Atomistic prediction on the composition- and configuration-dependent bandgap of Ga(As,Sb) using cluster expansion and ab initio thermodynamics

Abstract

The composition- and configuration-dependent bandgaps of pseudobinary Ga(As,Sb) are examined by the cluster expansion method and statistical thermodynamics based on density functional theory. The bandgaps and energetic stability of 330,000 configurations in the entire composition range show a consistent inverse relationship, in that a configuration with lower energy has a higher bandgap for a given composition. This inverse relation can be deduced from the opposite signs of effective cluster interaction coefficients for bandgap and energy, and can be quantified by the correlations of properties with short-range order parameters. The bandgap of GaAs0.5Sb0.5 varies from 0.02 to 0.93 eV depending on the atomic configuration, which suggests another tremendous chance to tune the bandgap by the configuration control. The average bandgap of a certain composition, calculated by the ab initio thermodynamics, decreases with increasing temperature. The calculated average bandgap shows feasible agreement with the experimental bandgap, reproducing the bandgap bowing.