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

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 cryst...

Full description

Saved in:
Bibliographic Details
Main Authors: Qingsi Lai, Fanjie Xu, Lin Yao, Zhifeng Gao, Siyuan Liu, Hongshuai Wang, Shuqi Lu, Di He, Liwei Wang, Linfeng Zhang, Cheng Wang, Guolin Ke
Format: Article
Language:English
Published: Wiley 2025-02-01
Series:Advanced Science
Subjects:
Online Access:https://doi.org/10.1002/advs.202410722
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849727292414099456
author Qingsi Lai
Fanjie Xu
Lin Yao
Zhifeng Gao
Siyuan Liu
Hongshuai Wang
Shuqi Lu
Di He
Liwei Wang
Linfeng Zhang
Cheng Wang
Guolin Ke
author_facet Qingsi Lai
Fanjie Xu
Lin Yao
Zhifeng Gao
Siyuan Liu
Hongshuai Wang
Shuqi Lu
Di He
Liwei Wang
Linfeng Zhang
Cheng Wang
Guolin Ke
author_sort Qingsi Lai
collection DOAJ
description 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.
format Article
id doaj-art-13e5cc9a143048d3bce69b9904d8870d
institution DOAJ
issn 2198-3844
language English
publishDate 2025-02-01
publisher Wiley
record_format Article
series Advanced Science
spelling doaj-art-13e5cc9a143048d3bce69b9904d8870d2025-08-20T03:09:54ZengWileyAdvanced Science2198-38442025-02-01128n/an/a10.1002/advs.202410722End‐to‐End Crystal Structure Prediction from Powder X‐Ray DiffractionQingsi Lai0Fanjie Xu1Lin Yao2Zhifeng Gao3Siyuan Liu4Hongshuai Wang5Shuqi Lu6Di He7Liwei Wang8Linfeng Zhang9Cheng Wang10Guolin Ke11DP Technology Beijing 100080 ChinaDP Technology Beijing 100080 ChinaDP Technology Beijing 100080 ChinaDP Technology Beijing 100080 ChinaDP Technology Beijing 100080 ChinaDP Technology Beijing 100080 ChinaDP Technology Beijing 100080 ChinaSchool of Intelligence Science and Technology Peking University Beijing 100871 ChinaCenter for Data Science Peking University Beijing 100871 ChinaDP Technology Beijing 100080 ChinaCollege of Chemistry and Chemical Engineering Xiamen University Xiamen 361005 ChinaDP Technology Beijing 100080 ChinaAbstract 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.https://doi.org/10.1002/advs.202410722crystal structure predictiondeep learningequivariant deep generative modelmetal–organic frameworks (MOFs)powder X‐ray diffraction
spellingShingle Qingsi Lai
Fanjie Xu
Lin Yao
Zhifeng Gao
Siyuan Liu
Hongshuai Wang
Shuqi Lu
Di He
Liwei Wang
Linfeng Zhang
Cheng Wang
Guolin Ke
End‐to‐End Crystal Structure Prediction from Powder X‐Ray Diffraction
Advanced Science
crystal structure prediction
deep learning
equivariant deep generative model
metal–organic frameworks (MOFs)
powder X‐ray diffraction
title End‐to‐End Crystal Structure Prediction from Powder X‐Ray Diffraction
title_full End‐to‐End Crystal Structure Prediction from Powder X‐Ray Diffraction
title_fullStr End‐to‐End Crystal Structure Prediction from Powder X‐Ray Diffraction
title_full_unstemmed End‐to‐End Crystal Structure Prediction from Powder X‐Ray Diffraction
title_short End‐to‐End Crystal Structure Prediction from Powder X‐Ray Diffraction
title_sort end to end crystal structure prediction from powder x ray diffraction
topic crystal structure prediction
deep learning
equivariant deep generative model
metal–organic frameworks (MOFs)
powder X‐ray diffraction
url https://doi.org/10.1002/advs.202410722
work_keys_str_mv AT qingsilai endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT fanjiexu endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT linyao endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT zhifenggao endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT siyuanliu endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT hongshuaiwang endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT shuqilu endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT dihe endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT liweiwang endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT linfengzhang endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT chengwang endtoendcrystalstructurepredictionfrompowderxraydiffraction
AT guolinke endtoendcrystalstructurepredictionfrompowderxraydiffraction