Physics guided deep learning for generative design of crystal materials with symmetry constraints
Physics guided deep learning for generative design of crystal materials with symmetry constraints
复制标题
DOI：
10.1038/s41524-023-00987-9
复制
发表时间：
2022-03
影响因子：
9.7
通讯作者：
Yong Zhao;E. M. Siriwardane;Zhenyao Wu;Nihang Fu;Mohammed Al-Fahdi;Ming Hu;Jianjun Hu
中科院分区：
文献类型：
--
作者：
Yong Zhao;E. M. Siriwardane;Zhenyao Wu;Nihang Fu;Mohammed Al-Fahdi;Ming Hu;Jianjun Hu
发现新材料是材料科学中的一项具有挑战性的任务，对人类社会的进步至关重要。传统的基于实验和模拟的方法是劳动密集型的或昂贵的，成功很大程度上取决于专家的启发式知识。在这里，我们提出了一个基于深度学习的物理引导晶体生成模型（PGCGM），用于具有高度结构多样性和对称性的高效晶体材料设计。与最新的结构生成器之一FTCP相比，我们的模型将生成有效性提高了700%以上，与我们以前的CubicGAN模型相比，提高了45%以上。使用密度泛函理论（DFT）计算来验证所生成的结构，其中2000种材料中的1869种材料被成功优化并存入卡罗莱纳材料数据库www.carolinamatdb.org，其中39.6%具有负形成能，5.3%具有小于0.25 eV/原子的船体以上能量，表明它们的热力学稳定性和潜在的可合成性。
Discovering new materials is a challenging task in materials science crucial to the progress of human society. Conventional approaches based on experiments and simulations are labor-intensive or costly with success heavily depending on experts’ heuristic knowledge. Here, we propose a deep learning based Physics Guided Crystal Generative Model (PGCGM) for efficient crystal material design with high structural diversity and symmetry. Our model increases the generation validity by more than 700% compared to FTCP, one of the latest structure generators and by more than 45% compared to our previous CubicGAN model. Density Functional Theory (DFT) calculations are used to validate the generated structures with 1869 materials out of 2000 are successfully optimized and deposited into the Carolina Materials Database www.carolinamatdb.org, of which 39.6% have negative formation energy and 5.3% have energy-above-hull less than 0.25 eV/atom, indicating their thermodynamic stability and potential synthesizability.