Computational approaches to point defect simulations for semiconductor solid solution alloys

Despite their technological importance, studying the properties of alloys with first principles methods remains challenging. In cases of Al_xGa_1-xN and Ba_xSr_x-1TiO₃ (BST), whose most important properties are governed by point defects, explicit simulation can be a computationally demanding task due to the random occupation of Al and Ga on cation sites in AlGaN and Ba and Sr on A-sites in BST. In this work, interpolation between end member compounds is used as a first approximation to defect properties and concentrations in intermediate alloy compositions in lieu of explicit simulation. In AlGaN, the efficacy of Si and Ge as dopants for n-type Al-rich AlGaN is explored by considering self-compensating defects such as multi-donor vacancy complexes and Si and Ge DX configurations. In BST, variation of the high temperature defect chemistry of Mg and Fe is examined. The approach presented here is expected to be generally appropriate for first approximation of defect properties in semiconductors and dielectrics where the alloy is a random solid solution of the end members.