Parts | Days | Selection | Search | Updates | Downloads | Help

TT: Fachverband Tiefe Temperaturen

TT 18: CE: Low-dimensional Systems - Materials 1

TT 18.1: Talk

Monday, March 14, 2011, 18:30–18:45, HSZ 201

Magnetic models from DFT cluster calculations - an alternative to supercell computations? — •Stefan Lebernegg¹, Miriam Schmitt², Ulrike Nitzsche³, and Helge Rosner² — ¹Universität Salzburg, 5081 Salzburg, Austria — ²MPI CPfS Dresden, 01187 Dresden, Germany — ³IFW Dresden, 01171 Dresden, Germany

For the treatment of strong electron correlation two standard approaches are commonly used: In principle, quantum chemistry can deal with strong correlations exactly, but only for small clusters. On the other hand, periodic 3D compounds can be calculated reliably in DFT codes, but the correlations are treated in a very approximate manner, often including external parameters like in the LSDA+U method. The far goal of this work is to find a way to construct reliable clusters capable to describe local properties of solids. In a first step toward this aim, magnetic properties of simple Cu²⁺ compounds with different ligands (N,O,F,Cl,Br) that exhibit edge sharing chains as their central building blocks are systematically studied. For the model clusters, effects of their finite size, embedding and termination as well as the substitution of Cu by formally nonmagnetic cations (broken chain) are investigated. The calculations are carried out within a full potential DFT code and are compared with corresponding known periodic 3D compounds (CuNCN, Li₂CuO₂, CuCl₂, CuBr₂). The presented cluster approach not only enables the application of highly accurate computational methods but might be used instead of elaborated supercell calculations. Moreover, additional degrees of freedom can provide deeper insight into magneto-structural correlations.