Metal clusters delocalized bonding

To sum up all the experimental evidence whereas AU11L7X3 and smaller molecular cluster compounds are covalently bonded and show no tendency toward metallic binding, the bonding in Au55(PPh3),2Cl6 is delocalized, non-directional, and substantially metallic in character. 

Whereas in ligand bridged dinuclear complexes, removal or addition of two electrons makes or breaks one metal-metal bond (15) this does not seem to be the case for clusters, presumably because of their delocalized bonding. At least for one case, however, two-electron reduction can induce a significant change in cluster shape (18,42) the 84-electron cluster Os6(CO),g with framework 1 is easily reduced to the 86-electron anion Os6(CO) g with framework 2, in accordance with skeletal electron counting rules. 

From all the applications we have done in recent years [10-12], we review those that show the essence of our methodology. After introducing the VB formalism, we study the four electrons problem, a cluster of hydrogen, in an unusual limit, in order to address the problem of insulator to metal transition in solid hydrogen under pressure. Then we proceed to the applications to neutral and anionic lithium clusters, which are systems with very delocalized bonds. 

Several earlier review articles are relevant to our subject. Slichter reviews the work done in his laboratory [16], most of it concerned with atoms or molecules adsorbed on the metal clusters, and the experimental techniques used in such studies [17]. Duncan s review [9] pays special attention to the C NMR of adsorbed CO. Most recently, one of us has given a rather detailed review of the held, in particular on metal NMR of supported metal catalysts [18]. While the topics and examples discussed in this chapter will inevitably have some overlap with these previous reviews, particular emphasis is directed toward highlighting the ability of metal NMR to access the iff-LDOS at both metal surfaces and molecular adsorbates. The iff-LDOS is an attractive concept, in that it contains information on both a spatial (local) and energy (electronic excitations) scale. It can bridge the conceptual gap between localized chemical descriptors (e.g., the active site or the surface bond) and the delocalized descriptors of condensed matter physics (e.g., the band structure of the metal surfaces). 

Non-Metal Clusters. - For clusters of metal atoms, the electrons are largely delocalized over the complete cluster and directional bonds between the atoms do hardly exist. For other types of atoms, this is not the case, and one has to treat the electronic interactions explicitly in order to obtain accurate descriptions of the systems. This means that simplified potentials as the ones we discussed in the preceding section are not sufficiently accurate. However, when the electronic interactions are included explicitly, the computational demands on the calculations grow, which easily can put severe limitations on the possibilities of theoretical studies. Nevertheless, many studies have been devoted to those systems and here we shall briefly review some few of those, concentrating on just three elements. 

Even when these adaptations are taken into account, there remain important classes of compounds for which the model does not work. These are systems with delocalized bonding electrons (e.g. metals, metal cluster compounds and aromatic rings) or compounds in which formal oxidation states are zero or unassignable (e.g. metal carbonyls and jt complexes). 

Hydride ligands can adopt terminal, bridging or (in metal clusters) interstitial modes of bonding (23.9-23.12). A localized 2c-2e M—H bond is an appropriate description for a terminal hydride, delocalized 3c-2e or 4c-2e interactions describe p-H and P3-H interactions respectively (Figures 23.3a and 23.3b), and a 7c-2e interaction is appropriate for an interstitial hydride in an octahedral cage (Figure 23.3c). 

---