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

Title: Learning reduced kinetic Monte Carlo models of complex chemistry from molecular dynamics

Abstract

Here, we propose a novel statistical learning framework for automatically and efficiently building reduced kinetic Monte Carlo (KMC) models of large-scale elementary reaction networks from data generated by a single or few molecular dynamics simulations (MD). Existing approaches for identifying species and reactions from molecular dynamics typically use bond length and duration criteria, where bond duration is a fixed parameter motivated by an understanding of bond vibrational frequencies. Conversely, we show that for highly reactive systems, bond duration should be a model parameter that is chosen to maximize the predictive power of the resulting statistical model. We demonstrate our method on a high temperature, high pressure system of reacting liquid methane, and show that the learned KMC model is able to extrapolate more than an order of magnitude in time for key molecules. Additionally, our KMC model of elementary reactions enables us to isolate the most important set of reactions governing the behavior of key molecules found in the MD simulation. We develop a new data-driven algorithm to reduce the chemical reaction network which can be solved either as an integer program or efficiently using L1 regularization, and compare our results with simple count-based reduction. For our liquid methane system,more » we discover that rare reactions do not play a significant role in the system, and find that less than 7% of the approximately 2000 reactions observed from molecular dynamics are necessary to reproduce the molecular concentration over time of methane. Furthermore, we describe a framework in this work that paves the way towards a genomic approach to studying complex chemical systems, where expensive MD simulation data can be reused to contribute to an increasingly large and accurate genome of elementary reactions and rates.« less

Authors:
ORCiD logo [1];  [2];  [3]
  1. Stanford Univ., CA (United States). Inst. for Computational and Mathematical Engineering
  2. Pontifical Catholic Univ. of Chile, Santiago (Chile). Mathematical and Computational Engineering
  3. Stanford Univ., CA (United States). Inst. for Computational and Mathematical Engineering and Dept. of Materials Science and Engineering
Publication Date:
Research Org.:
Washington State Univ., Pullman, WA (United States)
Sponsoring Org.:
USDOE National Nuclear Security Administration (NNSA)
OSTI Identifier:
1367888
Grant/Contract Number:  
NA0002007
Resource Type:
Accepted Manuscript
Journal Name:
Chemical Science
Additional Journal Information:
Journal Volume: 679; Journal ID: ISSN 2041-6520
Publisher:
Royal Society of Chemistry
Country of Publication:
United States
Language:
English
Subject:
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY

Citation Formats

Yang, Qian, Sing-Long, Carlos A., and Reed, Evan J. Learning reduced kinetic Monte Carlo models of complex chemistry from molecular dynamics. United States: N. p., 2017. Web. https://doi.org/10.1039/c7sc01052d.
Yang, Qian, Sing-Long, Carlos A., & Reed, Evan J. Learning reduced kinetic Monte Carlo models of complex chemistry from molecular dynamics. United States. https://doi.org/10.1039/c7sc01052d
Yang, Qian, Sing-Long, Carlos A., and Reed, Evan J. Mon . "Learning reduced kinetic Monte Carlo models of complex chemistry from molecular dynamics". United States. https://doi.org/10.1039/c7sc01052d. https://www.osti.gov/servlets/purl/1367888.
@article{osti_1367888,
title = {Learning reduced kinetic Monte Carlo models of complex chemistry from molecular dynamics},
author = {Yang, Qian and Sing-Long, Carlos A. and Reed, Evan J.},
abstractNote = {Here, we propose a novel statistical learning framework for automatically and efficiently building reduced kinetic Monte Carlo (KMC) models of large-scale elementary reaction networks from data generated by a single or few molecular dynamics simulations (MD). Existing approaches for identifying species and reactions from molecular dynamics typically use bond length and duration criteria, where bond duration is a fixed parameter motivated by an understanding of bond vibrational frequencies. Conversely, we show that for highly reactive systems, bond duration should be a model parameter that is chosen to maximize the predictive power of the resulting statistical model. We demonstrate our method on a high temperature, high pressure system of reacting liquid methane, and show that the learned KMC model is able to extrapolate more than an order of magnitude in time for key molecules. Additionally, our KMC model of elementary reactions enables us to isolate the most important set of reactions governing the behavior of key molecules found in the MD simulation. We develop a new data-driven algorithm to reduce the chemical reaction network which can be solved either as an integer program or efficiently using L1 regularization, and compare our results with simple count-based reduction. For our liquid methane system, we discover that rare reactions do not play a significant role in the system, and find that less than 7% of the approximately 2000 reactions observed from molecular dynamics are necessary to reproduce the molecular concentration over time of methane. Furthermore, we describe a framework in this work that paves the way towards a genomic approach to studying complex chemical systems, where expensive MD simulation data can be reused to contribute to an increasingly large and accurate genome of elementary reactions and rates.},
doi = {10.1039/c7sc01052d},
journal = {Chemical Science},
number = ,
volume = 679,
place = {United States},
year = {2017},
month = {6}
}

