FLO-SIC: Efficient Density Functional Theory Calculations without Self-Interaction

Density functional theory (DFT) is the foundation for much of the computational modeling done in chemistry and materials science today. Current approaches based on DFT are remarkably successful in describing molecules and solids when atoms are near their equilibrium positions. Bond lengths, bond angles, and vibrational frequencies are examples of properties that can be modeled very accurately by modern DFT methods. Yet self-interaction error (SIE) is an unavoidable flaw that poisons all existing DFT calculations, most remarkably when electrons are shared over stretched bonds, as in the transition states of chemical reactions, and in many transition metal complexes even near equilibrium geometries. Attempts to remove the SIE indicate that orbital-by-orbital self-interaction corrections (SIC) can largely eliminate these problems, but previous methodologies used to implement SIC were too computationally demanding to be practical for general use or even to perform tests on large systems. It is therefore important to develop a computational technology that can provide efficient and reliable self-interaction-free DFT calculations. A recent idea based on localized Fermi orbitals, shows promise towards this end, but the method has yet to be fully implemented. The over-arching goal of this project is to develop, test, and validate a user-friendly, open-source software package that implements the Fermi-Löwdin-orbital-based self-interaction-correction (FLO-SIC) to DFT. The open-source software development effort will be integrated with theoretical research investigating the use of FLO-SIC with sophisticated semilocal density functionals like the new 'strongly constrained and appropriately normed' SCAN functional and its proposed successor, the ultimate semilocal functional. Additionally, the FLO-SIC code will be tested and validated in combined computational and experimental research involving heterogeneous, transition-metal-based catalysts, and magnetic molecules. If it successfully passes these tests, the FLO-SIC formalism will replace conventional DFT methods with a predictive, self-interaction-free method for studying chemical systems.