Electronic Structure Studies of Cuprate Superconductors

Abstract

Since the discovery of the cuprate superconductors ¹ three years have gone and still the mechanism for superconductivity in this new class of materials is unclear. Strongly linked to this problem is the knowledge of the electronic structure of these materials, mainly at energies near the Fermi level. In this field a lot of experimental and theoretical work has been done bringing some insight into the problem. It seems to be established now that (maybe excluding the so-called electron conducting cuprates) superconductivity is connected with hole states on oxygen most probably in the CuO₂ planes or ribbons. But there is a strong debate whether the cuprates can be described within a Fermi liquid theory or whether due to the strong onsite Coulomb interaction (U~8 eV) of the Cu d electrons a more local description is required. Band-structure calculations in the local density approximation (LDA)² cannot describe the electronic structure of La_2−xSr_xCuO₄: The band-structure calculation predicts a half-filled band and hence a metal, even in the case of the undoped La₂CuO₄ which is an antiferromagnetic insulator. XPS investigations ³ showed that the density-of-states (DOS) of the valence band is at lower energy by about 2 eV compared to LDA band-structure calculations. On the other hand, band-structure calculations predict rather well the topography of the Fermi surface as found by angle resolved photoemission (UPS) investigations on Bi₂Sr₂CaCu₂O₈ ⁴. While the tools for band-structure calculations are rather elaborated giving excellent agreement with experimental results in “normal” cases it is not yet possible to calculate the electronic structure for the highly correlated systems such as La₂CuO₄. For the discussion of our results we may use a “band” picture ⁵ (Fig. 1) in which, due to the Coulomb interaction, the copper-oxygen band splits into two Cu3d Hubbard bands with an oxygen 2p band in between. In the case of the undoped La₂CuO₄ or YBa₂Cu₃O₆ the oxygen band is filled. We will argue in this paper that upon doping holes are formed in this band. Another model for the electronic structure in the doped case may be a description where impurity states are created upon doping at or near the Fermi level in between the filled oxygen band and the upper Hubbard band.