Chemical and Biological Engineering MS Thesis Defense by Ezgi Erdem








Title: Global Optimization of Pt-Cu Clusters using a Genetic Algorithm and Density Functional Theory


Speaker: Ezgi Erdem


Time: August 8, 2017, 10:30


Place: ENG 208

Koç University

Rumeli Feneri Yolu

Sariyer, Istanbul

Thesis Committee Members:

Prof. Can Erkey (Advisor, Koç University)

Prof. Ersin Yurtsever (Koç University)

Prof. Ramazan Yıldırım (Boğaziçi University)


Bimetallic nanoparticles are important in a wide variety of applications ranging from catalysis to optics. Bimetallic clusters are used as models for bimetallic nanoparticles due to their fewer number of atoms. Their structures and segregation properties have been studied by theoretical approaches. In this study, the atomic arrangements and structures of monometallic Pt and Cu clusters and bimetallic PtCu clusters composed of 2-40 atoms were determined using a genetic algorithm (GA) with multiple runs between 100 and 400. Atomic interactions in the cluster were represented by the Gupta-type potential energy. The algorithm was developed and implemented in MATLAB environment. After a large number of GA runs on Pt, Cu and PtCu clusters, the lowest energy configurations were successfully determined. To improve our understanding of the structural and energetic properties of the mono- and bimetallic Pt-Cu clusters, a wide range of energetic and structural properties (i.e., excess, binding energy and second difference in energy, effective coordination number, average weighted bond length, and second order parameter) were calculated. Energetic analysis provided strong evidence that the lowest energy structures of Pt, Cu and PtCu clusters were stable and energetically favorable. The optimum structures for every size of Pt and Cu clusters were symmetric, regular and mainly based on icosahedron structures while PtCu clusters above 20 atoms were distorted. The lowest energy structures of PtCu clusters were found as mixed distorted icosahedrons with Pt segregation in core region whereas Cu atoms were located on the surface.   Remarkably, 38 atoms of Pt, Cu and PtCu alloy clusters tended to be perfect truncated octahedrons which was the similar face centered cubic packing as in bulk Pt and Cu. Further, global optimization of each composition of 10 atoms of PtCu clusters were carried out at the DFT level in Gaussian 09. Energetically the lowest energy composition was obtained as Pt3Cu7. Hydrogen and OH adsorption was studied to investigate how hydrogen interacts with Pt−Cu clusters for improving our understanding of the factors influencing the catalytic activity of such materials for oxygen reduction reaction. Hydrogen and OH adsorption energy and adsorption properties were found in the different charge states and in different adsorption sites as top, bridge and hollow Pt10, Cu10 and Pt3Cu7 clusters. More negative adsorption energy indicates stronger adsorption since the adsorption energy measures the magnitude of the binding energy of the species to the cluster. Based on our calculations, the lowest energy structures for both H and OH adsorption on Pt3Cu7 cluster were reached in the neutral state. The most favorable adsorption site of one hydrogen atom was found as the top site for Pt10 and bridge site for Cu10 cluster. For hydrogen adsorption on Pt3Cu7 cluster, the hollow site from Pt-rich site was found as more favorable. This result indicates that H atoms show a preference to bind to Pt atoms rather than Cu atoms in Pt3Cu7 cluster.