Researchers model semiconductor defects across operating temperatures

Chalmers University of Technology researchers demonstrated a machine-learning-assisted framework for calculating how material defects change with temperature, with results matching experimental observations. Their study appeared in npj Computational Materials. The team modeled missing atoms across benchmark materials and tracked changes in the electronic energy levels those defects introduce. Several levels shifted substantially with temperature, changes that zero-Kelvin calculations would miss. The method follows defects across a temperature range, but the demonstration concerns simulations of missing-atom defects rather than manufactured device performance.