Cargando…

Machine learning enables long time scale molecular photodynamics simulations

Photo-induced processes are fundamental in nature but accurate simulations of their dynamics are seriously limited by the cost of the underlying quantum chemical calculations, hampering their application for long time scales. Here we introduce a method based on machine learning to overcome this bott...

Descripción completa

Detalles Bibliográficos
Autores principales: Westermayr, Julia, Gastegger, Michael, Menger, Maximilian F. S. J., Mai, Sebastian, González, Leticia, Marquetand, Philipp
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Royal Society of Chemistry 2019
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6849489/
https://www.ncbi.nlm.nih.gov/pubmed/31857878
http://dx.doi.org/10.1039/c9sc01742a
_version_ 1783469216496615424
author Westermayr, Julia
Gastegger, Michael
Menger, Maximilian F. S. J.
Mai, Sebastian
González, Leticia
Marquetand, Philipp
author_facet Westermayr, Julia
Gastegger, Michael
Menger, Maximilian F. S. J.
Mai, Sebastian
González, Leticia
Marquetand, Philipp
author_sort Westermayr, Julia
collection PubMed
description Photo-induced processes are fundamental in nature but accurate simulations of their dynamics are seriously limited by the cost of the underlying quantum chemical calculations, hampering their application for long time scales. Here we introduce a method based on machine learning to overcome this bottleneck and enable accurate photodynamics on nanosecond time scales, which are otherwise out of reach with contemporary approaches. Instead of expensive quantum chemistry during molecular dynamics simulations, we use deep neural networks to learn the relationship between a molecular geometry and its high-dimensional electronic properties. As an example, the time evolution of the methylenimmonium cation for one nanosecond is used to demonstrate that machine learning algorithms can outperform standard excited-state molecular dynamics approaches in their computational efficiency while delivering the same accuracy.
format Online
Article
Text
id pubmed-6849489
institution National Center for Biotechnology Information
language English
publishDate 2019
publisher Royal Society of Chemistry
record_format MEDLINE/PubMed
spelling pubmed-68494892019-12-19 Machine learning enables long time scale molecular photodynamics simulations Westermayr, Julia Gastegger, Michael Menger, Maximilian F. S. J. Mai, Sebastian González, Leticia Marquetand, Philipp Chem Sci Chemistry Photo-induced processes are fundamental in nature but accurate simulations of their dynamics are seriously limited by the cost of the underlying quantum chemical calculations, hampering their application for long time scales. Here we introduce a method based on machine learning to overcome this bottleneck and enable accurate photodynamics on nanosecond time scales, which are otherwise out of reach with contemporary approaches. Instead of expensive quantum chemistry during molecular dynamics simulations, we use deep neural networks to learn the relationship between a molecular geometry and its high-dimensional electronic properties. As an example, the time evolution of the methylenimmonium cation for one nanosecond is used to demonstrate that machine learning algorithms can outperform standard excited-state molecular dynamics approaches in their computational efficiency while delivering the same accuracy. Royal Society of Chemistry 2019-08-05 /pmc/articles/PMC6849489/ /pubmed/31857878 http://dx.doi.org/10.1039/c9sc01742a Text en This journal is © The Royal Society of Chemistry 2019 http://creativecommons.org/licenses/by/3.0/ This article is freely available. This article is licensed under a Creative Commons Attribution 3.0 Unported Licence (CC BY 3.0)
spellingShingle Chemistry
Westermayr, Julia
Gastegger, Michael
Menger, Maximilian F. S. J.
Mai, Sebastian
González, Leticia
Marquetand, Philipp
Machine learning enables long time scale molecular photodynamics simulations
title Machine learning enables long time scale molecular photodynamics simulations
title_full Machine learning enables long time scale molecular photodynamics simulations
title_fullStr Machine learning enables long time scale molecular photodynamics simulations
title_full_unstemmed Machine learning enables long time scale molecular photodynamics simulations
title_short Machine learning enables long time scale molecular photodynamics simulations
title_sort machine learning enables long time scale molecular photodynamics simulations
topic Chemistry
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6849489/
https://www.ncbi.nlm.nih.gov/pubmed/31857878
http://dx.doi.org/10.1039/c9sc01742a
work_keys_str_mv AT westermayrjulia machinelearningenableslongtimescalemolecularphotodynamicssimulations
AT gasteggermichael machinelearningenableslongtimescalemolecularphotodynamicssimulations
AT mengermaximilianfsj machinelearningenableslongtimescalemolecularphotodynamicssimulations
AT maisebastian machinelearningenableslongtimescalemolecularphotodynamicssimulations
AT gonzalezleticia machinelearningenableslongtimescalemolecularphotodynamicssimulations
AT marquetandphilipp machinelearningenableslongtimescalemolecularphotodynamicssimulations