Atomistic modeling at experimental strain rates and timescales

Abstract

Modeling physical phenomena with atomistic fidelity and at laboratory timescales is one of the holy grails of computational materials science. Conventional molecular dynamics (MD) simulations enable the elucidation of an astonishing array of phenomena inherent in the mechanical and chemical behavior of materials. However, conventional MD, with our current computational modalities, is incapable of resolving timescales longer than microseconds (at best). In this short review article, we briefly review a recently proposed approach - the so-called autonomous basin climbing (ABC) method - that in certain instances can provide valuable information on slow timescale processes. We provide a general summary of the principles underlying the ABC approach, with emphasis on recent methodological developments enabling the study of mechanically-driven processes at slow (experimental) strain rates and timescales. Specifically, we show that by combining a strong physical understanding of the underlying phenomena, kinetic Monte Carlo, transition state theory and minimum energy pathway methods, the ABC method has been found to be useful in a variety of mechanically-driven problems ranging from the prediction of creep-behavior in metals, constitutive laws for grain boundary sliding, void nucleation rates, diffusion in amorphous materials to protein unfolding. Aside from reviewing the basic ideas underlying this approach, we emphasize some of the key challenges encountered in our own personal research work and suggest future research avenues for exploration.