Theory of pseudogap in underdoped cuprates

Recent theoretical studies on the origin of the pseudogap emerging in underdoped cuprate superconductors are overviewed, based on insights obtained by a cluster extension of the dynamical mean field theory (cDMFT) for the doped two-dimensional (2D) Mott insulator. The pseudogap obtained in the cDMFT shows an s-wave-like full-gap structure, distinct from the d-wave superconducting gap. The zero-temperature electronic structure supports that a non-Fermi liquid phase exists and underlies the pseudogap. The non-Fermi liquid phase is separated from the larger-doped Fermi liquid by topological transitions of the Fermi surface and an emergence of zeros of Green's function. A coexisting evolution of the poles (Fermi surface) and zeros of the Green's function is a unique feature of the pseudogap phase. The spectra well reproduce the arc/pocket formation, together with basic experimental properties of the pseudogap phase in the cuprates. Furthermore, a full-gap structure is supported by a comparison with the results of Raman experiments. The overall feature supports the proximity of the Mott insulator and the significance of the quantum criticality of the Mott transition. These numerical results are further favorably interpreted by extending the exciton concept, known in semiconductors, to doped Mott insulators. In this composite fermion (CF) theory, the pseudogap emerges as a gap arising from a hybridization of the quasiparticle (QP) with the CF. The pairing channel opening between a QP and a CF solves the puzzle of the dichotomy between the d-wave superconductivity and the precursory insulating gap in the same antinodal region. A mechanism of superconductivity emerges from this pairing.