Unified differentiable learning of electric response

Abstract Predicting response of materials to external stimuli is a primary objective of computational materials science. However, current methods are limited to small-scale simulations due to the unfavorable scaling of computational costs. Here, we implement an equivariant machine-learning framework...

Full description

Saved in:
Bibliographic Details
Main Authors: Stefano Falletta, Andrea Cepellotti, Anders Johansson, Chuin Wei Tan, Marc L. Descoteaux, Albert Musaelian, Cameron J. Owen, Boris Kozinsky
Format: Article
Language:English
Published: Nature Portfolio 2025-04-01
Series:Nature Communications
Online Access:https://doi.org/10.1038/s41467-025-59304-1
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850284552524660736
author Stefano Falletta
Andrea Cepellotti
Anders Johansson
Chuin Wei Tan
Marc L. Descoteaux
Albert Musaelian
Cameron J. Owen
Boris Kozinsky
author_facet Stefano Falletta
Andrea Cepellotti
Anders Johansson
Chuin Wei Tan
Marc L. Descoteaux
Albert Musaelian
Cameron J. Owen
Boris Kozinsky
author_sort Stefano Falletta
collection DOAJ
description Abstract Predicting response of materials to external stimuli is a primary objective of computational materials science. However, current methods are limited to small-scale simulations due to the unfavorable scaling of computational costs. Here, we implement an equivariant machine-learning framework where response properties stem from exact differential relationships between a generalized potential function and applied external fields. Focusing on responses to electric fields, the method predicts electric enthalpy, forces, polarization, Born charges, and polarizability within a unified model enforcing the full set of exact physical constraints, symmetries and conservation laws. Through application to α−SiO2, we demonstrate that our approach can be used for predicting vibrational and dielectric properties of materials, and for conducting large-scale dynamics under arbitrary electric fields at unprecedented accuracy and scale. We apply our method to ferroelectric BaTiO3 and capture the temperature dependence, frequency dependence, and time evolution of the ferroelectric hysteresis, revealing the underlying intrinsic mechanisms of nucleation and growth that govern ferroelectric domain switching.
format Article
id doaj-art-7f71b57d50d249f2ac7a48e44dae4458
institution OA Journals
issn 2041-1723
language English
publishDate 2025-04-01
publisher Nature Portfolio
record_format Article
series Nature Communications
spelling doaj-art-7f71b57d50d249f2ac7a48e44dae44582025-08-20T01:47:32ZengNature PortfolioNature Communications2041-17232025-04-0116111210.1038/s41467-025-59304-1Unified differentiable learning of electric responseStefano Falletta0Andrea Cepellotti1Anders Johansson2Chuin Wei Tan3Marc L. Descoteaux4Albert Musaelian5Cameron J. Owen6Boris Kozinsky7John A. Paulson School of Engineering and Applied Sciences, Harvard UniversityJohn A. Paulson School of Engineering and Applied Sciences, Harvard UniversityJohn A. Paulson School of Engineering and Applied Sciences, Harvard UniversityJohn A. Paulson School of Engineering and Applied Sciences, Harvard UniversityJohn A. Paulson School of Engineering and Applied Sciences, Harvard UniversityJohn A. Paulson School of Engineering and Applied Sciences, Harvard UniversityDepartment of Chemistry and Chemical Biology, Harvard UniversityJohn A. Paulson School of Engineering and Applied Sciences, Harvard UniversityAbstract Predicting response of materials to external stimuli is a primary objective of computational materials science. However, current methods are limited to small-scale simulations due to the unfavorable scaling of computational costs. Here, we implement an equivariant machine-learning framework where response properties stem from exact differential relationships between a generalized potential function and applied external fields. Focusing on responses to electric fields, the method predicts electric enthalpy, forces, polarization, Born charges, and polarizability within a unified model enforcing the full set of exact physical constraints, symmetries and conservation laws. Through application to α−SiO2, we demonstrate that our approach can be used for predicting vibrational and dielectric properties of materials, and for conducting large-scale dynamics under arbitrary electric fields at unprecedented accuracy and scale. We apply our method to ferroelectric BaTiO3 and capture the temperature dependence, frequency dependence, and time evolution of the ferroelectric hysteresis, revealing the underlying intrinsic mechanisms of nucleation and growth that govern ferroelectric domain switching.https://doi.org/10.1038/s41467-025-59304-1
spellingShingle Stefano Falletta
Andrea Cepellotti
Anders Johansson
Chuin Wei Tan
Marc L. Descoteaux
Albert Musaelian
Cameron J. Owen
Boris Kozinsky
Unified differentiable learning of electric response
Nature Communications
title Unified differentiable learning of electric response
title_full Unified differentiable learning of electric response
title_fullStr Unified differentiable learning of electric response
title_full_unstemmed Unified differentiable learning of electric response
title_short Unified differentiable learning of electric response
title_sort unified differentiable learning of electric response
url https://doi.org/10.1038/s41467-025-59304-1
work_keys_str_mv AT stefanofalletta unifieddifferentiablelearningofelectricresponse
AT andreacepellotti unifieddifferentiablelearningofelectricresponse
AT andersjohansson unifieddifferentiablelearningofelectricresponse
AT chuinweitan unifieddifferentiablelearningofelectricresponse
AT marcldescoteaux unifieddifferentiablelearningofelectricresponse
AT albertmusaelian unifieddifferentiablelearningofelectricresponse
AT cameronjowen unifieddifferentiablelearningofelectricresponse
AT boriskozinsky unifieddifferentiablelearningofelectricresponse