Introduction

In earlier chapters, allusions were made to the effects of covalent bonding. For example, covalent interactions were invoked to account for the intensification of absorption bands in crystal field spectra when transition metal ions occupy tetrahedral sites (3.7.1); patterns of cation ordering for some transition metal ions in silicate crystal structures imply that covalency influences the intracrystalline (or intersite) partitioning of these cations (6.8.4); and, the apparent failure of the Goldschmidt Rules to accurately predict the fractionation of transition elements during magmatic crystallization was attributed to covalent bonding characteristics of these cations (8.3.2).

A fundamental assumption underlying the crystal field model of chemical bonding is that ligands may be treated as point negative charges with no overlap of metal and ligand orbitals. Thus, 3d electrons are assumed to remain entirely on the transition metal ion with no delocalization into ligand orbitals. This situation is never realized, even in ionic structures such as periclase (MgO) and forsterite (Mg2SiO4), let alone bunsenite (NiO), liebenbergite (Ni2SiO4) or fayalite (Fe2SiO4), in which metal–oxygen bonds have some degree of covalent character and electrons in metal orbitals participate in the bonding. Some of the fundamental features of crystal field theory are contrary to expectation or are impossible to derive using the point charge model (Cotton, 1964).