XRDiff: Crystal Structure Prediction from Powder X-Ray Diffraction Data Using Diffusion Models

2026-06-11

AI Summary

This work addresses the highly underdetermined inverse problem of reconstructing crystal structures from powder X-ray diffraction (PXRD) data, which suffers from the missing phase information. The authors propose, for the first time, an end-to-end diffusion-based generative model that synthesizes crystal structures using only the chemical composition and the total number of atoms in the unit cell. A novel diffraction peak descriptor is introduced to encode PXRD patterns, and its performance is rigorously compared against full-spectrum encoding under an evaluation protocol designed to preserve polymorphic diversity. Experimental results demonstrate that the proposed method accurately distinguishes between polymorphs on simulated data and achieves superior zero-shot crystal structure solution on real experimental data, with the peak descriptor significantly outperforming full-spectrum encoding.

Abstract

Determining the crystal structure of a material from its powder X-ray diffraction (PXRD) pattern is a central challenge in materials science. PXRD is an accessible and widely used characterization technique, yet recovering the atomic structure from diffraction data requires solving an underdetermined inverse problem due to the loss of phase information. Generative modeling can provide a prior over atomic structure and learn the mapping from PXRD patterns to crystal structures via simulated structure-spectrum pairs. We present XRDiff, a diffusion model that recovers crystal structures from PXRD given either the stoichiometry or, in a more challenging setting, the elemental constituents and total number of atoms in the unit cell. We evaluate on datasets where each stoichiometry has multiple polymorphs and all polymorphs of a given composition are held out together, ensuring that high performance reflects genuine use of the diffraction signal. XRDiff achieves strong structure recovery rates on simulated benchmarks, indicating that the model learns a spectrum-to-structure mapping precise enough to differentiate between polymorphs. To address generalization to experimental data, we compare a full-spectrum encoding against an encoding based on peak descriptors. The peak-based encoding generalizes substantially better, outperforming even a model trained on full spectra with augmentations fitted to the experimental noise distribution. These results demonstrate that representations robust to the noise and artifacts present in real-world PXRD offer a practical and scalable path toward closing the simulation-to-experiment gap, enabling zero-shot crystal structure solution from experimental PXRD with full or partial chemical composition input.

Problem

• crystal structure prediction
• powder X-ray diffraction
• phase problem
• inverse problem
• polymorph differentiation

Innovation

• diffusion model
• crystal structure prediction
• powder X-ray diffraction
• polymorph differentiation
• simulation-to-experiment generalization