Powder diffraction crystal structure determination using generative models
Q Li, Rui Jiao, Liming Wu, Tiannian Zhu, Wenbing Huang, Shifeng Jin, Junyang Liu, Hongming Weng, Xiaolong Chen
Abstract
Accurate crystal structure determination is critical across all scientific disciplines involving crystalline materials. However, solving and refining crystal structures from powder X-ray diffraction (PXRD) data is traditionally a labor-intensive process that demands substantial expertise. Here we introduce PXRDGen, an end-to-end neural network that determines crystal structures by learning joint structural distributions from experimentally stable crystals and their PXRD, producing atomically accurate structures refined through PXRD data. PXRDGen integrates a pretrained XRD encoder, a diffusion/flow-based structure generator, and a Rietveld refinement module, solving structures with unparalleled accuracy in seconds. Evaluation on MP-20 dataset reveals a record high matching rate of 82% (1-sample) and 96% (20-samples) for valid compounds, with Root Mean Square Error (RMSE) approaching the precision limits of Rietveld refinement. PXRDGen effectively tackles key challenges in PXRD, such as the resolution of overlapping peaks, localization of light atoms, and differentiation of neighboring elements. Crystal structure determination from powder X-ray diffraction is challenging but vital for materials research. Here, authors develop PXRDGen, an AI system that automatically solves crystal structures with 96% accuracy across thousands of compounds.