skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Guidelines for creating artificial neural network empirical interatomic potential from first-principles molecular dynamics data under specific conditions and its application to $α$-Ag2Se

Abstract

First-principles molecular dynamics (FPMD) simulations are highly accurate, but due to their high calculation cost, the computational scale is often limited to hundreds of atoms and few picoseconds under specific temperature and pressure conditions. As such, we present here the guidelines for creating artificial neural network empirical interatomic potential (ANN potential) trained with such a limited FPMD data, which can perform long time scale MD simulations at least under the same conditions. The FPMD data for training are prepared on the basis of the convergence of radial distribution function [$g(r)$]. While training the ANN using total energy and atomic forces of the FPMD data, the error of pressure is also monitored and minimized. To create further robust potential, we add a small amount of FPMD data to reproduce the interaction between two atoms that are close to each other. ANN potentials for $α$-Ag2Se were created as an application example, and it has been confirmed that not only $g(r)$ and mean square displacements but also the specific heat requiring a long time scale simulation matched the FPMD and the experimental values. In addition, the MD simulation using the ANN potential achieved over 104 acceleration over the FPMD one. The guidelines proposedmore » here mitigate the creation difficulty of the ANN potential, and a lot of FPMD data sleeping on the hard disk after the research may be put on the front stage again.« less

Authors:
ORCiD logo [1];  [2];  [2];  [2];  [3];  [4]; ORCiD logo [4];  [4];  [1];  [1]
  1. Kobe Univ. (Japan)
  2. Kumamoto Univ. (Japan)
  3. Kyushu Sangyo Univ., Fukuoka (Japan)
  4. Univ. of Southern California, Los Angeles, CA (United States)
Publication Date:
Research Org.:
Univ. of Southern California, Los Angeles, CA (United States)
Sponsoring Org.:
Ministry of Education, Culture, Sports, Science and Technology (MEXT); Japan Society for the Promotion of Science (JSPS); Japan Science and Technology Agency (JST); USDOE Office of Science (SC), Basic Energy Sciences (BES) (SC-22). Materials Sciences & Engineering Division
OSTI Identifier:
1612871
Grant/Contract Number:  
SC0018195; 16K05478; 17H06353; 18K03825; 19K14676; JPMJCR18I2
Resource Type:
Accepted Manuscript
Journal Name:
Journal of Chemical Physics
Additional Journal Information:
Journal Volume: 151; Journal Issue: 12; Journal ID: ISSN 0021-9606
Publisher:
American Institute of Physics (AIP)
Country of Publication:
United States
Language:
English
Subject:
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSICS; 37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY; Chemistry; Physics; Artificial neural networks; Potential energy surfaces; Symmetry functions; Molecular dynamics; Interatomic potentials; First-principle calculations; Thermodynamic properties

