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

Title: A coarse-grained deep neural network model for liquid water

Abstract

We introduce a coarse-grained deep neural network (CG-DNN) model for liquid water that utilizes 50 rotational and translational invariant coordinates and is trained exclusively against energies of ~30 000 bulk water configurations. Our CG-DNN potential accurately predicts both the energies and the molecular forces of water, within 0.9 meV/molecule and 54 meV/Å of a reference (coarse-grained bond-order potential) model. The CG-DNN water model also provides good prediction of several structural, thermodynamic, and temperature dependent properties of liquid water, with values close to those obtained from the reference model. More importantly, CG-DNN captures the well-known density anomaly of liquid water observed in experiments. Our work lays the groundwork for a scheme where existing empirical water models can be utilized to develop a fully flexible neural network framework that can subsequently be trained against sparse data from high-fidelity albeit expensive beyond-DFT calculations.

Authors:
ORCiD logo [1];  [1];  [1]; ORCiD logo [1];  [2];  [3]
  1. Argonne National Lab. (ANL), Argonne, IL (United States). Center for Nanoscale Materials
  2. Univ. of Louisville, KY (United States)
  3. Argonne National Lab. (ANL), Argonne, IL (United States). Center for Nanoscale Materials; Univ. of Illinois, Chicago, IL (United States)
Publication Date:
Research Org.:
Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States). National Energy Research Scientific Computing Center (NERSC); Univ. of California, Oakland, CA (United States); Argonne National Lab. (ANL), Argonne, IL (United States)
Sponsoring Org.:
USDOE Office of Science (SC)
OSTI Identifier:
1577785
Alternate Identifier(s):
OSTI ID: 1573079
Grant/Contract Number:  
AC02-05CH11231; AC02-06CH11357
Resource Type:
Accepted Manuscript
Journal Name:
Applied Physics Letters
Additional Journal Information:
Journal Volume: 115; Journal Issue: 19; Journal ID: ISSN 0003-6951
Publisher:
American Institute of Physics (AIP)
Country of Publication:
United States
Language:
English
Subject:
42 ENGINEERING; Physics

Citation Formats

Patra, Tarak K., Loeffler, Troy D., Chan, Henry, Cherukara, Mathew J., Narayanan, Badri, and Sankaranarayanan, Subramanian K. R. S. A coarse-grained deep neural network model for liquid water. United States: N. p., 2019. Web. doi:10.1063/1.5116591.
Patra, Tarak K., Loeffler, Troy D., Chan, Henry, Cherukara, Mathew J., Narayanan, Badri, & Sankaranarayanan, Subramanian K. R. S. A coarse-grained deep neural network model for liquid water. United States. doi:10.1063/1.5116591.
Patra, Tarak K., Loeffler, Troy D., Chan, Henry, Cherukara, Mathew J., Narayanan, Badri, and Sankaranarayanan, Subramanian K. R. S. Mon . "A coarse-grained deep neural network model for liquid water". United States. doi:10.1063/1.5116591. https://www.osti.gov/servlets/purl/1577785.
@article{osti_1577785,
title = {A coarse-grained deep neural network model for liquid water},
author = {Patra, Tarak K. and Loeffler, Troy D. and Chan, Henry and Cherukara, Mathew J. and Narayanan, Badri and Sankaranarayanan, Subramanian K. R. S.},
abstractNote = {We introduce a coarse-grained deep neural network (CG-DNN) model for liquid water that utilizes 50 rotational and translational invariant coordinates and is trained exclusively against energies of ~30 000 bulk water configurations. Our CG-DNN potential accurately predicts both the energies and the molecular forces of water, within 0.9 meV/molecule and 54 meV/Å of a reference (coarse-grained bond-order potential) model. The CG-DNN water model also provides good prediction of several structural, thermodynamic, and temperature dependent properties of liquid water, with values close to those obtained from the reference model. More importantly, CG-DNN captures the well-known density anomaly of liquid water observed in experiments. Our work lays the groundwork for a scheme where existing empirical water models can be utilized to develop a fully flexible neural network framework that can subsequently be trained against sparse data from high-fidelity albeit expensive beyond-DFT calculations.},
doi = {10.1063/1.5116591},
journal = {Applied Physics Letters},
number = 19,
volume = 115,
place = {United States},
year = {2019},
month = {11}
}

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

Citation Metrics:
Cited by: 3 works
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

Temperature-dependent self-diffusion coefficients of water and six selected molecular liquids for calibration in accurate 1H NMR PFG measurements
journal, January 2000

  • Holz, Manfred; Heil, Stefan R.; Sacco, Antonio
  • Physical Chemistry Chemical Physics, Vol. 2, Issue 20
  • DOI: 10.1039/b005319h

Machine learning coarse grained models for water
journal, January 2019


Recent advances and applications of machine learning in solid-state materials science
journal, August 2019

  • Schmidt, Jonathan; Marques, Mário R. G.; Botti, Silvana
  • npj Computational Materials, Vol. 5, Issue 1
  • DOI: 10.1038/s41524-019-0221-0

Machine Learning Classical Interatomic Potentials for Molecular Dynamics from First-Principles Training Data
journal, January 2019

  • Chan, Henry; Narayanan, Badri; Cherukara, Mathew J.
  • The Journal of Physical Chemistry C, Vol. 123, Issue 12
  • DOI: 10.1021/acs.jpcc.8b09917

Screening billions of candidates for solid lithium-ion conductors: A transfer learning approach for small data
journal, June 2019

  • Cubuk, Ekin D.; Sendek, Austin D.; Reed, Evan J.
  • The Journal of Chemical Physics, Vol. 150, Issue 21
  • DOI: 10.1063/1.5093220

DeePCG: Constructing coarse-grained models via deep neural networks
journal, July 2018

  • Zhang, Linfeng; Han, Jiequn; Wang, Han
  • The Journal of Chemical Physics, Vol. 149, Issue 3
  • DOI: 10.1063/1.5027645

The structure of water around the compressibility minimum
journal, December 2014

  • Skinner, L. B.; Benmore, C. J.; Neuefeind, J. C.
  • The Journal of Chemical Physics, Vol. 141, Issue 21
  • DOI: 10.1063/1.4902412

A method for the solution of certain non-linear problems in least squares
journal, January 1944

  • Levenberg, Kenneth
  • Quarterly of Applied Mathematics, Vol. 2, Issue 2
  • DOI: 10.1090/qam/10666

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


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

A reappraisal of what we have learnt during three decades of computer simulations on water
journal, November 2002


A molecular level explanation of the density maximum of liquid water from computer simulations with a polarizable potential model
journal, February 2000


The atomic simulation environment—a Python library for working with atoms
journal, June 2017

  • Hjorth Larsen, Ask; Jørgen Mortensen, Jens; Blomqvist, Jakob
  • Journal of Physics: Condensed Matter, Vol. 29, Issue 27
  • DOI: 10.1088/1361-648X/aa680e

Physically informed artificial neural networks for atomistic modeling of materials
journal, May 2019


A survey of transfer learning
journal, May 2016


Construction of high-dimensional neural network potentials using environment-dependent atom pairs
journal, May 2012

  • Jose, K. V. Jovan; Artrith, Nongnuch; Behler, Jörg
  • The Journal of Chemical Physics, Vol. 136, Issue 19
  • DOI: 10.1063/1.4712397

A Density-Functional Theory-Based Neural Network Potential for Water Clusters Including van der Waals Corrections
journal, April 2013

  • Morawietz, Tobias; Behler, Jörg
  • The Journal of Physical Chemistry A, Vol. 117, Issue 32
  • DOI: 10.1021/jp401225b

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

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