The nature of the bonding in the above carbonyls was studied using the analysis of domain averaged Fermi holes (DAFH). The results straightforwardly confirm the conclusions of earlier theoretical studies in which the existence of direct metal-metal bond, anticipated for the above carbonyls on the basis of 18-electron rule, was questioned. In addition to indicating the lack of direct metal-metal bond, the DAFH analysis also allowed to characterize the nature of the electron pairs involved in the bonding of the bridging ligands. The analysis has shown that because the number of available electron pairs is not sufficient for the formation of ordinary localized 2c-2e bonds between terminal M(CO)(3) fragments and the bridging ligands, the bonding in both carbonyls exhibits typical features of electron deficiency and one bonding electron pair is effectively involved in multicenter 3c-2e bonding. Because of the symmetry of the complexes the bridging ligands are not distinguishable and all M-C-M bridges have a partial 3c-2e nature via resonance of the localized structures.

This article reports the numerical comparison of the quantities characterizing the extent of electron fluctuation and pair localization in the domains determined by the direct minimization of electron fluctuation with the domains resulting from the partitioning of the molecules based on the topological analysis of the so-called electron localization function (ELF). Such a comparison demonstrates that the ELF partitioning can be regarded as a feasible alternative to computationally much more demanding direct optimization of minimum fluctuation domains. This opened the possibility of the systematic scrutiny of the electron pair model of the chemical bond, and as it was demonstrated, the previous pessimistic claims about the applicability of this model are not completely justified.

This article reports the application of a recently proposed formalism of domain averaged Fermi holes to the problem of the localization of electron pairs in electron localization function (ELF) domains and its possible implications for the electron pair model of chemical bond. The main focus was on the systems, such as H 2 O or N 2 , in which the ''unphysical'' population of ELF domains makes the parallel between these domains and chemical bond questionable. On the basis of the results of the Fermi-hole analysis, we propose that the above problems could be due to the fact that in some cases the boundaries of the ELF domains need not be determined precisely enough.

No abstract