Concealed Mott Criticality: Unifying the Kondo Breakdown and Doped Charge-Transfer Insulators

TL;DR

This work proposes that the quantum critical points observed in heavy fermions and doped cuprates arise from concealed Mott criticality, where one electron species undergoes Mott localization in the presence of metallic carriers. By introducing a two-band model with conduction ($c$) and flat ($f$) electrons and a hybridization $t_ot$, the paper demonstrates a Fermi-surface jump and diverging effective mass as the transition is tuned, with the critical behavior inherited from the underlying Mott QCP of the $f$ sector. The analysis emphasizes a charge-first perspective, arguing that spin degrees of freedom play a secondary role in driving the observed critical features. The results provide a unifying framework for understanding Kondo breakdown and QCPs in cuprates, and motivate exploring beyond-DMFT treatments and moiré systems to further elucidate concealed Mott criticality and its connection to strange metal behavior and high-temperature superconductivity.

Abstract

I show that the quantum critical points observed in heavy fermions (the `Kondo breakdown') and in doped cuprates can be understood in terms of concealed Mott criticality. In this picture, one species of electrons undergoes a Mott localization transition, in the presence of metallic charge carries. As is shown in a simple toy model, this results in a Fermi surface jump at the transition, as well as mass enhancement on both the `large' and `small' Fermi surface side of the transition, consistent with the experimental observations.

Figures (4)

• Figure 1: The simplified picture of regular Mott criticality. a. Regular Mott criticality describes the continuous zero-temperature phase transition from a Fermi liquid to a Mott insulator in a single band system. On the Fermi liquid side, the quasiparticle weight $Z$ continuously goes to zero as a function of the tuning parameter $\delta$. On the Mott insulator side, the Mott gap $\Delta$ goes to zero as we approach the transition. b. The effective quasiparticle dispersion is 'flattened' due to the mass enhancement $m^*/m = 1/Z$ as we approach the Mott critical point from the Fermi liquid side. c. On the Mott insulator side, the pole in the self-energy (Eq. \ref{['Eq:MottSelfEnergy']}) causes a split of the bare dispersion (gray dashed line) into an Upper and Lower Hubbard band.
• Figure 2: The Fermi surfaces in the different phases of our model. a. In the absence of interactions and the inter-orbital hopping $t_\perp$, there is a half-filled Fermi surface associated with the $f$ electrons and a small Fermi surface for the $c$ electrons. b. If the $f$ electrons are in the Fermi liquid regime, the resulting bandstructure has a single large Fermi surface with a volume given by electron density $n = 1+n_c'$. c. When the $f$ electrons localize, only the small Fermi surface remains with density given by the $c$-electron density.
• Figure 3: In the regime with a large Fermi surface, the hybridized band structure given by Eq. \ref{['Eq:LargeFSBS']} inherits the mass enhancement of the $f$ electrons. This figure shows the dispersions for $Z =0.2$ to $Z=1$ in steps of $0.2$.
• Figure 4: When there is a small Fermi surface due to the Mott localization of the $f$ electrons, the small Fermi surface exhibits a mass enhancement proportional to the inverse of the Mott gap $\Delta$. This figure shows the dispersions for a range of gap values of the form $\Delta = 2^n$, from $n=2$ to 64. In the limit of an infinite Mott gap, the $c$ electron dispersion is regained. The thin lines indicate the lower and upper Hubbard bands associated with the $f$ electrons.