Structural optimization of Fe nanoclusters based on multi-populations differential evolution algorithm

Abstract

Structural optimization of Fe nanoclusters is of importance to their applications since their magnetic and catalytic properties are strongly dependent on their structures. In this article, the lowest-energy structures of Fe nanoclusters have been investigated using a multi-populations differential evolution algorithm. Finnis–Sinclair potentials and Johnson potentials have been, respectively, used to describe the interatomic interactions. The algorithm performance has been first evaluated by comparing the results of Finnis–Sinclair potentials with Cambridge Cluster Database. Two lower-energy structures of Fe n clusters (N = 83 and 84) have been discovered. Our results show that many of the structural configurations optimized with the two potentials are different, indicating that the potential model plays an important role in structural optimization. However, the structures of Fe clusters exhibit the similar growth patterns as the cluster increasing for the two potentials, that is, their lowest-energy structures contain many icosahedra, and the number of the icosahedral rings increases with their sizes. Furthermore, the stability analysis by investigating the finite difference and the second finite difference of energy indicates that the two potentials predict the same relative stability at the magic numbers. The comparison of optimized structures of Fe n clusters at magic numbers (N = 13, 19, 23, 26, 29, 55), respectively, using Finnis–Sinclair and Johnson potentials.