Citation Formats

Shimamura, Kohei, Fukushima, Shogo, Koura, Akihide, Shimojo, Fuyuki, Misawa, Masaaki, Kalia, Rajiv K., Nakano, Aiichiro, Vashishta, Priya, Matsubara, Takashi, and Tanaka, Shigenori. Guidelines for creating artificial neural network empirical interatomic potential from first-principles molecular dynamics data under specific conditions and its application to $α$-Ag2Se. United States: N. p., 2019. Web. https://doi.org/10.1063/1.5116420.
Shimamura, Kohei, Fukushima, Shogo, Koura, Akihide, Shimojo, Fuyuki, Misawa, Masaaki, Kalia, Rajiv K., Nakano, Aiichiro, Vashishta, Priya, Matsubara, Takashi, & Tanaka, Shigenori. Guidelines for creating artificial neural network empirical interatomic potential from first-principles molecular dynamics data under specific conditions and its application to $α$-Ag2Se. United States. https://doi.org/10.1063/1.5116420
Shimamura, Kohei, Fukushima, Shogo, Koura, Akihide, Shimojo, Fuyuki, Misawa, Masaaki, Kalia, Rajiv K., Nakano, Aiichiro, Vashishta, Priya, Matsubara, Takashi, and Tanaka, Shigenori. Wed . "Guidelines for creating artificial neural network empirical interatomic potential from first-principles molecular dynamics data under specific conditions and its application to $α$-Ag2Se". United States. https://doi.org/10.1063/1.5116420. https://www.osti.gov/servlets/purl/1612871.
@article{osti_1612871,
title = {Guidelines for creating artificial neural network empirical interatomic potential from first-principles molecular dynamics data under specific conditions and its application to $α$-Ag2Se},
author = {Shimamura, Kohei and Fukushima, Shogo and Koura, Akihide and Shimojo, Fuyuki and Misawa, Masaaki and Kalia, Rajiv K. and Nakano, Aiichiro and Vashishta, Priya and Matsubara, Takashi and Tanaka, Shigenori},
abstractNote = {First-principles molecular dynamics (FPMD) simulations are highly accurate, but due to their high calculation cost, the computational scale is often limited to hundreds of atoms and few picoseconds under specific temperature and pressure conditions. As such, we present here the guidelines for creating artificial neural network empirical interatomic potential (ANN potential) trained with such a limited FPMD data, which can perform long time scale MD simulations at least under the same conditions. The FPMD data for training are prepared on the basis of the convergence of radial distribution function [$g(r)$]. While training the ANN using total energy and atomic forces of the FPMD data, the error of pressure is also monitored and minimized. To create further robust potential, we add a small amount of FPMD data to reproduce the interaction between two atoms that are close to each other. ANN potentials for $α$-Ag2Se were created as an application example, and it has been confirmed that not only $g(r)$ and mean square displacements but also the specific heat requiring a long time scale simulation matched the FPMD and the experimental values. In addition, the MD simulation using the ANN potential achieved over 104 acceleration over the FPMD one. The guidelines proposed here mitigate the creation difficulty of the ANN potential, and a lot of FPMD data sleeping on the hard disk after the research may be put on the front stage again.},
doi = {10.1063/1.5116420},
journal = {Journal of Chemical Physics},
number = 12,
volume = 151,
place = {United States},
year = {2019},
month = {9}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Citation Metrics:
Cited by: 1 work
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

Deviation from the Einstein Relation in Average Crystals Self-Diffusion of Ag + Ions in α-Ag 2 S and α-Ag 2 Se
journal, August 1967

  • Okazaki, Hideo
  • Journal of the Physical Society of Japan, Vol. 23, Issue 2
  • DOI: 10.1143/jpsj.23.355

A molecular dynamics method for simulations in the canonical ensemble
journal, June 1984


Surface reconstructions and premelting of the (100) CaF 2 surface
journal, January 2019

  • Faraji, Somayeh; Ghasemi, S. Alireza; Parsaeifard, Behnam
  • Physical Chemistry Chemical Physics, Vol. 21, Issue 29
  • DOI: 10.1039/c9cp02213a

Predicting molecular properties with covariant compositional networks
journal, June 2018

  • Hy, Truong Son; Trivedi, Shubhendu; Pan, Horace
  • The Journal of Chemical Physics, Vol. 148, Issue 24
  • DOI: 10.1063/1.5024797

Deep learning
journal, May 2015

  • LeCun, Yann; Bengio, Yoshua; Hinton, Geoffrey
  • Nature, Vol. 521, Issue 7553
  • DOI: 10.1038/nature14539

Multilayer feedforward networks are universal approximators
journal, January 1989


How van der Waals interactions determine the unique properties of water
journal, July 2016

  • Morawietz, Tobias; Singraber, Andreas; Dellago, Christoph
  • Proceedings of the National Academy of Sciences, Vol. 113, Issue 30
  • DOI: 10.1073/pnas.1602375113

Structure, dynamics, and electronic structure of liquid Ag-Se alloys investigated by ab initio simulation
journal, July 1996


Ab initio thermodynamics of liquid and solid water
journal, January 2019

  • Cheng, Bingqing; Engel, Edgar A.; Behler, Jörg
  • Proceedings of the National Academy of Sciences, Vol. 116, Issue 4
  • DOI: 10.1073/pnas.1815117116

A Limited Memory Algorithm for Bound Constrained Optimization
journal, September 1995

  • Byrd, Richard H.; Lu, Peihuang; Nocedal, Jorge
  • SIAM Journal on Scientific Computing, Vol. 16, Issue 5
  • DOI: 10.1137/0916069

Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces
journal, April 2007


Projector augmented-wave method
journal, December 1994


How to improve the accuracy of equilibrium molecular dynamics for computation of thermal conductivity?
journal, May 2010


