Partially filled d bands

The transition metals are also good conductors as they have a similar sp band as the free-electron metals, plus a partially filled d band. The Group IB metals, copper, silver and gold, represent borderline cases, as the d band is filled and located a few eV below the Fermi level. Their sp band, however, ensures that these metals are good conductors. 

The activity of the transition metals, especially for the chemisorption of molecular hydrogen and in hydrogenation reactions has been correlated, in the past, with the existence of partially filled d bands. Many alloy studies were prompted by the expectation that catalytic activity would change abruptly once these vacancies were filled by alloying with a group IB metal. Examples of such behavior have been collected together for the Pd-Au system (1). It is to be expected also that various complications might superimpose on the simple activity patterns observed for primitive... 

The partially filled d band of the transition metals, or the d states in clusters, are described well by the tight binding (TB) approximation11 using a linear combination of atomic d orbitals. The basic concepts of the method are as... 

Although this distinction appears to explain the conductivities of most materials, the fact that ionic compounds containing transition elements are generally insulators or semiconductors, even though treatment of the d electrons by the band approach would call for partially filled d bands, supports the assertion that there is an Rc such that if R > Rc, the band approach is not applicable. 

From the discussion of Chapter I, it follows that metallic conduction is to be associated with partially filled bands of collective-electron states. Since the s-p bands of an ionic compound are either full or empty, metallic conduction implies partially filled d bands, and collective d electrons imply Rtt < Rc(n,d). From the requirement Rtt < Rc(n4) it is apparent that the metallic conduction in ionic compounds must be restricted either to transition element compounds in which the anions are relatively small or to compounds with a cation/anion ratio > 1. Also Rc(n,d) decreases, for a given n, with increasing atomic number, that is with increasing nuclear charge, and the presence of eQ electrons increases the effective size of an octahedral cation (627) (see Fig. 66) and similarly UQ electrons the size of a tetrahedral cation. It follows that If the cation/anion ratio < 1, MO d electrons are more probable in ionic compounds with octahedral-site cations if the cations contain three or less d electrons MO d electrons are more probable in ionic compounds with tetrahedral-site cations if the cations contain two or less d electrons. 

Transition Metal Oxides with Partially Filled d Bands... 

The atom also feels a van der Waals attraction but that is negligible on this scale where the hydrogen atoms are simply falling into a chemisorbed state. Where the crossover occurs determines the size of the barrier for dissociation. Since all metals basically have this feature how do the differences arise The d-electrons have not yet been considered and they are the key to the difference. The d-electrons form a considerably narrower d-band since these orbitals are more localized than the sp-orbitals. Remember the width of the band is proportional to the overlap of the orbitals. The effect of the d-band is illustrated schematically in Figure 4.16 where we now let our hydrogen molecule approach a transition metal with a partially filled d-band, which naturally also has a broad sp-band. 

Fig. 3. A schematic representation of the DOS in APS. (a) A partially filled d-band for transition metals. N E) is the one-electron DOS. The two-electron density of conduction states N2c ) is given by the selfconvolution of N E) above Ep. The derivative of this function is the dotted curve and is characterized by a sharp peak at Ep followed by a negative dip. (b) The d-band is just filled for noble metals. The derivative of N2c(E) is a step-like function (Park 1975).

The leading volume term in this expansion, E oh as well as the two-, three-, and four-ion interatomic potentials, V2, V3, and V4, are volume-dependent, but structure-independent quantities and thus transferable to arbitrary bulk ion configurations. The angular-force multi-ion potentials V3 and V4 reflect directional-bonding contributions from partially filled d bands and are important for mid-period transition metals. In the full GPT, however, these potentials are multidimensional functions, so that V3 and V4 cannot be readily tabulated for application purposes. This has led to the development of a simplified MGPT, which achieves short-ranged, analytic potential forms that can be applied to large-scale atomistic simulations [50]. 

---