Bridging the Gap Between Quantum Mechanics and Large-Scale Atomistic Simulation

Abstract

The prospect of modeling across disparate length and time scales to achieve a predictive multiscale description of real materials properties has attracted widespread research interest in the last decade [1]. To be sure, the challenges in such multiscale modeling are many, and in demanding cases, such as mechanical properties or dynamic phase transitions, multiple bridges extending from the atomic level all the way to the continuum level must be built. Although often overlooked in this process, one of the most fundamental and important problems in multiscale modeling is that of bridging the gap between first-principles quantum mechanics, from which true predictive power for real materials emanates, and the large-scale atomistic simulation of thousands or millions of atoms, which is usually essential to describe the complex atomic processes that link to higher length and time scales. For example, to model single-crystal plasticity at micron length scales via dislocation-dynamics simulations that evolve the detailed dislocation microstructure requires accurate large-scale atomistic information on the mobility and interaction of individual dislocations. Similarly, modeling the kinetics of structural phase transitions requires linking accurate large-scale atomistic information on nucleation processes with higher length and time scale growth processes.