Optimized Tersoff potential parameters for tetrahedrally bonded III-V semiconductors

We address the issue of accurate parametrization for the Abell-Tersoff empirical potential applied to tetrahedrally bonded semiconductor materials. Empirical potential methods for structural relaxation are widely used for group IV semiconductors while, with few notable exceptions, work on III-V materials has not been extensive. In the case of the Abell-Tersoff potential parametrizations exist only for III-As and III-N, and are designed to correctly predict only a limited number of cohesive and elastic properties. In this work we show how by fitting to a larger set of cohesive and elastic properties calculated from density functional theory, we are able to obtain parameters for III-As, III-N, III-P, and III-Sb zinc blende semiconductors, which can also correctly predict important nonlinear effects in the strain.

Figure 1

Phase diagrams for III-As, III-N, III-P, and III-Sb semiconductors with the zinc blende (solid line), simple cubic (dash), and face centered cubic (dot). Correctly the zinc blende is the lowest energy structure.

Figure 2

Dependence of the sublattice internal displacement parameter $\zeta$ on shear strain for III-As, III-N, III-P, and III-Sb semiconductors. Ga (squares), In (circles), and Al (triangles) compounds are shown. For GaAs and InAs theoretical data (Ref. 59) (hollow symbols) is also shown for comparison.