Datensatz

Zusammenfassung

The accurate description of magnetic level energetics in oligonuclear exchange-coupled transition-metal complexes remains a formidable challenge for quantum chemistry. The density matrix renormalization group (DMRG) brings such systems for the
first time easily within reach of multireference wave
function methods by enabling the use of unprecedentedly large
active spaces. But does this guarantee systematic improvement in
predictive ability and, if so, under which conditions? We identify
operational parameters in the use of DMRG using as a test system
an experimentally characterized mixed-valence bis-μ-oxo/μ-acetato
Mn(III,IV) dimer, a model for the oxygen-evolving complex of
photosystem II. A complete active space of all metal 3d and bridge 2p orbitals proved to be the smallest meaningful starting
point; this is readily accessible with DMRG and greatly improves on the unrealistic metal-only configuration interaction or
complete active space self-consistent field (CASSCF) values. Orbital optimization is critical for stabilizing the antiferromagnetic
state, while a state-averaged approach over all spin states involved is required to avoid artificial deviations from isotropic behavior
that are associated with state-specific calculations. Selective inclusion of localized orbital subspaces enables probing the relative
contributions of different ligands and distinct superexchange pathways. Overall, however, full-valence DMRG-CASSCF
calculations fall short of providing a quantitative description of the exchange coupling owing to insufficient recovery of dynamic
correlation. Quantitatively accurate results can be achieved through a DMRG implementation of second order
N-electron valence
perturbation theory (NEVPT2) in conjunction with a full-valence metal and ligand active space. Perspectives for future
applications of DMRG-CASSCF/NEVPT2 to exchange coupling in oligonuclear clusters are discussed.