Article (Scientific journals)
Accurate Many-Body Repulsive Potentials for Density-Functional Tight Binding from Deep Tensor Neural Networks
Stoehr, Martin; Medrano Sandonas, Leonardo; Tkatchenko, Alexandre
2020In Journal of Physical Chemistry Letters, 11 (16), p. 6835–6843
Peer Reviewed verified by ORBi
 

Files


Full Text
arXiv-manuscript.pdf
Author preprint (1.62 MB)
Download

All documents in ORBilu are protected by a user license.

Send to



Details



Abstract :
[en] We combine density-functional tight binding (DFTB) with deep tensor neural networks (DTNN) to maximize the strengths of both approaches in predicting structural, energetic, and vibrational molecular properties. The DTNN is used to construct a nonlinear model for the localized many-body interatomic repulsive energy, which so far has been treated in an atom-pairwise manner in DFTB. Substantially improving upon standard DFTB and DTNN, the resulting DFTB-NNrep model yields accurate predictions of atomization and isomerization energies, equilibrium geometries, vibrational frequencies, and dihedral rotation profiles for a large variety of organic molecules compared to the hybrid DFT-PBE0 functional. Our results highlight the potential of combining semiempirical electronic-structure methods with physically motivated machine learning approaches for predicting localized many-body interactions. We conclude by discussing future advancements of the DFTB-NNrep approach that could enable chemically accurate electronic-structure calculations for systems with tens of thousands of atoms.
Disciplines :
Physics
Author, co-author :
Stoehr, Martin ;  University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Physics and Materials Science (DPHYMS)
Medrano Sandonas, Leonardo ;  University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Physics and Materials Science (DPHYMS)
Tkatchenko, Alexandre ;  University of Luxembourg > Faculty of Science, Technology and Medicine (FSTM) > Department of Physics and Materials Science (DPHYMS)
External co-authors :
no
Language :
English
Title :
Accurate Many-Body Repulsive Potentials for Density-Functional Tight Binding from Deep Tensor Neural Networks
Publication date :
30 July 2020
Journal title :
Journal of Physical Chemistry Letters
ISSN :
1948-7185
Publisher :
American Chemical Society, Washington, United States - District of Columbia
Volume :
11
Issue :
16
Pages :
6835–6843
Peer reviewed :
Peer Reviewed verified by ORBi
Focus Area :
Physics and Materials Science
European Projects :
H2020 - 725291 - BeStMo - Beyond Static Molecules: Modeling Quantum Fluctuations in Complex Molecular Environments
FnR Project :
FNR11274975 - Coupling Nuclear Dynamics To Electronic Correlation In Molecular Materials, 2016 (01/10/2016-30/09/2020) - Martin Stöhr
Funders :
CE - Commission Européenne [BE]
Available on ORBilu :
since 10 December 2020

Statistics


Number of views
327 (80 by Unilu)
Number of downloads
96 (23 by Unilu)

Scopus citations®
 
56
Scopus citations®
without self-citations
48
OpenCitations
 
45
WoS citations
 
54

Bibliography


Similar publications



Contact ORBilu