First-principles simulations of atomic and electronic structure in solids offer a powerful route to predict and understand material properties. This is particularly relevant for point defects, which can dramatically affect material properties yet remain challenging to characterize experimentally.

Recent years have seen significant advances in both computational methodologies and associated toolkits for modeling defect behavior. In this webinar, we’ll introduce the typical defect modeling workflow using Density Functional Theory (DFT) and explore collaborative efforts that are making defect calculations more robust, reproducible, and efficient.

The webinar will cover the ShakeNBreak approach for navigating complex defect configurational landscapes and identifying ground-state defect structures, as well as doped, a Python package designed for robust and reproducible defect supercell calculations. We’ll also explore the modelling of non-radiative charge capture processes at defects in semiconductors.

Time permitting, we’ll finish with a look at recent efforts to accelerate these workflows using machine-learning approaches—highlighting their exciting potential while also considering current limitations in accuracy and reliability.