PROJECTS

Constructing density functionals compatible with exact exchange, utilizing the strong interaction limit

PRINCIPAL INVESTIGATOR

PROJECT RESEARCH ASSISTANT

ABSTRACT

Electronic-structure phenomena such as chemical reactions are expected to become a routine subject of computer simulations in the near future. One of the most promising numerical strategies for this purpose is density functional theory (DFT). Crucial for achieving ``chemical accuracy'' in a DFT calculation is the use of the exact exchange functional. Therefore, constructing correlation functionals that are compatible with exact exchange has become an ambitious goal in recent years. One such compatible correlation functional has been developed within an earlier phase of the present project. Using the information on electronic correlations provided by the strong-interaction limit of DFT, this ISI functional predicts the atomization energies of 18 standard test molecules with competitive accuracy. For routine practical use, however, the new functional needs two major improvements which are the main goal of this project. First, the PC model for the strong-interaction limit must be refined for the case of strong density gradients. The second improvement concerns the opposite weak-interaction limit where an approximation to the expensive second-order correlation energy shall be constructed from an extrapolation to negative interaction strengths with attractive electrons.