End-to-End Crystal Structure Prediction from Powder X-Ray Diffraction

2024-01-08

Advancement of science

AI Summary

To address the longstanding challenges of manual intervention and coarse-grained resolution in PXRD data-driven crystal structure prediction, this paper introduces XtalNet—the first end-to-end, database-free, equivariant deep generative model for fine-grained crystal structure determination. Methodologically, XtalNet comprises two synergistic modules: (i) PXRD–crystal contrastive pretraining and (ii) conditional diffusion-based structural generation, integrating equivariant graph neural networks with explicit PXRD–structure space alignment to directly generate 3D atomic coordinates from experimental PXRD patterns. For the first time, XtalNet achieves fully automated, fine-grained structure solution on complex organic crystals containing up to 400 atoms per unit cell (e.g., hMOFs), attaining top-10 matching rates of 90.2% and 79% on the hMOF-100 and hMOF-400 benchmarks, respectively. This eliminates manual intervention entirely and overcomes critical bottlenecks in both accuracy and automation for PXRD-based structural elucidation.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Natural Language Processing: Generation

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Powder X-ray diffraction (PXRD) is a prevalent technique in materials characterization. While the analysis of PXRD often requires extensive human manual intervention, and most automated method only achieved at coarse-grained level. The more difficult and important task of fine-grained crystal structure prediction from PXRD remains unaddressed. This study introduces XtalNet, the first equivariant deep generative model for end-to-end crystal structure prediction from PXRD. Unlike previous crystal structure prediction methods that rely solely on composition, XtalNet leverages PXRD as an additional condition, eliminating ambiguity and enabling the generation of complex organic structures with up to 400 atoms in the unit cell. XtalNet comprises two modules: a Contrastive PXRD-Crystal Pretraining (CPCP) module that aligns PXRD space with crystal structure space, and a Conditional Crystal Structure Generation (CCSG) module that generates candidate crystal structures conditioned on PXRD patterns. Evaluation on two MOF datasets (hMOF-100 and hMOF-400) demonstrates XtalNet's effectiveness. XtalNet achieves a top-10 Match Rate of 90.2% and 79% for hMOF-100 and hMOF-400 in conditional crystal structure prediction task, respectively. XtalNet enables the direct prediction of crystal structures from experimental measurements, eliminating the need for manual intervention and external databases. This opens up new possibilities for automated crystal structure determination and the accelerated discovery of novel materials.

Problem

• Automates fine-grained crystal structure prediction
• Uses PXRD for complex organic structures
• Eliminates manual intervention in material discovery

Innovation

• Equivariant deep generative model
• PXRD as additional condition
• End-to-end crystal structure prediction