For example, look at the co-ordination of copper in one of the new ceramic oxide superconductor ¶YBa2Cu3O7 (YBCO). The copper (Cu) are the green atoms, the oxygen are as usual red and barium is blue. Clearly there are two kinds of copper atom - those that are co-ordinated by 4 oxygen atoms (green squares), typical of divalent Cu++, and those that have a fifth oxygen atom (green pyramids). This material has zero electrical resistance even above the temperature of liquid air - cold but easy to produce and handle. This is truly amazing, and a few years ago would have been thought impossible.
If we heat this remarkable superconductor in the absence of oxygen it loses one of its oxygen atoms and becomes the insulator ¶YBa2Cu3O6 with a very similar structure. The oxygen is lost from one particular site; the chains of CuO4 squares. Copper in these squares is left with only two oxygen atoms, typical of monovalent Cu+. Copper is said to have been "reduced" from Cu++ to Cu+. Oxygen and superconductivity can be restored by "oxidising" the copper again from Cu+ to Cu++. This solid state chemistry is clearly responsible for the unusual electrical properties.
Yet it is not the CuO4 chains that are responsible for the superconductivity in YBCO. Many similar materials, which conduct at even higher temperatures, can be made by replacing these chains by layers of other materials, such as heavy metal oxides. Neutron diffraction from oxide superconductors indicated that oxidation of these ¶charge reservoir layers results in the formal oxidation of the planes of copper oxide pyramids (Cu++ to Cu+++), due to "charge transfer". This empirical understanding of solid state chemistry in YBCO, which resulted from the ILL Grenoble's most cited paper, lead directly to the discovery of many other similar superconducting materials.
The highest superconducting temperature (Tc) so far obtained is for a material, where the charge reservoir consists of ¶mercury oxide; here Tc is 50% higher than in YBCO ! The mercury oxide (HgO) layers are drawn as a yellow "rock salt" type structure, but actually the charge reservoir structure is much more complex, being typically not quite "commensurate" with the copper oxide layers; the so-called "lone-pair electrons" on mercury further complicate the real structure. By understanding this subtle crystal chemistry we hope to make materials with even higher Tc's.
These superconductor structures may appear complex, but in fact we can easily understand how they are related to the simple perovskite structure. The formula YBa2Cu3O7 may be considered as ¶(YBa2)Cu3O9 or 3 units of perovskite ¶3x(A.B.X3) with 2 of the oxygen atoms removed. Oxygen atoms must be removed to preserve charge/valence balance for the formula Y+++1Ba++2Cu+++1 Cu++2O--7. When further oxygen is removed the Cu+++ is reduced to Cu++ in the non-superconducting material, and the formula becomes Y+++1Ba++2 Cu++2Cu+1O--6
The original high temperature superconductor ¶La2CuO4, for which Bednorz and Muller received the Nobel prize, can be recognised immediately as a member of the great perovskite family. If you are particularly interested in these incredible materials, you can generate your own 3D VRML structures of oxide superconductors from the ILL's Superconductor page.
We believe that superconductivity is due to an interaction between electrons, the details of
which are not yet fully understood. Interactions between magnetic moments, which
are produced by the movement of electrons, result in other fascinating and useful materials -
ceramic oxide magnets and giant magneto-resistive materials.