Toward Full Configuration Interaction for Transition-Metal Complexes

An efficacious approximation to full configuration interaction (FCI) is adapted to calculate singlet-triplet gaps for transition-metal complexes. This strategy, incremental FCI (iFCI), uses a many-body expansion to systematically add correlation to a simple reference wave function and therefore achieves greatly reduced computational costs compared to FCI. iFCI through the 3-body expansion is demonstrated on four model transition-metal complexes involving the metals Zn, V, and Cu. Screening techniques to increase the computational efficiency of iFCI are proposed and tested, showing reduction in the number of 3-body terms by more than 90% with controlled errors. The largest complex treated herein by iFCI has 142 valence electrons, all of which are correlated among the full set of 444 active orbitals. Computed spin gaps approach experimental results for the four complexes, though room for improvement remains.

Automated summary

The efficient incremental full configuration interaction (iFCI) method uses a many-body expansion to add correlation systematically, reducing computational costs. iFCI is demonstrated on model transition-metal complexes through a 3-body expansion, with proposed screening techniques to increase computational efficiency. Computed spin gaps approach experimental results for the complexes, but there is room for improvement.