Pseudogap transition within the superconducting phase in the three-band Hubbard model

Abstract

The onset of the pseudogap in high-$T_c$ superconducting cuprates (HTSC) is marked by the $T^*$ line in the doping-temperature phase diagram, which ends at a point $p^*$ at zero temperature within the superconducting dome. Although various theoretical and experimental studies indicate a competition between the pseudogap and superconductivity, there is no general consensus on the effects of the pseudogap within the superconducting phase. We use cluster dynamical mean field theory on a three-band Hubbard model for the HTSC to study the superconducting phase at $T=0$, obtained when doping the charge-transfer insulator, for several values of $U$. We observe a first-order transition within the superconducting phase, which separates the underdoped and overdoped solutions. The transition to the underdoped solution is marked by a discontinuous increase in the spectral gap, and on further underdoping the spectral gap increases while the superconducting order parameter decreases. We conclude that this is due to the onset of the pseudogap in the underdoped region, which contributes to the increasing spectral gap; this is further consistent with the appearance of a pole in the normal component of the self-energy, in the antinodal region, in the underdoped solution. This is accompanied by a change in the source of the condensation energy from potential energy, in the overdoped region, to kinetic energy in the underdoped region. Further, we also observe that the $d$-wave node vanishes smoothly within the superconducting phase at low values of hole doping, within the underdoped region. We see this as a manifestation of Mott physics operating at very low doping. Various aspects of the results and their implications are discussed.