Deep learning tight-binding approach for large-scale electronic simulations at finite temperatures with $ab$ $initio$ accuracy
Published in Nature Communications 15, 6772 (2024), 2024
Recommended citation: Gu, Qiangqiang, Zhanghao Zhouyin, Shishir Kumar Pandey, Peng Zhang, Linfeng Zhang, and Weinan E. "Deep learning tight-binding approach for large-scale electronic simulations at finite temperatures with ab initio accuracy." Nature Communications 15, 6772 (2024). https://www.nature.com/articles/s41467-024-51006-4
Simulating electronic behavior in materials and devices with realistic large system sizes remains a formidable task within the $ab$ $initio$ framework. We propose DeePTB, an efficient deep learning-based tight-binding (TB) approach with $ab$ $initio$ accuracy to address this issue. By training with $ab$ $initio$ eigenvalues, our method can efficiently predict TB Hamiltonians for unseen structures. This capability facilitates efficient simulation of large-size systems under external perturbations like strain, which are vital for semiconductor band gap engineering. Moreover, DeePTB, combined with molecular dynamics, can be used to perform efficient and accurate finite temperature simulations of both atomic and electronic behavior simultaneously. This is demonstrated by computing the temperature-dependent properties of a GaP system with 106 atoms.
Recommended citation: Gu, Qiangqiang, et al. “Deep learning tight-binding approach for large-scale electronic simulations at finite temperatures with ab initio accuracy.” Nature Communications 15, 6772 (2024).