COMPARED ELECTRONIC-STRUCTURES OF NEGATIVE-IONS MPCN- .2. TRANSITION-METALS IN HUCKEL THEORY

SIMS experiments on transition metal carbides produce negative cluster ions M(p)C(n-) (n < 10) where the transition metal M is Ti, Zr, V, Cr, W, Fe and Ni with p = 1 and V with p = 2. The M(p)C(n-) ions show very pronounced alternations in their emission intensities I(M(p)C(n-)) versus n with strong maxima for even n whatever M. Since such phenomena are due to the stability properties of the clusters themselves (correspondence rule), it means that the negative ions are the most stable ones for even n. It is thus possible to get the general outlines of their electronic structures from the Pitzer and Clementi model (sp hybridization in Huckel approximation): the clusters are assumed to be linear chains of "cumulene"-type : = C = .. C = C = M and the alternations in the relative stabilities of these chains are due to the fact that the HOMO (highest occupied molecular orbital) of the clusters lies in a double degenerate pi level band. Now HOMO may be either full (or almost full) or half-filled (or nearly half-filled), and an aggregate with a complete (or almost complete) HOMO is more stable than an aggregate with a half-filled HOMO. Consequently, the number of pi electrons is governing the parity effect in the stability alternations. However, this number is depending on the number of sigma electrons of the chain and on the existence of one d-delta level (due to the M atom) which is either empty for deficient d electron elements (Ti, Zr, V, Cr, W) or filled for rich d electron elements (half-filled for Fe or full up for Ni). As the MC(n-) chain must have a full (or nearly full) HOMO if n is even, it is then possible to infer a likely electronic configurations of these clusters, and hence to determine their relative stabilities, and to verify that "even" clusters are more stable than "odd" ones. Thus such Huckel model results can be used as a guide for more sophisticated calculations (ab initio, etc...).