DOPING DEPENDENCE OF CORRELATION EFFECTS IN K_1-x Fe_2-y Se₂ SUPERCONDUCTORS: LDA+ DMFT INVESTIGATION
Nekrasov I.A., Pavlov N.S., Sadovskii M.V.

Received: April 23, 2013

DOI: 10.7868/S0044451013110163

DJVU (518.6K)

PDF (2725.9K)

We present a detailed LDA+DMFT investigation of the doping dependence of correlation effects in the novel K_1-x Fe_2-y Se₂ superconductor. Calculations are performed at four different hole doping levels, starting from a hypothetical stoichiometric composition with the total number of electrons equal to 29 per unit cell through 28 and 27.2 electrons toward the case of 26.52, which corresponds to the chemical composition K_0.76 Fe_1.72 Se₂ studied in recent ARPES experiments. In the general case, the increase in hole doping leads to quasiparticle bands in a wide energy window eV around the Fermi level becoming more broadened by lifetime effects, while correlation-induced compression of Fe-3d LDA bandwidths stays almost the same, of the order of 1.3 for all hole concentrations. However, close to the Fermi level, the situation is more complicated. In the energy interval from -1.0 eV to 0.4 eV, the bare Fe-3d LDA bands are compressed by significantly larger renormalization factors up to 5 with increased hole doping, while the value of Coulomb interaction remains the same. This fact manifests the increase in correlation effects with hole doping in the K_1-x Fe_2-y Se₂ system. Moreover, in contrast to typical pnictides, K_1-x Fe_2-y Se₂ does not have well-defined quasiparticle bands on the Fermi levels, but has a ``pseudogap''-like dark region instead. We also find that with the growth of hole doping Fe-3d orbitals of various symmetries are affected by correlations differently in different parts of the Brillouin zone. To illustrate this, we determine the quasiparticle mass renormalization factors and energy shifts that transform the bare Fe-3d LDA bands of various symmetries into LDA+DMFT quasiparticle bands. These renormalization factors effectively mimic more complicated energy-dependent self-energy effects and can be used to analyze the available ARPES data.