Can the structure of the Ti or V Magneli binary oxides host superconductivity?

A strategy for tailoring new superconducting systems is based on band structure calculations and the use of the empirical criteria suggested by L.F. Mattheiss. These criteria allow one to evaluate the capability of new materials to host superconductivity. Promising host structures might be found among those materials undergoing metal-insulator transitions as a function of temperature. Materials with metallic ground states can then be generated by substitution. In this article, we suggest that simple binary compounds, such as titanium or vanadium oxides belonging to the homologous series MnO2n-1 may be good candidates. They exhibit structures which can be derived from rutile by removing one oxygen atom every n M cation via a shearing mechanism, which besides creating n-octahedron-thick rutile blocks along the pseudo tetragonal c-axis, generates a non-stoichiometry with respect to the MO2 formula. Because of the mixed valence state, these materials generally exhibit metallic conductivity and undergo metal-to-insulator transitions as a function of temperature. For example, Ti4O7 undergoes a transition at 150 K, at which temperature a drop of conductivity by three orders of magnitude and one of magnetic susceptibility by a factor of three are observed. On further cooling, a second drop of conductivity by two orders of magnitude is observed at 130 K when the compound undergoes a semiconductor-to-semiconductor transition. The V counterpart seems to be a somewhat better candidate than Ti4O7. The behavior of V4O7 as function of temperature has some similarity to that of undoped La2CuO4 . The vanadium Magneli phase with 50% V3+ and 50% V3+, has a metallic conductivity above 250 K, becomes a semiconductor below this temperature and orders antiferromagnetically at 40 K. (C) 2000 Elsevier Science B.V. All rights reserved.