Structural evolution and stability of PtCo clusters from global optimization and DFT insights

Abstract

The structural optimization of bimetallic clusters requires the simultaneous exploration of atomic coordinates and chemical ordering. Here, an improved differential evolution algorithm with a multi-subpopulation collaborative architecture is applied to PtCo clusters at two representative magic sizes, N= 13 and N= 38. The algorithm combines differentiated mutation strategies, adaptive Pt–Co atom exchange, inter-subpopulation information sharing, and local relaxation. Benchmark and ablation tests show that the atom-exchange operation is particularly important for improving the reliability and convergence rate of the search. The optimized structures reveal a marked contrast between the two sizes: the N =13 clusters retain an icosahedral framework across the compositional range, whereas the N = 38 clusters exhibit composition-dependent fcc-like, mixed, and disordered motifs. The calculated excess energies indicate favorable Pt–Co mixing over a broad range of compositions. Structural descriptors and electronic-structure calculations further associate the enhanced stability of mixed clusters with abundant heteroatomic coordination, charge redistribution, and Pt–Co d-state interactions. These results provide a computational framework for examining how composition influences structure and stability in bimetallic clusters.