Effects of doping in Kondo insulators (invited)

Publisher Website
Google Scholar
Abstract
Kondo insulators like Ce3Bi4Pt3 and CeNiSn are compounds with small-gap semiconductor properties. Nonmagnetic impurities, so-called Kondo holes, break the translational invariance and hence the coherence of the ground state. Impurity states can be introduced by (i) substituting the rare earth (actinide) ion or (ii) by replacing (or adding, removing) one of the ligand atoms. Isolated impurities usually give rise to bound states in the gap. Depending on the nature of the impurity (charge neutral or a dopand) the Fermi level is pinned by the impurity level or lies in the gap. In the former case the Kondo hole has magnetic properties (Curie susceptibility and Schottky anomaly in the specific heat), while in the latter situation the properties are nonmagnetic. For a finite concentration of Kondo holes the situations (i) and (ii) are qualitatively different. In (i) it gives rise to an impurity band inside the gap of the semiconductor. The height and width of the impurity band in the f-electron density of states are proportional to c1/2 for small concentrations. If the impurities are charge neutral the Fermi level lies in the impurity band giving rise to a specific heat proportional to T and a Pauli-like susceptibility. If the impurities dope the bands the properties remain semiconducting with a strongly reduced gap. For ligand impurities [case (ii)] tails of impurity states develop close to the gap edges, suppressing in this way the gap. The system remains a semiconductor if the impurities are charge neutral, but C∝T and χ is finite if they are dopands.

This publication has 22 references indexed in Scilit: