Resonance Raman Scattering in Short Period GaAs-AlAs Superlattices

Abstract

New semiconductor growth techniques such as molecular beam epitaxy provide a powerful tool for band structure manipulation.¹ The first optical experiments in the early seventies demonstrated the possibility of fabricating structures whose band gaps differ from those of the bulk parent materials due to quantum confinement effects.² Moreover, it soon became apparent that even more dramatic effects could be expected, such as the appearance of new direct optical transitions produced by the folding of the “old” Brillouin zone (BZ) under the superlattice periodicity.³ This perspective is particularly exciting in the case of indirect gap materials, which might form direct gap superlattices. Recently, Pearsall et al. reported the observation of superlattice-induced optical transitions in Si-Ge structures.⁴ These transitions have small oscillator strengths, a carryover of their forbidden nature in the bulk which implies that they can only be observed in superlattices whose periods do not exceed a few atomic layers. This poses a serious material problem, because the superlattice period becomes of the order of the interface roughness. In the specific case of Si-Ge structures, the growth difficulty is compounded by the large lattice mismatch, the different optimal growth temperatures for Si and Ge, and the interchangeability of the two group-IV atoms in the crystal lattice. It is therefore not surprising that efforts in this area have concentrated on GaAs-AlAs structures, whose epitaxial growth is much simpler. Several groups have reported the observation, in very thin GaAs-AlAs superlattices, of electronic structure features qualitatively different from those found in thicker superlattices: non-monotonical thickness-dependence of the E₀-band gap,⁵ new direct optical transitions,⁶ Γ-x mixing effects,⁷ etc. In this paper, we study the problem of the superlattice-induced direct transitions and their oscillator strength using resonance Raman scattering.