Abstract Scope |
AlGaN alloys are critically important for high-electron-mobility transistor (HEMT) devices, yet their defect behavior remains insufficiently studied. Understanding defect interactions and their properties, particularly under radiation environments, is essential for predicting how they would affect device performance. To address this, we investigate defect formation and migration energetics in AlGaN using a machine learning-based interatomic potential. This model accurately reproduces fundamental properties of GaN and AlN and predicts defect formation energies for Frenkel pairs and Schottky defects, consistent with first-principles calculations. In the alloy systems, our analysis highlights the significant impact of local atomic environments, which is driven by combined strain and chemical effects, on defect energetics. While cation defect formation energetics show minimal variation across different Al compositions, nitrogen defects are highly sensitive to local atomic surroundings, especially the number of neighboring Al atoms. These findings provide insights towards understanding defect behavior in AlGaN-based electronic devices. |