eReaxFF: A Pseudoclassical Treatment of Explicit Electrons within Reactive Force Field Simulations

Abstract

We present a computational tool, eReaxFF, for simulating explicit electrons within the framework of the standard ReaxFF reactive force field method. We treat electrons explicitly in a pseudoclassical manner that enables simulation several orders of magnitude faster than quantum chemistry (QC) methods, while retaining the ReaxFF transferability. We delineate here the fundamental concepts of the eReaxFF method and the integration of the Atom-condensed Kohn-Sham DFT approximated to second order (ACKS2) charge calculation scheme into the eReaxFF. We trained our force field to capture electron affinities (EA) of various species. As a proof-of-principle, we performed a set of molecular dynamics (MD) simulations with an explicit electron model for representative hydrocarbon radicals. We establish a good qualitative agreement of EAs of various species with experimental data, and MD simulations with eReaxFF agree well with the corresponding Ehrenfest dynamics simulations. The standard ReaxFF parameters available in the literature are transferrable to the eReaxFF method. The computationally economic eReaxFF method will be a useful tool for studying large-scale chemical and physical systems with explicit electrons as an alternative to computationally demanding QC methods.

Modeling Diffusion in Functional Materials: From Density Functional Theory to Artificial Intelligence

“…Thus, the picture that emerges from the recent progress in the field is that further development and application of multiscale approaches that take into account several important length and time‐scales holds great promise toward the simulation‐based design of better functional materials with less trial and error. With the development of ever sophisticated force fields, molecular dynamics simulations will meet the challenge by providing more accurate, atomistic‐scale descriptions of sub‐micrometer phenomena.…”

Machine Learning Force Field Parameters from Ab Initio Data

“…Here, to overcome the challenge encountered with force field parameterization using ML techniques, we have used the genetic algorithm (GA), which is an evolutionary algorithm that mimics the process of natural selection. 37 Previous successful applications of the genetic algorithm in similar contexts include the determination of the parameters for the ReaxFF reactive force field, 2–3, 5 and various force fields including Morse+QEq charge transfer ionic potential (CTIP), 6 and a new hybrid bond-order potential (HyBOP) for materials system. 2–3, 5–6, 38–39 …”

The role of shock waves on the biodamage induced by ion beam radiation

“…• From the theoretical point of view, classical molecular dynamics is incorporating the new tools of reactive force fields (Sushko et al 2016a;van Duin et al 2001;Abolfath et al 2011Abolfath et al , 2013 and irradiation-driven molecular dynamics (Sushko et al 2016b). These methods allow the simulation of the dynamics and reactivity of large and complex systems [and even the consideration of electrons within classical simulations (Su and Goddard 2007;Islam et al 2016)], taking advantage of the detailed information which can be obtained from ab initio approaches, either pure quantum calculations or first-principles molecular dynamics for the smaller systems (Fraile et al 2019;Smyth and Kohanoff 2012;Gaigeot et al 2010;Kohanoff et al 2017).…”

“…• This requires that reactive force fields (Sushko et al 2016a;van Duin et al 2001;Abolfath et al 2011Abolfath et al , 2013 include all the relevant chemical species in water radiolysis, not only free radicals, but also aqueous electrons. Some force fields are starting to be developed, where electrons can be explicitly considered (Su and Goddard 2007;Islam et al 2016). • New tools where the radiation (quantum) effects can be coupled to the system's (classical) dynamics, such as irradiation driven molecular dynamics (Sushko et al 2016b), need to be further developed.…”