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

Title: Interatomic Potential Model Development: Finite-Temperature Dynamics Machine Learning

Abstract

Developing an accurate interatomic potential model is a prerequisite for achieving reliable results from classical molecular dynamics (CMD) simulations; however, most of the potentials are biased as specific simulation purposes or conditions are considered in the parameterization. Here, for developing an unbiased potential, a finite-temperature dynamics machine learning (FTD-ML) approach is proposed, and its processes and feasibility are demonstrated using the Buckingham potential model and aluminum (Al) as an example. Compared with conventional machine learning approaches, FTD-ML exhibits three distinguished features: 1) FTD-ML intrinsically incorporates more extensive configurational and conditional space for enhancing the transferability of developed potentials; 2) FTD-ML employs various properties calculated directly from CMD, for ML model training and prediction validation against experimental data instead of first-principles data; 3) FTD-ML is much more computationally cost effective than first-principles simulations, especially when the system size increases over 103 atoms as employed in this research for ensuring reliable training data. The Al Buckingham potential developed by the FTD-ML approach exhibits good performance for general simulation purposes. Thus, the FTD-ML approach is expected to contribute to a fast development of interatomic potential model suitable for various simulation purposes and conditions, without limitation of model type, while maintaining experimental-level accuracy.

Authors:
 [1]; ORCiD logo [1];  [2]
  1. Univ. of Tennessee, Knoxville, TN (United States). Dept. of Mechanical, Aerospace, and Biomedical Engineering
  2. Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States). Computer Science and Mathematics Division
Publication Date:
Research Org.:
Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
Sponsoring Org.:
USDOE; National Science Foundation (NSF)
OSTI Identifier:
1659562
Grant/Contract Number:  
AC05-00OR22725; ACI-1053575
Resource Type:
Accepted Manuscript
Journal Name:
Advanced Theory and Simulations
Additional Journal Information:
Journal Volume: 3; Journal Issue: 2; Journal ID: ISSN 2513-0390
Publisher:
Wiley
Country of Publication:
United States
Language:
English
Subject:
42 ENGINEERING; aluminum; Buckingham potential; finite-temperature dynamics; interatomic potential development; machine learning

Citation Formats

Wang, Jiaqi, Shin, Seungha, and Lee, Sangkeun. Interatomic Potential Model Development: Finite-Temperature Dynamics Machine Learning. United States: N. p., 2019. Web. https://doi.org/10.1002/adts.201900210.
Wang, Jiaqi, Shin, Seungha, & Lee, Sangkeun. Interatomic Potential Model Development: Finite-Temperature Dynamics Machine Learning. United States. https://doi.org/10.1002/adts.201900210
Wang, Jiaqi, Shin, Seungha, and Lee, Sangkeun. Tue . "Interatomic Potential Model Development: Finite-Temperature Dynamics Machine Learning". United States. https://doi.org/10.1002/adts.201900210. https://www.osti.gov/servlets/purl/1659562.
@article{osti_1659562,
title = {Interatomic Potential Model Development: Finite-Temperature Dynamics Machine Learning},
author = {Wang, Jiaqi and Shin, Seungha and Lee, Sangkeun},
abstractNote = {Developing an accurate interatomic potential model is a prerequisite for achieving reliable results from classical molecular dynamics (CMD) simulations; however, most of the potentials are biased as specific simulation purposes or conditions are considered in the parameterization. Here, for developing an unbiased potential, a finite-temperature dynamics machine learning (FTD-ML) approach is proposed, and its processes and feasibility are demonstrated using the Buckingham potential model and aluminum (Al) as an example. Compared with conventional machine learning approaches, FTD-ML exhibits three distinguished features: 1) FTD-ML intrinsically incorporates more extensive configurational and conditional space for enhancing the transferability of developed potentials; 2) FTD-ML employs various properties calculated directly from CMD, for ML model training and prediction validation against experimental data instead of first-principles data; 3) FTD-ML is much more computationally cost effective than first-principles simulations, especially when the system size increases over 103 atoms as employed in this research for ensuring reliable training data. The Al Buckingham potential developed by the FTD-ML approach exhibits good performance for general simulation purposes. Thus, the FTD-ML approach is expected to contribute to a fast development of interatomic potential model suitable for various simulation purposes and conditions, without limitation of model type, while maintaining experimental-level accuracy.},
doi = {10.1002/adts.201900210},
journal = {Advanced Theory and Simulations},
number = 2,
volume = 3,
place = {United States},
year = {2019},
month = {12}
}

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

Save / Share:

Works referenced in this record:

SPFP: Speed without compromise—A mixed precision model for GPU accelerated molecular dynamics simulations
journal, February 2013

  • Le Grand, Scott; Götz, Andreas W.; Walker, Ross C.
  • Computer Physics Communications, Vol. 184, Issue 2
  • DOI: 10.1016/j.cpc.2012.09.022

Probing the limits of metal plasticity with molecular dynamics simulations
journal, September 2017

  • Zepeda-Ruiz, Luis A.; Stukowski, Alexander; Oppelstrup, Tomas
  • Nature, Vol. 550, Issue 7677
  • DOI: 10.1038/nature23472

A note on a general definition of the coefficient of determination
journal, January 1991


Cramming More Components Onto Integrated Circuits
journal, January 1998


Molecular Dynamics Simulation of Poly(ethylene terephthalate) Oligomers
journal, January 2010

  • Wang, Qifei; Keffer, David J.; Petrovan, Simioan
  • The Journal of Physical Chemistry B, Vol. 114, Issue 2
  • DOI: 10.1021/jp909762j

