Ab Initio Cluster Studies of La2CuO4

Abstract

The discovery of the first high temperature cuprate superconductor by Bednorz and Muller1 has spurred enormous experimental and theoretical activity aimed at determining the “unique” aspects of the electronic structure of these materials2. The straightforward application of local-density-functional (LDF) band theory3 predicts the parent compound La2CuO4 to be a metal, when it is in fact an anti-ferromagnetic insulator with a gap of ~2eV. The local Coulomb interactions which drive the superexchange interaction4 are certainly at the heart of the anti-ferromagnetism and so one goal of electronic structure theory has been to develop simple models which capture these effects and can then be extended to the infinite system. At one extreme in the first principles approaches to this problem are the constrained LDF methods5,6. They essentially carve out a local region of space within the LDF band structure in order to generate the parameters for a tight-binding model including the appropriate Coulomb interactions. The latter are determined by monitoring the total energy of the system as a function of the charge constrained to reside within the local region. At the other extreme lie first principles cluster approaches7,8,9 whose essential philosophy is that the parameters characterizing a small cluster should be transferrable to the solid and largely determine its properties. Although the local interactions can be treated with great sophistication in this approach, approximations must be made concerning the treatment of the background used to imbed the cluster. Most efforts utilize a point-charge background for these materials; an extremely different environment from the metallic background of the constrained LDF approaches. The “truth” presumably lies somewhere between these two extremes.