Electronic Structure Tools for Transition Metal Complexes with Many Open Shells

Abstract

In this thesis we present three different tools to model open shell transition metal electronic structure, commonly referred to as multi-reference (MR) or multi-configurational electronic structure. The first tool allows us to incorporate spin-orbit coupling into a method that also rigorously treats the multiple configurations arising from electrons in open shells. We call this method the state interaction spin-orbit (SISO) coupling method using density matrix renormalization group (DMRG) wavefunctions. We implement the DMRG-SISO scheme using a spin-adapted DMRG algorithm that computes transition density matrices between arbitrary wavefunctions of the interacting electronic states. To demonstrate the potential of this method, we present accurate benchmark calculations for the zero-field splitting (ZFS) of the copper and gold atoms and also calculate, for the first time, the ZFS of a [2Fe-2S] complex.

The second tool builds on the first to allow the calculation of electron paramagnetic resonance (EPR) $g$-values, one of the main spectroscopic methods used to characterize paramagnetic metals in transition metal complexes. We apply this to several benchmark systems such as \ce{TiF3} and \ce{CuCl4^2-}, as well as to determine the $g$-tensor for a [2Fe-2S] complex. This work opens up the possibility to model the $g$-tensor of the active site metals in bioinorganic systems.

Finally, we introduce the atomic valence active space (AVAS), which is a simple automated technique to identify the important orbitals to treat in a multi-configurational electronic structure method. We discuss the background, theory, and implementation of the idea, and several of its variations are tested. To demonstrate the performance and accuracy, we calculate the excitation energies for various transition metal complexes in typical application scenarios. The described technique makes MR calculations easier to execute, easier to reproduce by any user, and simplifies the determination of the appropriate size of the active space required for accurate results.