A genetic algorithm based inverse band structure method for semiconductor alloys

We present an efficient and accurate method for searching for atomic configurations with target band structure properties. Our approach to this inverse problem is to search the atomic configuration space by repeatedly applying a forward solver, guiding the search toward the optimal configuration using a genetic algorithm. For the forward solver, we relax the atomic positions, then solve the Schrِdinger equation using a fast empirical pseudopotential method. We employ a hierarchical parallelism for the combined forward solver and genetic algorithm. This enables the optimization process to run on many more processors than would otherwise be possible. We have optimized AlGaAs alloys for maximum bandgap and minimum bandgap for several given compositions and discuss the results. This approach can be generalized to a wide range of applications in material design.