Utilização de algoritmo genético para a determinação da estabilidade de aglomerados atômicos e as aplicações destes com grafeno

Abstract

In this thesis we propose a improvement of the Genetic Algorithm (GA) method that is used to search the stability of geometry of atomic clusters. The proposed improvements consist by introducing two new operators: the topological index and the discretization of the configurational space via multi-grid technique. This study using a semi-empirical Tight-Binding potential for the 19-atom Ti cluster showed that the multi-grid technique applied in the initial steps of the AG search selects regions on the potential energy surface accessible to the problem, however, there is no guarantee that the selected regions will be those associated with the global minimum. The use of the topological index results in a higher success rate in relation to the non use of this index, that is, a greater number of independent trajectories in the configurational space presents energies closer to the best solution found. This implementation of AG with topological index was applied to the study of Ti, Pd, Pt and Au clusters with up to 6 (six) atoms. Our structural results for these six-atom clusters demonstrate good agreement with the literature. Then We study the interaction of these 6 and 13 Ti, Pd, Pt and Au clusters with a pristine graphene sheet and containing a Stone-Wales (SW) defect. We show that the presence of the defect increases the binding energy of the six-metal clusters with the graphene sheet even when these clusters are deposited outside from the defect region. We also show that the magnitude of the clustergraphene interactions obey the d-band model. Finally, using the lower energy configuration for the Pt13-graphene interacting system (with SW defect), we studied (i) the cluster oxidation of Pt13 and (ii) we started the study of the process of transforming the CO molecule intoCO2 when the interaction of the CO molecule with the oxidized Pt13 cluster on the graphene sheet.