Unsupervised identification of crystal defects from atomistic potential descriptors

Abstract Identifying crystal defects is vital for unraveling the origins of many physical phenomena. Traditionally used order parameters are system-dependent and can be computationally expensive to calculate for long molecular dynamics simulations. Unsupervised algorithms offer an alternative indepe...

Full description

Saved in:
Bibliographic Details
Main Authors: Lukáš Kývala, Pablo Montero de Hijes, Christoph Dellago
Format: Article
Language:English
Published: Nature Portfolio 2025-02-01
Series:npj Computational Materials
Online Access:https://doi.org/10.1038/s41524-025-01544-2
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850045898797613056
author Lukáš Kývala
Pablo Montero de Hijes
Christoph Dellago
author_facet Lukáš Kývala
Pablo Montero de Hijes
Christoph Dellago
author_sort Lukáš Kývala
collection DOAJ
description Abstract Identifying crystal defects is vital for unraveling the origins of many physical phenomena. Traditionally used order parameters are system-dependent and can be computationally expensive to calculate for long molecular dynamics simulations. Unsupervised algorithms offer an alternative independent of the studied system and can utilize precalculated atomistic potential descriptors from molecular dynamics simulations. We compare the performance of three such algorithms (PCA, UMAP, and PaCMAP) on silicon and water systems. Initially, we evaluate the algorithms for recognizing phases, including crystal polymorphs and the melt, followed by an extension of our analysis to identify interstitials, vacancies, and interfaces. While PCA is found unsuitable for effective classification, it has been shown to be a suitable initialization for UMAP and PaCMAP. Both UMAP and PaCMAP show promising results overall, with PaCMAP proving more robust in classification, except in cases of significant class imbalance, where UMAP performs better. Notably, both algorithms successfully identify nuclei in supercooled water, demonstrating their applicability to ice nucleation in water.
format Article
id doaj-art-c0eb503e8dfe4430bf094f2227a09de7
institution DOAJ
issn 2057-3960
language English
publishDate 2025-02-01
publisher Nature Portfolio
record_format Article
series npj Computational Materials
spelling doaj-art-c0eb503e8dfe4430bf094f2227a09de72025-08-20T02:54:36ZengNature Portfolionpj Computational Materials2057-39602025-02-011111810.1038/s41524-025-01544-2Unsupervised identification of crystal defects from atomistic potential descriptorsLukáš Kývala0Pablo Montero de Hijes1Christoph Dellago2Faculty of Physics, University of ViennaFaculty of Physics, University of ViennaFaculty of Physics, University of ViennaAbstract Identifying crystal defects is vital for unraveling the origins of many physical phenomena. Traditionally used order parameters are system-dependent and can be computationally expensive to calculate for long molecular dynamics simulations. Unsupervised algorithms offer an alternative independent of the studied system and can utilize precalculated atomistic potential descriptors from molecular dynamics simulations. We compare the performance of three such algorithms (PCA, UMAP, and PaCMAP) on silicon and water systems. Initially, we evaluate the algorithms for recognizing phases, including crystal polymorphs and the melt, followed by an extension of our analysis to identify interstitials, vacancies, and interfaces. While PCA is found unsuitable for effective classification, it has been shown to be a suitable initialization for UMAP and PaCMAP. Both UMAP and PaCMAP show promising results overall, with PaCMAP proving more robust in classification, except in cases of significant class imbalance, where UMAP performs better. Notably, both algorithms successfully identify nuclei in supercooled water, demonstrating their applicability to ice nucleation in water.https://doi.org/10.1038/s41524-025-01544-2
spellingShingle Lukáš Kývala
Pablo Montero de Hijes
Christoph Dellago
Unsupervised identification of crystal defects from atomistic potential descriptors
npj Computational Materials
title Unsupervised identification of crystal defects from atomistic potential descriptors
title_full Unsupervised identification of crystal defects from atomistic potential descriptors
title_fullStr Unsupervised identification of crystal defects from atomistic potential descriptors
title_full_unstemmed Unsupervised identification of crystal defects from atomistic potential descriptors
title_short Unsupervised identification of crystal defects from atomistic potential descriptors
title_sort unsupervised identification of crystal defects from atomistic potential descriptors
url https://doi.org/10.1038/s41524-025-01544-2
work_keys_str_mv AT lukaskyvala unsupervisedidentificationofcrystaldefectsfromatomisticpotentialdescriptors
AT pablomonterodehijes unsupervisedidentificationofcrystaldefectsfromatomisticpotentialdescriptors
AT christophdellago unsupervisedidentificationofcrystaldefectsfromatomisticpotentialdescriptors