Advancing nonadiabatic molecular dynamics simulations in solids with E(3) equivariant deep neural hamiltonians

Abstract Non-adiabatic molecular dynamics (NAMD) simulations have become an indispensable tool for investigating excited-state dynamics in solids. In this work, we propose a general framework, N2AMD (Neural-Network Non-Adiabatic Molecular Dynamics), which employs an E(3)-equivariant deep neural Hami...

Full description

Saved in:
Bibliographic Details
Main Authors: Changwei Zhang, Yang Zhong, Zhi-Guo Tao, Xinming Qin, Honghui Shang, Zhenggang Lan, Oleg V. Prezhdo, Xin-Gao Gong, Weibin Chu, Hongjun Xiang
Format: Article
Language:English
Published: Nature Portfolio 2025-02-01
Series:Nature Communications
Online Access:https://doi.org/10.1038/s41467-025-57328-1
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850029179903410176
author Changwei Zhang
Yang Zhong
Zhi-Guo Tao
Xinming Qin
Honghui Shang
Zhenggang Lan
Oleg V. Prezhdo
Xin-Gao Gong
Weibin Chu
Hongjun Xiang
author_facet Changwei Zhang
Yang Zhong
Zhi-Guo Tao
Xinming Qin
Honghui Shang
Zhenggang Lan
Oleg V. Prezhdo
Xin-Gao Gong
Weibin Chu
Hongjun Xiang
author_sort Changwei Zhang
collection DOAJ
description Abstract Non-adiabatic molecular dynamics (NAMD) simulations have become an indispensable tool for investigating excited-state dynamics in solids. In this work, we propose a general framework, N2AMD (Neural-Network Non-Adiabatic Molecular Dynamics), which employs an E(3)-equivariant deep neural Hamiltonian to boost the accuracy and efficiency of NAMD simulations. Distinct from conventional machine learning methods that predict key quantities in NAMD, N2AMD computes these quantities directly with a deep neural Hamiltonian, ensuring excellent accuracy, efficiency, and consistency. N2AMD not only achieves impressive efficiency in performing NAMD simulations at the hybrid functional level within the framework of the classical path approximation (CPA), but also demonstrates great potential in predicting non-adiabatic coupling vectors and suggests a method to go beyond CPA. Furthermore, N2AMD demonstrates excellent generalizability and enables seamless integration with advanced NAMD techniques and infrastructures. Taking several extensively investigated semiconductors as the prototypical system, we successfully simulate carrier recombination in both pristine and defective systems at large scales where conventional NAMD often significantly underestimates or even qualitatively incorrectly predicts lifetimes. This framework offers a reliable and efficient approach for conducting accurate NAMD simulations across various condensed materials.
format Article
id doaj-art-9a248ada2b6e4bf7bfd441f722c66b69
institution DOAJ
issn 2041-1723
language English
publishDate 2025-02-01
publisher Nature Portfolio
record_format Article
series Nature Communications
spelling doaj-art-9a248ada2b6e4bf7bfd441f722c66b692025-08-20T02:59:35ZengNature PortfolioNature Communications2041-17232025-02-0116111310.1038/s41467-025-57328-1Advancing nonadiabatic molecular dynamics simulations in solids with E(3) equivariant deep neural hamiltoniansChangwei Zhang0Yang Zhong1Zhi-Guo Tao2Xinming Qin3Honghui Shang4Zhenggang Lan5Oleg V. Prezhdo6Xin-Gao Gong7Weibin Chu8Hongjun Xiang9Key Laboratory of Computational Physical Sciences (Ministry of Education), Institute of Computational Physical Sciences, State Key Laboratory of Surface Physics, and Department of Physics, Fudan UniversityKey Laboratory of Computational Physical Sciences (Ministry of Education), Institute of Computational Physical Sciences, State Key Laboratory of Surface Physics, and Department of Physics, Fudan UniversityKey Laboratory of Computational Physical Sciences (Ministry of Education), Institute of Computational Physical Sciences, State Key Laboratory of Surface Physics, and Department of Physics, Fudan UniversityKey Laboratory of Precision and Intelligent Chemistry, Hefei National Research Center for Physical Sciences at the Microscale, University of Science and Technology of ChinaKey Laboratory of Precision and Intelligent Chemistry, Hefei National Research Center for Physical Sciences at the Microscale, University of Science and Technology of ChinaSCNU Environmental