Physics Guided Deep Learning For Generative Design Of Crystal Materials With Symmetry Constraints

This article presents a deep learning based Physics Guided Crystal Generative Model (PGCGM) for efficient crystal material design. The model increases the generation validity by more than 700% compared to FTCP and by more than 45% compared to the CubicGAN model. Density Functional Theory (DFT) calculations are used to validate the generated structures, with 1869 materials out of 2000 successfully optimized and deposited into the Carolina Materials Database. 39.6% of these materials have negative formation energy and 5.3% have energy-above-hull less than 0.25 eV/atom, indicating their thermodynamic stability and potential synthesis.