were estimated from their valence behaviors, photoemission core-level shifts, and the heats of formation of related compounds. They are found to be only weakly dependent on the valence and the atomic number of the rare-earth metals, consistent with experimental photoemission results for multiple-layer coverages of these metals on semiconductor substrates.