An Off-Lattice Kinetic Monte Carlo Framework For Long-Time Atomistic Simulations

Abstract

The goal of this thesis is to develop an off-lattice Kinetic Monte Carlo (KMC) framework to simulate the atomistic dynamics of materials at extreme conditions over long time scales. Despite the dramatic increase in computational power over the last few decades, rigorous approaches such as classical Molecular Dynamics (MD) techniques cannot access the engineering and experimental time scales due to the fundamental scaling limitation constrained by atomic vibrations. KMC approaches are powerful stochastic computational techniques that focus on the simulation of rare atomistic events in order to analyze the coarse-grained dynamics of condensed matter systems and replicate non-equilibrium phenomena in a statistical fashion. However, their application to problems at extreme conditions — such as those encountered in materials science under high pressure, temperature, and radiation — has been limited by the complexity of atomistic interactions, by the variability and instability of underlying structures, and by the computational cost of simulating large systems over sufficiently long time-scales.

To address such challenges, this thesis proposes an off-lattice, modular and scalable KMC framework that features adaptive inferred structures, efficient process sampling and dynamic rate constant calculations, together with the corresponding Julia implementation. The developed KMC framework is justified theoretically, described step-by-step methodologically, and then validated against MD results for early-time dynamics.