| Abstract Scope |
Theoretical models for polycrystalline grain growth, such as the Srolovitz-McPherson law, are deterministic. However, computer simulations based on these models do not reliably predict experimental growth trajectories for individual grains. To understand this disagreement, we must examine the sources of uncertainty in simulation and experiment. All simulations include epistemic and aleatoric uncertainty. In contrast, experimental uncertainty is assumed to be primarily epistemic, limited by our knowledge of the initial state and by measurement resolution. The possibility of innate, aleatoric uncertainty is rarely considered. We perform grain growth simulations using the Molecular Dynamics method. Even when starting from identical initial structures, we find that growth trajectories of individual grains can vary significantly from run to run. We trace these differences to discrete, stochastic evolution events, including defect annihilation and topological transformations. The implication is that polycrystalline grain growth has a fundamental, aleatoric uncertainty that limits our ability to predict its outcomes. |