Defects design in 2D materials via high-throughput calculation and machine learning

Employing high throughput DFT calculations, we study the crystal structure, stability, and electronic structures of defects in 2D materials such as hexagonal boron nitride and transition metal dichalcogenides. The interaction of defects was evaluated by comparing the formation energies of defect complexes and individual defects. A mean-field theory model was constructed to understand the interaction dynamics of defects. Machine learning models were trained to predict the stability and electronic properties of defects.