Abstract
We introduce a new kinetic Monte Carlo (KMC) algorithm for off-lattice simulation. In off-lattice KMC, one needs to calculate the rates for all possible moves from the current state by searching the energy landscape for index 1 saddle points surrounding the current basin of attraction. We introduce a rejection scheme in which the true rates, are replaced by rate estimates. This is done by first associating each saddle point with the atom that would move the most if that transition were to take place and then constructing an estimate for the total rate associated with each atom by using a nearest-neighbor bond count. These estimates allow one to select a set of possible transitions, one of which is accepted or rejected on the basis of a localized saddle point search focused on a particular atom. In principle, this allows a performance boost that scales with the number of particles in the system. We test the method on a growing two-species nanocluster and find we can reduce the computation time by 90% for clusters that contain similar to 55 particles and 96% for clusters that contain similar to 65 particles.