Empirical case for two pseudogaps in cuprate superconductors

Abstract

Superconductivity in cuprates is achieved by doping holes into a correlated charge-transfer insulator. While the correlated character of the parent insulator is now understood, there is no accepted theory for the "normal" state of the doped insulator. I present a mostly empirical analysis of a large range of experimental characterizations, making the case for two pseudogaps: (1) a large pseudogap resulting from the competition between the energy of superexchange-coupled local Cu moments and the kinetic energy of doped holes; (2) a small pseudogap that results from dopant disorder and consequent variations in local charge density, leading to a distribution of local superconducting onset temperatures. The large pseudogap closes as hole kinetic energy dominates at higher doping and the dynamic antiferromagnetic correlations become overdamped. Establishing spatially-homogeneous $d$-wave superconductivity is limited by those regions with the weakest superconducting phase coherence, which tends to be limited by low-energy spin fluctuations. The magnitude of the small pseudogap is correlated with the doping-dependent energy $E_{\rm cross}$ associated with the neck of the hour-glass dispersion of spin excitations. The consequences of this picture are discussed.