Superconductivity narrow-band systems

A striking similarity between the charge transfer salts and oxides is the importance of electron correlations,and hence the proximity of superconducting and magnetic states in the phase diagrams, as a function of band filling or structural modifications. This arises from the fact that both types of compound are narrow-band systems. 

At this point we want to stress that the heavy fermion systems (22) are related to a very large spin enhanced susceptibility or localized magnetic moments, very narrow bands of elementary excitations at the Fermi level, and a new type of pairing, in the case where they become superconducting, at low temperatures even if they were spin fluctuators above the critical superconducting transition temperature Tg. 

The alkali metals act here as donors, which make a metal of the semiconductor Ceo as acceptor, with an energy band gap of ca. 2.3 eV via half-filling of the conduction band. The system behaves similarly to the radical-anion salts which we have already treated (Sects. 9.2 and 9.3). Ceo is a good electron acceptor. It consists entirely of carbon atoms and is thus not actually an organic molecule. Superconducting Fullerene salts however have properties which are like those of the organic molecular salts. These are above all the important role played by the Jt electrons in charge transport, and the existence of relatively narrow bands with a low electron density. 

Ceo fullerides exhibit several interesting phenomena related with the presence of strong correlations such as high temperature superconductivity or antiferromagnetism. The existence of non-conventional behaviors can be anticipated from the fact that both the electron-phonon and electron-electron interactions are, respectively, comparable and much larger than the narrow bandwidth predicted by standard electronic structure calculations (a few hundreds of meV). These systems are therefore close to a metal-insula-tor transition of the Mott-Hubbard type and the validity of the adiabatic approximation, assumed in most electronic structure calculations, can be questioned. From the experimental point of view the band derived from the molecular LUMO is usually seen as a quite broad feature, much wider that the theoretical estimates, in photoemission experiments. However, the observation of the band dispersion has proven elusive in ARPES studies until very recently. In a recent joint experimental and theoretical paper, Yang et al. [158] have reported the first photoemission measurement of the band dispersion for a K-doped Qo monolayer deposited on a Ag(lll) substrate. The results have been compared with ab initio calculations performed with SIESTA. In those calculations the Ag substrate was modeled by a slab con-... 

---