Superconductivity by local pairs in the metal oxide semiconductors

Theoretical arguments and experimental evidence are presented in favour of a general model to explain the overall properties of the various types of metal oxide superconductors. These materials are polar semi-conductors, with intrinsic gaps of a few eV. Doping or self-doping leads to p-or n-type extrinsic conductivity in, respectively, an acceptor or a donor impurity band inside the gap. The metal-nonmetal transition at low doping levels is of the Mott-Anderson type. The gap-states are charged molecular defects that can be seen as mixed-valence small-polarons. The defects may carry a single or double charge, depending on whether the metal atom involved has a tendency to disproportionate or not. The mobility of these polarons and bipolarons is governed by the principles of mixed-valence (polaronic) charge transfer. The moving bipolarons can be viewed as weakly interacting, charged Bose quasiparticles hopping on a lattice, and will Bose-condense into a superfluid state at sufficiently low temperatures. The polarons considered here are basically dielectric polarons, in view of the very strong electron- lattice interactions that are to be expected when extra carriers are introduced into the highly polar metal-oxide lattices. However, their effective mass may still be reasonably small (10 - 100 me), since these (point) defects can be viewed as nonlinear (solitonic) excitations. In the copper-oxides the dielectric polarons are spinless defects, and thus they are at the same time also spin-polarons, since the presence of a spinless defect in an antiferromagnetic background will distort the antiferromagnetic arrangement of the nearest-neighbour spins. 161 refs