Fermi arcs and pseudogap emerging from dimensional crossover at the Fermi surface in La _2−x Sr _x CuO ₄

Abstract

The doping mechanism and realistic Fermi surface (FS) evolution of La2−xSrxCuO4 (LSCO) are modelled within an extensive ab initio framework including advanced band-unfolding techniques. We show that ordinary Kohn-Sham DFT+U can reproduce the observed metalinsulator transition, when not restricted to the paramagnetic solution space. Arcs are self-doped by orbital charge transfer within the Cu-O planes, while the introduced Sr charge is strongly localized. Arc protection and the inadequacy of the rigid-band picture are consequences of a rapid change in orbital symmetry at the Fermi energy: the material undergoes a dimensional crossover along the Fermi surface, between the nodal (2D) and antinodal (3D) regions. In LSCO, this crossover accounts for FS arcs, the antinodal pseudogap, and insulating behavior in c-axis conductivity, all ubiquitous phenomena in high-Tc cuprates. Ligand Coulomb integrals involving out-of-plane sites are principally responsible for the most striking effects observed by ARPES in LSCO.

High-Temperature Superconductors as Ionic Metals

“…This interpretation of "arc protection" [31] is verified analytically [24] within the four-band model mentioned above. The secular polynomial of this model factorizes along the zone diagonal (k x = k y ), so that the dispersion along the zone diagonal is uncoupled from the effective Cu 4s orbital which is the physical conduit of out-of-plane effects on the plane.…”

“…As it stands, the correction penalizes departures from the idempotent occupation numbers, zero and one, in the d orbitals only, thus pushing the Cu orbitals towards the ionic limit. With a value U = 4 eV, known to give the correct formation energy of LSCO, the measured optical gap of 1.8 eV was obtained at half-filling, a useful reality check on the calculation [24]. In the context of Hund's rule, the success of the +U correction means that the Cu orbital has compensated the strong intraatomic correlation radially, by transferring charge to the ligand oxygen, which is itself not strongly correlated.…”

“…(1), whether it can be derived adiabatically or not. It will be pointed out below that the Fermi arcs [23] have in fact been obtained in a one-body microscopic model, once the out-of-plane Coulomb integrals were properly taken into account [24].…”

“…The small Brillouin zone pertaining to the large dopand unit cell is not observed because of dopand disorder, which was not included in the calculation [24]. This is just the disordered-alloy situation for which the bandunfolding algorithm [29] was originally developed.…”

Fundamental Building Blocks of Strongly Correlated Wave Functions

“…a fascinating mixture of ionicity and metallicity [1,2] which remains to be unravelled. New tools and approaches are constantly being sought [3] for the description of electrons which inhabit active (open) orbitals in these materials, for which the paradigm "strongly correlated electrons" has been coined long ago.…”