D-orbital energy levels in planar [MIIF4]2-, [MII(NH3)4]2+ and [MII(CN)4]2- complexes: the nature of M-L π bonding and the implications for ligand field theory.

Abstract

Qualitative MO theory predicts degenerate dπ orbitals for planar coordination complexes with formally σ-only ligands and the splitting energy, ΔEπ = E(dxy) - E(dxz,dyz), should be zero. For π-donor ligands, ΔEπ should be positive (dxy > dxz,yz) while for π-acceptors, ΔEπ should be negative (dxy < dxz,yz). However, experimental d-d spectra, ab initio ligand field theory (AI LFT) and crystal field theory for σ-only [M(NH3)4]2+ complexes give pronounced dπ splittings with ΔEπ around +2500 cm-1 for first-row, divalent metal ions. AI LFT further suggests ΔEπ values around +4500 cm-1 for [MF4]2- and +1000 cm-1 for [M(CN)4]2- species. The origins of these dπ orbital splittings can be traced to the effects of the ligand field potential surrounding the metal centre which includes not only the intrinsic metal-ligand π bonding but also substantial contributions from the 'void' regions above and below the molecular plane. The π component of the 'void cell' potentials increases ΔEπ which, if not explicitly taken into account, artificially enhances the apparent π-donor strength of the ligands. With the inclusion of void cell π interactions, even though the AI LFT d orbital sequence always places dxz,yz below dxy, the ligand field analysis provides a chemically-reasonable description of the M-L π interactions with cyanide being a weak π acceptor, ammonia being π-neutral and fluoride being a strong π donor. In the case of [Ni(CN)4]2-, ligand field calculations further show that, contrary to the recent claims of Oppenheim et al. (Inorg. Chem., 2019, 58, 15202) the sequence of the many-electron excited states is not a definitive guide to the underlying order of one-electron d orbital energies and that the observed sequence of nA2g > nEg > nB1g, n = 1 or 3, does not guarantee a d-orbital sequence of dxy < dxz,yz < dz2.