Time-Scaling in Atomistics and the Rate-Dependent Mechanical Behavior of Nanostructures

Conventional molecular dynamics simulations enable the elucidation of an astonishing array of phenomena inherent in the mechanical and chemical behavior of materials. Unfortunately, current computational limitations preclude accounting for processes whose transition times exceed, at best, microseconds. This limitation severely impacts, among others, a realistic assessment of slow-strain-rate mechanical behavior. In this work, using a simple paradigmatical model of a metallic nanopillar that is often the subject of experimental works, we attempt to circumvent the time-scale bottleneck of conventional molecular dynamics and provide novel physical insights into the rate-dependence of mechanical behavior of nanostructures. Using a collection of algorithms that include a recently developed potential energy surface sampling method - the so-called autonomous basin climbing approach, kinetic Monte Carlo, and others, we assess the nanopillar mechanical behavior under strain rates ranging from 1 to 10⁸ s^-1. While our results for high-strain rate behavior are consistent with conventional molecular dynamics, we find that the response of nanostructures to slow compression is "liquid-like" and accompanied by extensive surface reconstructions.