PXRD-Bench: How Close Are We to Determining Crystal Structures from Powder Diffraction?
Abstract
Determining crystal structures from powder X-ray diffraction (PXRD) would enable routine characterization of newly synthesized materials, including those absent from reference databases. Generative models offer a promising approach, but existing evaluations remain fragmented and largely assume single-phase samples with known composition. To fill this gap, we introduce PXRD-Bench, a unified benchmark that extends evaluation to unknown per-element atom counts and recovery of both structures in two-phase mixtures. It provides approximately 4.1 million training structures with simulated PXRD, tests generalization to unseen compositions and structural prototypes, and evaluates recovery from experimental PXRD after excluding structurally matching training entries. To provide a strong reference baseline, we also introduce Bragglite, a PXRD-to-structure model developed through a systematic study of PXRD modeling choices. It achieves a 44.2\% Top-1 match rate on experimental RRUFF patterns with known composition, 24.6\% above the strongest prior baseline. Further analysis yields several key findings: some models fail to benefit from experimental PXRD, while generalization to unseen structural prototypes and recovery from two-phase mixtures remain challenging. Together, our benchmark, systematic studies, and Bragglite clarify the capabilities and limitations of current PXRD-to-structure models.
Related
- PXtal: Learning to Align Powder X-Ray Diffraction and Crystal Structures under Information Asymmetry across Modalities
- X2SBench: an open benchmark for evaluating crystal structure determination from powder diffraction
- Differentiable powder X-ray diffraction enables analysis of evolving structures
- Bridging powder and multi-crystal diffraction with basis-adaptive texture tomography
- Machine learning magnetic interactions from neutron powder diffraction data
BibTeX
@misc{woo2026pxrdbench,
title = {{PXRD-Bench: How Close Are We to Determining Crystal Structures from Powder Diffraction?}},
author = {Dongyeop Woo and Nofit Segal and Nayoung Kim and Minkyu Kim and Rafael Gomez-Bombarelli and Sungsoo Ahn},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.16480},
eprint = {2610.16480},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.16480},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - PXRD-Bench: How Close Are We to Determining Crystal Structures from Powder Diffraction?
AU - Dongyeop Woo
AU - Nofit Segal
AU - Nayoung Kim
AU - Minkyu Kim
AU - Rafael Gomez-Bombarelli
AU - Sungsoo Ahn
PY - 2026
DA - 2026/10/12
PB - arXiv
DO - 10.48550/arxiv.2610.16480
UR - https://arxiv.org/abs/2610.16480
AB - Determining crystal structures from powder X-ray diffraction (PXRD) would enable routine characterization of newly synthesized materials, including those absent from reference databases. Generative models offer a promising approach, but existing evaluations remain fragmented and largely assume single-phase samples with known composition. To fill this gap, we introduce PXRD-Bench, a unified benchmark that extends evaluation to unknown per-element atom counts and recovery of both structures in two-phase mixtures. It provides approximately 4.1 million training structures with simulated PXRD, tests generalization to unseen compositions and structural prototypes, and evaluates recovery from experimental PXRD after excluding structurally matching training entries. To provide a strong reference baseline, we also introduce Bragglite, a PXRD-to-structure model developed through a systematic study of PXRD modeling choices. It achieves a 44.2\% Top-1 match rate on experimental RRUFF patterns with known composition, 24.6\% above the strongest prior baseline. Further analysis yields several key findings: some models fail to benefit from experimental PXRD, while generalization to unseen structural prototypes and recovery from two-phase mixtures remain challenging. Together, our benchmark, systematic studies, and Bragglite clarify the capabilities and limitations of current PXRD-to-structure models.
ER -