On the nature of antiferromagnetism in the CuO <Emphasis Type="Bold"/> planes of oxide superconductors

Recent results regarding electron- and hole-doped CuO₂ planes can be rather easily explained by the marked covalency of CuO bonding, suggesting a band picture of long and short range antiferromagnetism. The maxima of superconductive T_cversus doping can be related to the crossing by the Fermi level of the edges of the pseudogap due to antiferromagnetic short range order (bonding edge for hole-doping, antibonding for electron-doping). The symmetry of the superconductive gap can be related to the Bragg scattering of electronic Bloch states near the edges of the antiferromagnetism (AF) pseudogap. Assuming a standard phonon coupling, one then predicts for commensurate AF a pure d_{x2 - y2} symmetry for the superconductive gap for underdoped samples and d_{x2 - y2} symmetry plus an imaginary contribution of s or d_xy symmetry contribution increasing linearly with overdoping. This seems in agreement with recent measurements of gap symmetry for YBCO, but should be more fully tested, especially for electron-doped samples. Incommensurate AF, as in LSCO, is not considered here. The simple Hartree-Fock band approximation used could no doubt be made more realistic by specific inclusion of electron correlations and by a better description of the AF short range order. This weak atomic repulsion U model, standard in transitional metals, is complementary to the strong U models usually assumed in oxides. It considers specifically the possible effects, in doped samples, of a short range AF which could be slowly dynamical or static, possibly including in that case recent evidences of nanostructures of two different phases.