An unbiased optimization approach to study the structural evolution of gold nanoclusters
顏聰文1*, 賴山強1
1物理所, 中央大學, Taiwan
* presenting author:顏聰文, email:peter308@gmail.com
In this work, we propose two modifications to the well-known basin hopping (BH) optimization algorithm [D. J. Wales and J.P. Doye, J. Phys. Chem. A 101, 5111(1997)]
to study the structural evolution of gold nanoclusters (N=3-20). The first modification in this modified BH (MBH) algorithm is to define a confined spatial volume within which the cluster’s lowest energy structure is to be searched. The second modification is to introduce a cut-and-splice genetic operator to improve the searching performance of the energy minimum than the original BH technique. The valence electrons in a gold clusters Aun of size n were known to manifest relativistic effects and together with strong s-d hybridizations cause the structures to become planar even at a large size n. Predicting the structural evolution of gold nanoclusters as well as the 2D to 3D transition as a function of n is the motivation of this work. We have chosen the density-functional tight-binding (DFTB) theory to describe the electronic interactions semi-empirically mainly because electronic orbitals in Aun are calculated self-consistently with the s-d hybridization effect taken into account in the total potential energy function and hence are appropriate for the global structure optimization. With MBH/DFTB, the computation time is significantly reduced compared to numerical calculations utilizing the ab initio density-functional theory. In addition to structural predictions, electronic properties such as charge transfer and magnetic properties are studied as well. The optimized structures obtained in the preceding MBH/DFTB method are considered as initial structures for subsequent full density functional level calculation.


Keywords: optimization , density-functional tight-binding, magnetic properties, gold nanoclusters