Application of FLO-SIC to f-Electron Systems: Sixth Row Elements and Ligated Molecules

Alexander Irun Johnson, University of Texas at El Paso


Density functional theory - the most widely used theoretical method to study atoms, molecules, and solids - suffers from the well-known self-interaction error. A solution to the problem was suggested by Perdew and Zunger, who showed the self-interaction error can be removed with self-interaction correction. In 2014, Pederson showed a unitary transformation can be performed on the Kohn-Sham orbitals to generate Fermi-Löwdin orbitals which improve atomization energies, and avoid the computational costs of solving the localization equations. This method is known as the Fermi-Löwdin Orbital Self-Interaction Correction (FLO-SIC). Until now, the FLO-SIC methodology has been used for atoms not containing f-electrons, because f-electrons were not implemented in the FLOSIC code, which is based on NRLMOL. This work presents an implementation of an f-electron capable NRLMOL including FLOSIC. Difficulties and strategies of FLOSIC with f-electrons are discussed, such as generating parameters known as Fermi-orbital descriptors used to define Fermi-Löwdin orbitals. Highest occupied molecular orbital energies are compared to experimental ionization potentials for several 6th row elements, which are particularly affected by self-interaction error. For some of the open shell elements, DFT predicts incorrect ground state valence configurations which can be recovered with FLOSIC. The results suggest that FLOSIC is a useful and efficient method to cure self-interaction error for systems containing f-electrons. Additionally, potential applications to molecular magnets are discussed, which are in dire need of an effective ab initio theory for accurate predictions.

Subject Area

Computational physics|Physics

Recommended Citation

Johnson, Alexander Irun, "Application of FLO-SIC to f-Electron Systems: Sixth Row Elements and Ligated Molecules" (2022). ETD Collection for University of Texas, El Paso. AAI30241366.