A significant challenge in rare-event simulations is dealing with superbasins, or groups of recurrent free-energy minima separated from the full phase space of the system by high barriers that can considerably reduce the efficiency. I will discuss two different ways that we have recently dealt with this problem. In the first way, we discuss consolidation of neighboring free-energy minima using a bias potential in hyperdynamics simulations. We demonstrate the application of this protocol to understanding structural transformations of GaAs(001). We also present an algorithm for identifying local superbasins in kinetic Monte Carlo (kMC) simulations and we discuss the integration of this algorithm into kMC simulations containing one or more local superbasins, as well as non-superbasin events. We demonstrate, in several examples, that significant increases in computational efficiency can be achieved using these algorithms.
Copyright. All Rights Reserved.