Research Institute, Guangdong Provincial Key Laboratory of Chemical Pollution and Environmental Safety & MOE Key Laboratory of Environmental Theoretical Chemistry, South China Normal UniversityDepartment of Chemistry and Department of Physics & Astronomy, University of Southern CaliforniaKey Laboratory of Computational Physical Sciences (Ministry of Education), Institute of Computational Physical Sciences, State Key Laboratory of Surface Physics, and Department of Physics, Fudan UniversityKey Laboratory of Computational Physical Sciences (Ministry of Education), Institute of Computational Physical Sciences, State Key Laboratory of Surface Physics, and Department of Physics, Fudan UniversityKey Laboratory of Computational Physical Sciences (Ministry of Education), Institute of Computational Physical Sciences, State Key Laboratory of Surface Physics, and Department of Physics, Fudan UniversityAbstract Non-adiabatic molecular dynamics (NAMD) simulations have become an indispensable tool for investigating excited-state dynamics in solids. In this work, we propose a general framework, N2AMD (Neural-Network Non-Adiabatic Molecular Dynamics), which employs an E(3)-equivariant deep neural Hamiltonian to boost the accuracy and efficiency of NAMD simulations. Distinct from conventional machine learning methods that predict key quantities in NAMD, N2AMD computes these quantities directly with a deep neural Hamiltonian, ensuring excellent accuracy, efficiency, and consistency. N2AMD not only achieves impressive efficiency in performing NAMD simulations at the hybrid functional level within the framework of the classical path approximation (CPA), but also demonstrates great potential in predicting non-adiabatic coupling vectors and suggests a method to go beyond CPA. Furthermore, N2AMD demonstrates excellent generalizability and enables seamless integration with advanced NAMD techniques and infrastructures. Taking several extensively investigated semiconductors as the prototypical system, we successfully simulate carrier recombination in both pristine and defective systems at large scales where conventional NAMD often significantly underestimates or even qualitatively incorrectly predicts lifetimes. This framework offers a reliable and efficient approach for conducting accurate NAMD simulations across various condensed materials.https://doi.org/10.1038/s41467-025-57328-1
spellingShingle Changwei Zhang
Yang Zhong
Zhi-Guo Tao
Xinming Qin
Honghui Shang
Zhenggang Lan
Oleg V. Prezhdo
Xin-Gao Gong
Weibin Chu
Hongjun Xiang
Advancing nonadiabatic molecular dynamics simulations in solids with E(3) equivariant deep neural hamiltonians
Nature Communications
title Advancing nonadiabatic molecular dynamics simulations in solids with E(3) equivariant deep neural hamiltonians
title_full Advancing nonadiabatic molecular dynamics simulations in solids with E(3) equivariant deep neural hamiltonians
title_fullStr Advancing nonadiabatic molecular dynamics simulations in solids with E(3) equivariant deep neural hamiltonians
title_full_unstemmed Advancing nonadiabatic molecular dynamics simulations in solids with E(3) equivariant deep neural hamiltonians
title_short Advancing nonadiabatic molecular dynamics simulations in solids with E(3) equivariant deep neural hamiltonians
title_sort advancing nonadiabatic molecular dynamics simulations in solids with e 3 equivariant deep neural hamiltonians
url https://doi.org/10.1038/s41467-025-57328-1
work_keys_str_mv AT changweizhang advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT yangzhong advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT zhiguotao advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT xinmingqin advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT honghuishang advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT zhengganglan advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT olegvprezhdo advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT xingaogong advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT weibinchu advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians
AT hongjunxiang advancingnonadiabaticmoleculardynamicssimulationsinsolidswithe3equivariantdeepneuralhamiltonians