Theoretical aspects of metal cluster chemistry

During the last twenty years the Polyhedral Skeletal Electron Pair Theory and the isolobal analogy have provided a theoretical basis for the rapid experimental developments, which have occurred in metal cluster chemistry. These theoretical principles have been underpinned by m.o. calculations on specific molecules and more generally by the Tensor Surface Harmonic Theory. This paper will review the important theoretical developments and relate them to the experimental and structural data which have been obtained for cluster compounds in our own and other laboratories. In particular the application of theoretical principles for rationalising and predicting the structures of cluster compounds of the platinum metals and gold are discussed. The bonding requirements of interstitial atoms and fragments are discussed, particularly in the context of interstitial C, B, N and transition metal atoms and diatomic fragments, e.g. C2 and C-H. POLYHEDRAL SKELETAL ELECTRON PAIR THEORY The historical development of the theoretical ideas which have contributed to the Polyhedral Skeletal Electron Pair Theory can be traced back to the pioneering work of Longuet-Higgins forty years ago. In the 1950's the contributions of Longuet-Higgins proved decisive since he not only provided a molecular orbital description of the bonding in diborane (ref. l), but also pioneered the application of molecular orbital theory ideas to deltahedral borides and boranes (ref. 2). The three-centre two-electron description of B-H-B and B-B-B bonds was elegantly generalised into the styx formalism and applied to all known boron hydrides by Lipscomb (ref. 3). The molecular orbital analysis of polyhedral boranes resulted in the successful prediction of the octahedral and icosahedral borane anions some years before they were structurally characterised (ref. 4). Hoffmann and Lipscomb (ref. 5) also developed the molecular orbital methodology of L nguet-Higgins in a general molecular orbital analysis of polyhedral borane anions, BnH;-, in 1962. In the 1960's Cotton and Haas (ref. 6) pioneered the development of molecular orbital ideas to metal cluster compounds of the early transition metals stabilised by halide ligands. Their analysis emphasised the important interactions which can result in such clusters from the overlap of the transition metal d orbitals. An alternative localised description of the bonding in such clusters, which resembled the s t y x methodology, was proposed by Kettle (ref. 7 ) . These theoretical models tended to emphasise the inherent differences between metal and main group clusters and in common with ligand field theory tended to stress the role of the metal d orbitals. By the mid-nineteen sixties the major classes of transition metal IT -donor and acceptor clusters, and main group polyhedral borane and Zintl 'naked' clusters had been established but were viewed as distinct areas of inorganic chemistry. Experimental studies had, however, begun to indicate the artificiality of these subdivisions. The synthesis of polyhedral organometallic compounds from the reactions of acetylenes with metal carbonyls by HUbel and Braye and transition metallocarboranes by Hawthorne's group (ref. 8) clearly demonstrated that it was possible to synthesise polyhedral molecules with transition metal and main group atoms at the vertices. In addition the structural characterisation of Rh6(C0Il6 by Dahl (ref. 9) provided a real difficulty for the theoretical models which were most widely used. The bonding in Rh6(C0)16 could not be explained by the effective atomic number rule and was not amenable to a molecular orb'tal and [Ta6Cl12] by Cotton and Haas (ref. 6). 6 8 analysis base$+on an extension of the molecular orbital schemes proposed for [Mo C1 ] 4+