The Study Of Structural Properties Of Clusters And The Efficiency Of Different Methods

Clusters are composed by two to thousands of atoms,molecules or ions through physical or chemical interaction. The global minimum structure properties of clusters are no doubt the most important one to study their other behaviors. In this paper, we first use Gupta potential and Sutton-Chen potential combined with genetic algorithm to perform a systematic study of the ground state structural properties of Irn(n=2-60) clusters. Our results show that the average binding energy per atom from Gupta potential is a little higher than that of from Sutton-Chen potential at n>4. For Irn clusters at n=22,29,35,37, the different global minimum geometrical structures are detected by using these two different potentials while others the same. Results from the two potentials both show that there exists structures competition between the icosahedron-like,the fcc-like and the quindecahedron-like in the evolution of cluster sizes, the average binding energy,the average coordination number and the average nearest-neighbor distance per atom of iridium clusters generally increase as the cluster sizes increase. The same magic numbers (n=13,19,23,38,55)of Irn(n=2-60) clusters are obtained for the two different potentials.In addition,in this papers'second part, The ground state structural properties of Con(n=2-60) clusters were studied by using Gupta potential combined with Random number method,Random gridding method,Molecular dynamics simulated quenching method,Molecular dynamics simulated annealing method and Genetic algorithm. The efficiency of these different methods is discussed. By comparison with standard database, our results show that Genetic algorithm has the best efficiency and all the ground state structural of Con (n=2-60) clusters have been found. Molecular dynamics simulated annealing takes second place and just few global minimum geometrical structures have not been found .The two different methods, Random number method and Random gridding method, both have well efficiency for Con clusters at n<35 . In addition, Influence of different random number,different gridding and different quenching temperature on results are also discussed.