EXPANDING THE BOND-VALENCE MODEL

The bond-valence model (BVM) has been successful for rationalizing combinations of bond lengths, but the domain of applicability in which the model is quantitatively accurate is limited. First, the model has typically been applied to ionic and polar covalent, rather than fully covalent or metallic, bonds. Second, bonds between a given atom pair in crystals are usually restricted to a relatively narrow range of lengths, so very simple exponential and power-law forms of the bond valence-length equation are usually adequate for application to most crystals. We show, however, that the usual forms of the equation are inadequate if we expand the model to address molecular structures. The extent of this inadequacy depends strongly on the bond character (ionic, covalent, metallic), and we propose new forms of the bond valence-length equation to address it. We also show that the BVM can be extended to address even very weak fully covalent bonds. Doing so actually improves bond-valence sums in complex materials, and provides a way to model full structures in a consistent, but simpler manner than has been possible in the past. To do so, however, we must take into account known differences in bond energy for the same nominal bond valence, as a function of bond character.