An Modified Dynamic Lattice Searching Method For The Application Of Au Clusters

Clusters is a steady congery which is formed by force of several to thousand atoms. Most of propertices of Clusters are concerned with the structure of the atoms. The prediction for the lowest energy and structure of clusters is a foundamental and significative study. In the theory, the prediction for steady structure of clusters is a problem to optimize the multivariable function on the condition that potential function is confirmed. Therefore, two aspects should be done. Firstly, a reliable potential function which could describe the force of clusters must be confirmed a s an object function. Secondly,an efficient algorithm which could fast optimize the object function should be proposed.Gupta potential and Sutton-Chen potential are used to describe the force of Au atoms on the previous studies. Gupta potential is adopted there. Many algorithms have been proposed to optimize the energy of clusters, including basin-hopping algorithm, simulate annealing(SA), hierarchical greedy algorithm(HGA), genetic-symbiotic algorithm, heuristic algorithm with surface and interior(HA-SIO), dynamic lattice searching algorithm(DLS), simple greedy method with energy-based perturbation and a small population(EP-SGM).A modified dynamic lattice searching algorithm was proposed, and many new lowest energies have been found with the algorithm. Two sets of parameters which could represent Au clusters are taken in this paper. N=38-100 clusters are optimized with the first set of parameters(A=0.11844, B=1, p=10.15, q=4.13), and the energy of N=38,55 we got are better than the previous. On the other hand, N=100-200 clusters are optimized with the second set of parameters(A=0.2061, B=1.79, p=10.229, q=4.036), and the results of N=100,110,120,130,140,150,160,170,180,190,200 are better than the result s of Wu Xia.