Generalized Gradient Approximation Made Simple
journal, October 1996

  • Perdew, John P.; Burke, Kieron; Ernzerhof, Matthias
  • Physical Review Letters, Vol. 77, Issue 18, p. 3865-3868
  • DOI: 10.1103/physrevlett.77.3865

Potential Energy Surfaces Fitted by Artificial Neural Networks
journal, March 2010

  • Handley, Chris M.; Popelier, Paul L. A.
  • The Journal of Physical Chemistry A, Vol. 114, Issue 10
  • DOI: 10.1021/jp9105585

Accelerating crystal structure prediction by machine-learning interatomic potentials with active learning
journal, February 2019

  • Podryabinkin, Evgeny V.; Tikhonov, Evgeny V.; Shapeev, Alexander V.
  • Physical Review B, Vol. 99, Issue 6
  • DOI: 10.1103/physrevb.99.064114

Approximation by superpositions of a sigmoidal function
journal, December 1989

  • Cybenko, G.
  • Mathematics of Control, Signals, and Systems, Vol. 2, Issue 4
  • DOI: 10.1007/bf02551274

Interatomic potentials for ionic systems with density functional accuracy based on charge densities obtained by a neural network
journal, July 2015


Representing high-dimensional potential-energy surfaces for reactions at surfaces by neural networks
journal, September 2004


Grand canonical molecular dynamics simulations of Cu–Au nanoalloys in thermal equilibrium using reactive ANN potentials
journal, December 2015


An implementation of artificial neural-network potentials for atomistic materials simulations: Performance for TiO2
journal, March 2016


Structure of aqueous NaOH solutions: insights from neural-network-based molecular dynamics simulations
journal, January 2017

  • Hellström, Matti; Behler, Jörg
  • Physical Chemistry Chemical Physics, Vol. 19, Issue 1
  • DOI: 10.1039/c6cp06547c

Proton-Transfer Mechanisms at the Water–ZnO Interface: The Role of Presolvation
journal, March 2017

  • Quaranta, Vanessa; Hellström, Matti; Behler, Jörg
  • The Journal of Physical Chemistry Letters, Vol. 8, Issue 7
  • DOI: 10.1021/acs.jpclett.7b00358

SchNetPack: A Deep Learning Toolbox For Atomistic Systems
journal, November 2018

  • Schütt, K. T.; Kessel, P.; Gastegger, M.
  • Journal of Chemical Theory and Computation, Vol. 15, Issue 1
  • DOI: 10.1021/acs.jctc.8b00908

Enhanced thermoelectric performance in the very low thermal conductivity Ag 2 Se 0.5 Te 0.5
journal, September 2013

  • Drymiotis, Fivos; Day, Tristan W.; Brown, David R.
  • Applied Physics Letters, Vol. 103, Issue 14
  • DOI: 10.1063/1.4824353

High performance n-type Ag2Se film on nylon membrane for flexible thermoelectric power generator
journal, February 2019


Active learning of uniformly accurate interatomic potentials for materials simulation
journal, February 2019


Constant pressure molecular dynamics algorithms
journal, September 1994

  • Martyna, Glenn J.; Tobias, Douglas J.; Klein, Michael L.
  • The Journal of Chemical Physics, Vol. 101, Issue 5
  • DOI: 10.1063/1.467468

Thermal Conductivity during Phase Transitions
journal, December 2018


Explicit reversible integrators for extended systems dynamics
journal, April 1996

  • Martyna, Glenn J.; Tuckerman, Mark E.; Tobias, Douglas J.
  • Molecular Physics, Vol. 87, Issue 5
  • DOI: 10.1080/00268979600100761

QXMD: An open-source program for nonadiabatic quantum molecular dynamics
journal, July 2019


First Principles Neural Network Potentials for Reactive Simulations of Large Molecular and Condensed Systems
journal, August 2017


Deep Potential Molecular Dynamics: A Scalable Model with the Accuracy of Quantum Mechanics
journal, April 2018


From ultrasoft pseudopotentials to the projector augmented-wave method
journal, January 1999


Review on structural and dynamical properties of silver chalcogenides
journal, July 1990


Active learning of linearly parametrized interatomic potentials
journal, December 2017


Atom-centered symmetry functions for constructing high-dimensional neural network potentials
journal, February 2011

  • Behler, Jörg
  • The Journal of Chemical Physics, Vol. 134, Issue 7
  • DOI: 10.1063/1.3553717