Inverse Materials Design of Doping Strategies with AI, Thermodynamics, and Density Functional Theory

Inverse material design is, and has been, a long-standing goal of computational materials science. In this work, the authors present a novel approach for inverse materials design as it relates to point defect engineering: a particle swarm optimizer to aid in the design of doping profiles, particularly when working with multiple dopants and compound or wide band gap materials. This optimizer evaluates the full set of point defect and charge carrier concentrations of a material under high-temperature processing and subsequent quench to room temperature as part of its fitness function evaluation, using dopant concentrations as its parameter space. The approach is then applied to strontium titanate, an insulator widely used in capacitors in energy storage and energy filtering applications.