From mixed valence to the Kondo lattice regime

TL;DR

This paper investigates the crossover from the Kondo lattice to mixed-valent regimes in heavy fermion materials using the periodic Anderson model and dynamical mean field theory, revealing key changes in dynamics, transport, and optical properties.

Contribution

It provides a detailed theoretical analysis of the valence crossover in heavy fermion systems, highlighting the limitations of simplified models and comparing results with experimental data.

Findings

Kondo resonance broadens as valence decreases

Resistivity shows a two-peak structure in mixed-valent regime

Optical conductivity exhibits a growing mid-infrared shoulder

Abstract

Many heavy fermion materials are known to crossover from the Kondo lattice regime to the mixed-valent regime or vice-versa as a function of pressure or doping. We study this crossover theoretically by employing the periodic Anderson model within the framework of the dynamical mean field theory. Changes occurring in the dynamics and transport across this crossover are highlighted. As the valence is decreased (increased) relative to the Kondo lattice regime, the Kondo resonance broadens significantly, while the lower (upper) Hubbard band moves closer to the Fermi level. The resistivity develops a two peak structure in the mixed valent regime: a low temperature coherence peak and a high temperature 'Hubbard band' peak. These two peaks merge yielding a broad shallow maximum upon decreasing the valence further. The optical conductivity, likewise exhibits an unusual absorption feature (shoulder) in the deep mid-infrared region, which grows in intensity with decreasing valence. The involvement of the Hubbard bands in dc transport, and of the effective f-level in the optical conductivity are shown to be responsible for the anomalous transport properties. A two-band hybridization-gap model, which neglects incoherent effects due to many-body scattering, commonly employed to understand the optical response in these materials is shown to be inadequate, especially in the mixed-valent regime. Comparison of theory with experiment carried out for (a) dc resistivities of CeRhIn5, Ce2Ni3Si5, CeFeGe3 and YbIr2Si2; (b) pressure dependent resistivity of YbInAu2 and CeCu6; and (c) optical conductivity measurements in YbIr2Si2 yields excellent agreement.