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:

Neural Networks for the Prediction of Organic Chemistry Reactions
journal, October 2016


Exact stochastic simulation of coupled chemical reactions
journal, December 1977

  • Gillespie, Daniel T.
  • The Journal of Physical Chemistry, Vol. 81, Issue 25
  • DOI: 10.1021/j100540a008

Optimally-reduced kinetic models: reaction elimination in large-scale kinetic mechanisms
journal, November 2003


Efficient maximum likelihood parameterization of continuous-time Markov processes
journal, July 2015

  • McGibbon, Robert T.; Pande, Vijay S.
  • The Journal of Chemical Physics, Vol. 143, Issue 3
  • DOI: 10.1063/1.4926516

Model complexity reduction of chemical reaction networks using mixed-integer quadratic programming
journal, May 2013

  • Hannemann-Tamás, Ralf; Gábor, Attila; Szederkényi, Gábor
  • Computers & Mathematics with Applications, Vol. 65, Issue 10
  • DOI: 10.1016/j.camwa.2012.11.024

L1 Regularization-Based Model Reduction of Complex Chemistry Molecular Dynamics for Statistical Learning of Kinetic Monte Carlo Models
journal, January 2016

  • Yang, Qian; Sing-Long, Carlos A.; Reed, Evan J.
  • MRS Advances, Vol. 1, Issue 24
  • DOI: 10.1557/adv.2016.124

First principles simulation of a superionic phase of hydrogen fluoride (HF) at high pressures and temperatures
journal, July 2006

  • Goldman, Nir; Fried, Laurence E.
  • The Journal of Chemical Physics, Vol. 125, Issue 4
  • DOI: 10.1063/1.2220036

Modeling and Simulating Chemical Reactions
journal, January 2008


Kinetic mechanism reduction based on an integer programming approach
journal, February 2000


A Genetic Algorithm-Based Method for the Automatic Reduction of Reaction Mechanisms: AUTOMATIC REDUCTION OF REACTION MECHANISMS
journal, November 2013

  • Sikalo, N.; Hasemann, O.; Schulz, C.
  • International Journal of Chemical Kinetics, Vol. 46, Issue 1
  • DOI: 10.1002/kin.20826

Examination of the concept of degree of rate control by first-principles kinetic Monte Carlo simulations
journal, June 2009


A general method for numerically simulating the stochastic time evolution of coupled chemical reactions
journal, December 1976


Discovering chemistry with an ab initio nanoreactor
journal, November 2014

  • Wang, Lee-Ping; Titov, Alexey; McGibbon, Robert
  • Nature Chemistry, Vol. 6, Issue 12
  • DOI: 10.1038/nchem.2099

Model reduction for chemical kinetics: An optimization approach
journal, April 1999


Reduction of very large reaction mechanisms using methods based on simulation error minimization
journal, February 2009


Approximate accelerated stochastic simulation of chemically reacting systems
journal, July 2001

  • Gillespie, Daniel T.
  • The Journal of Chemical Physics, Vol. 115, Issue 4
  • DOI: 10.1063/1.1378322

Multiscale modeling of energetic materials: Easy to say, harder to do
conference, January 2012

  • Rice, Betsy M.
  • SHOCK COMPRESSION OF CONDENSED MATTER - 2011: Proceedings of the Conference of the American Physical Society Topical Group on Shock Compression of Condensed Matter, AIP Conference Proceedings
  • DOI: 10.1063/1.3686505

Stability of hydrocarbons at deep Earth pressures and temperatures
journal, April 2011

  • Spanu, L.; Donadio, D.; Hohl, D.
  • Proceedings of the National Academy of Sciences, Vol. 108, Issue 17
  • DOI: 10.1073/pnas.1014804108

Reduction of dynamical biochemical reactions networks in computational biology
journal, January 2012


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


First-principles and classical molecular dynamics simulation of shocked polymers
journal, February 2010


Simulations of Shocked Methane Including Self-Consistent Semiclassical Quantum Nuclear Effects
journal, October 2012

  • Qi, Tingting; Reed, Evan J.
  • The Journal of Physical Chemistry A, Vol. 116, Issue 42
  • DOI: 10.1021/jp308068c

The asymptotic properties of ML estimators when sampling from associated populations
journal, January 1962


    Works referencing / citing this record:

    New frontiers for the materials genome initiative
    journal, April 2019

    • de Pablo, Juan J.; Jackson, Nicholas E.; Webb, Michael A.
    • npj Computational Materials, Vol. 5, Issue 1
    • DOI: 10.1038/s41524-019-0173-4

    NOMAD: The FAIR concept for big data-driven materials science
    journal, September 2018