Antiferromagnetic metal to heavy-fermion metal quantum phase transition in the Kondo lattice model: A strong-coupling approach

摘要

We study the quantum phase transition from an antiferromagnetic metal to a heavy fermion metal in the Kondo lattice model. Based on the strong-coupling approach we first diagonalize the Kondo coupling term. Since this strong-coupling approach makes the resulting Kondo term relevant, the Kondo hybridization persists even in the antiferromagnetic metal, indicating that fluctuations of Kondo singlets are not critical in the phase transition. We find that the quantum transition in our strong coupling approach results from softening of antiferromagnetic spin fluctuations of localized spins, driven by the Kondo interaction. Thus, the volume change of the Fermi surface becomes continuous across the transition. Using the boson representation of the localized spin n_i = 1/2z_(iσ)τ_(σσ')z~(iσ') with the spin-fractionalized excitation z_(iσ), we derive an effective U(1) gauge Lagrangian in terms of renormalized conduction electrons and fractionalized local-spin excitations interacting via U(1) gauge fluctuations, where the renormalized conduction electrons are given by composites of the conduction electrons and fractionalized spin excitations. Performing a mean field analysis based on this effective Kondo action, we find a mean field phase diagram as a function of J_K/D with various densities of conduction electrons, where J_K is the Kondo coupling strength and D the half-bandwidth of conduction electrons. The phase diagram shows a quantum transition, resulting from condensation of the spin-fractionalized bosons, from an antiferromagnetic metal to a heavy fermion metal away from half-filling. We show that beyond the mean field approximation our critical field theory characterized by the dynamic critical exponent z =2 can explain the observed non-Fermi liquid physics such as the specific heat coefficient γ ≡ C_υ/T～-ln T near the quantum critical point. Furthermore, we argue that if our scenario is applicable, there can exist a narrow region of an anomalous metallic phase with the spin gap near the quantum critical point.