Geometry Optimisation of Aluminium Clusters Using a Genetic Algorithm

The application of a Genetic Algorithm, for optimising the geometry of aluminium clusters with 21–55 atoms bound by the many‐body Murrell–Mottram potential, is described. In this size regime, a number of different structural motifs are identified—face‐centred cubic, hexagonal close packed, decahedral and icosahedral structures. The larger clusters consist of hollow icosahedral geometric shells, with Al55 having a centred icosahedral structure. Evolutionary Progress Plots for Al19 and Al38 reveal how the best structure evolves from generation to generation upon operation of the Genetic Algorithm. For aluminium clusters in the 21–55 atom size regime, a number of different structural motifs are identified—fcc, hcp, decahedral and icosahedral structures. The larger clusters consist of hollow icosahedral geometric shells, with Al55 having a centred icosahedral structure. The progress of the genetic algorithm is shown: Two selected parent clusters are cut at random and spliced to form the child cluster, whose fitness to act as a parent itself depends on its stability. Evolutionary Progress Plots reveals how these descendants evolve generation by generation towards the global minium.