Detecting Novel Associations in Large Data Sets
journal, December 2011


Machine Learning for Thermal Transport Analysis of Aluminum Alloys with Precipitate Morphology
journal, January 2019

  • Wang, Jiaqi; Yousefzadi Nobakht, Ali; Blanks, James Dean
  • Advanced Theory and Simulations, Vol. 2, Issue 4
  • DOI: 10.1002/adts.201800196

Development of a machine learning potential for graphene
journal, February 2018


High accuracy and transferability of a neural network potential through charge equilibration for calcium fluoride
journal, March 2017


The General Utility Lattice Program ( GULP )
journal, May 2003


Connectivity-Based Parallel Replica Dynamics for Chemically Reactive Systems: From Femtoseconds to Microseconds
journal, October 2013

  • Joshi, Kaushik L.; Raman, Sumathy; van Duin, Adri C. T.
  • The Journal of Physical Chemistry Letters, Vol. 4, Issue 21
  • DOI: 10.1021/jz4019223

A thermodynamic approach to determine accurate potentials for molecular dynamics simulations: thermoelastic response of aluminum
journal, June 2009

  • Winey, J. M.; Kubota, Alison; Gupta, Y. M.
  • Modelling and Simulation in Materials Science and Engineering, Vol. 17, Issue 5
  • DOI: 10.1088/0965-0393/17/5/055004

Models for predicting temperature dependence of material properties of aluminum
journal, February 2014


Roles of atomic restructuring in interfacial phonon transport
journal, August 2010


Interatomic potential for the Al-Cu system
journal, February 2011


Ideal strength and structural instability of aluminum at finite temperatures
journal, August 2012


Geometrical Effects on Sintering Dynamics of Cu–Ag Core–Shell Nanoparticles
journal, July 2016

  • Wang, Jiaqi; Shin, Seungha; Hu, Anming
  • The Journal of Physical Chemistry C, Vol. 120, Issue 31
  • DOI: 10.1021/acs.jpcc.6b05515

Modelling defects in Ni–Al with EAM and DFT calculations
journal, April 2016

  • Bianchini, F.; Kermode, J. R.; De Vita, A.
  • Modelling and Simulation in Materials Science and Engineering, Vol. 24, Issue 4
  • DOI: 10.1088/0965-0393/24/4/045012

Fast Parallel Algorithms for Short-Range Molecular Dynamics
journal, March 1995


Periodic ab initio determination of interatomic potentials for alumina
journal, October 1992

  • Gale, J. D.; Catlow, C. R. A.; Mackrodt, W. C.
  • Modelling and Simulation in Materials Science and Engineering, Vol. 1, Issue 1
  • DOI: 10.1088/0965-0393/1/1/007

Extending the Time Scale in Atomistic Simulation of Materials
journal, August 2002


XSEDE: Accelerating Scientific Discovery
journal, September 2014

  • Towns, John; Cockerill, Timothy; Dahan, Maytal
  • Computing in Science & Engineering, Vol. 16, Issue 5
  • DOI: 10.1109/MCSE.2014.80

Self-consistent interatomic potentials for the simulation of binary and ternary oxides
journal, January 1994

  • Bush, Timothy S.; Gale, Julian D.; Catlow, C. Richard A.
  • Journal of Materials Chemistry, Vol. 4, Issue 6
  • DOI: 10.1039/jm9940400831

Switch interactions control energy frustration and multiple flagellar filament structures
journal, March 2006

  • Kitao, A.; Yonekura, K.; Maki-Yonekura, S.
  • Proceedings of the National Academy of Sciences, Vol. 103, Issue 13
  • DOI: 10.1073/pnas.0510285103

Perspective: Machine learning potentials for atomistic simulations
journal, November 2016

  • Behler, Jörg
  • The Journal of Chemical Physics, Vol. 145, Issue 17
  • DOI: 10.1063/1.4966192

Machine Learning Force Fields: Construction, Validation, and Outlook
journal, December 2016


The Thermal Expansion of Pure Metals: Copper, Gold, Aluminum, Nickel, and Iron
journal, October 1941


Machine learning of accurate energy-conserving molecular force fields
journal, May 2017

  • Chmiela, Stefan; Tkatchenko, Alexandre; Sauceda, Huziel E.
  • Science Advances, Vol. 3, Issue 5
  • DOI: 10.1126/sciadv.1603015

Accurate interatomic force fields via machine learning with covariant kernels
journal, June 2017


Atomistic simulation of the premelting of iron and aluminum: Implications for high-pressure melting-curve measurements
journal, December 2009


Atomic interaction of the MEAM type for the study of intermetallics in the Al–U alloy
journal, December 2015


A Note on a General Definition of the Coefficient of Determination
journal, September 1991


Comparison of Model Selection for Regression
journal, July 2003


The Heat Capacities and Entropies of Aluminum and Copper from 15 to 300°K.
journal, July 1941

  • Giauque, W. F.; Meads, P. F.
  • Journal of the American Chemical Society, Vol. 63, Issue 7
  • DOI: 10.1021/ja01852a027

Large-Scale Molecular Dynamics Simulations of Self-Assembling Systems
journal, August 2008


Structure evolution of nanoparticulate Fe 2 O 3
journal, January 2015

  • Erlebach, Andreas; Kurland, Heinz-Dieter; Grabow, Janet
  • Nanoscale, Vol. 7, Issue 7
  • DOI: 10.1039/C4NR06989G