Book of Abstracts Book of Abstracts - Universität Konstanz

A - 2
104
Gold Clusters with Density-Functional Based Tight-Binding Method
Pekka Koskinen 1 , Hannu Häkkinen 1 , Gotthard Seifert 2 , and Michael Moseler 3
1 Department of Physics, NanoScience Center, 40014 University of Jyväskylä, Finland
2 Institut für Physikalische Chemie, TU Dresden, Mommsenstrasse. 13, 01062 Dresden, Germany
3 Fraunhofer Institut für Werkstoffmechanik, Wöhlerstrasse 11, 79108 Freiburg, Germany
In this work we employ the density-functional based tight-binding method (DFTB)[1] in the
study of structures and energies of neutral and charged gold clusters. These clusters have been
studied extensively with the density-functional approach (DFT), but for clusters of larger size or
for molecular dynamics simulations of more complex structures one should have a more
efficient yet accurate computational tool. In this work we demonstrate that this efficient method
might indeed be a useful tool in the study of gold clusters.
As an example, Fig. 1 shows the reproducibility of the relative energies of the isomers of Au - 7clusters,
as compared with previous DFT studies [2]. Even better than the energy, the method
gives mainly the same stable cluster geometries as density-functional calculations. The
electronic structure is well described, as can be seen in the vertical detachment energies in Fig 2.
The DFTB-method is able to show that the structure of the ground state of small gold anions
transforms from two-dimensional to three-dimensional at approximately correct cluster size. We
studied larger Au - 55-cluster isomers and found the same stable structures compared to previous
studies, as well as reasonable isomer energies [3].
Fig.1 Energies and structures of a few Au - 7-clusters.
The energies are calculated with self-consistent
charge (SCC) DFTB as well as with DFT and plotted
relative to the corresponding ground state.
References
Fig. 2 The vertical detachment energies of the Au - 7clusters
shown. The values are consistently lower
than gradient-corrected DFT (GGA), and the error is
of the same order than with local-density DFT
(LDA).
[1] M. Elstner, D. Porezag, G. Jungnickel, J. Elstner, M Haug, Th. Frauenheim, S. Suhai, and G.
Seifert, Phys. Rev. B 58, 7260 (1998).
[2] H. Häkkinen, M. Moseler, and U. Landman, Phys. Rev. Lett. 89, 033401 (2002).
[3] H. Häkkinen, M. Moseler, O. Kostko, N. Morgner, M.A. Hoffmann, and B.v.Issendorff, Phys. Rev.
Lett. 93, 093401 (2004)