MLHybridReaxPaper
Supporting information for the manuscript titled "Machine Learning Assisted Hybrid ReaxFF Simulations "
- Authors:
- Dundar E. Yilmaz
Department of Mechanical Engineering,
The Pennsylvania State University,
University Park, PA 16802 - William Hunter Woodward DOW Analytical Sciences, The Dow Chemical Company, Midland, MI, 48674, USA
- Adri C. T. van Duin
Department of Mechanical Engineering,
The Pennsylvania State University,
University Park, PA 16802
-
corresponding author: duy42@psu.edu
-
Abstract:
We have developed a Machine Learning Assisted Hybrid ReaxFF Simulation method ("Hybrid/Reax"), which alternates reactive and non-reactive molecular dynamics simulations with the assistance of machine learning (ML) models to simulate phenomena that require longer time scales and/or larger systems than are typically accessible to ReaxFF. Hybrid/Reax uses a specialized tracking tool during the reactive simulations to further accelerate chemical reactions. Non-reactive simulations are used to equilibrate the system after the reactive simulations stage. ML models are used between reactive and non-reactive stages to predict non-reactive force field parameters of the system based on the updated bond topology. Hybrid/Reax simulation cycles can be continued until the desired chemical reactions are observed. As a case study, this method was used to study the crosslinking of a polyethylene matrix analogue (decane) with crosslinking agent dicumyl-peroxide. We were able to run relatively long simulations (>20 million MD steps) on a small test system (4660 atoms) to simulate crosslinking reactions of PE in the presence of dicumyl peroxide. Starting with 80 PE molecules, more than half of them crosslinked by the end of the Hybrid/Reax cycles on a single Xeon processor in under 48 hours. This simulation would take approximately one month if run with pure ReaxFF MD on the same machine.