Multicanonical Monte Carlo simulations combined with neural network potentials for structure evolution of core-shell gold nanoclusters

Abstract

Evolution of gold nanoclusters have attracted great attention in the past decade due to their remarkable catalytic activities, size- and supportdependent CO oxidation, and structure-activity relationships. Here, we performed a global structure search of the low-energy cluster of Aun (n = 59, 60, 68, 70, 92, 106, 112, 138, 156, 166, 168) exhibit core-shell type structures by using multicanonical Monte Carlo simulations combined with artificial neural network potentials. We found that the most stable structure of gold nanoclusters exhibit icosahedral type core structure.