Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:cond-mat/0605149 · Physical Review B

Avoided criticality in near-optimally doped high-temperature superconductors

Kristjan Haule, Gabriel Kotliar

Published Sep 24, 2007• 39 citations
arXivHTMLPDF

Abstract

We study the crossover from the underdoped to the overdoped regime of the $t\text{\ensuremath{-}}J$ model within a plaquette dynamical mean field approach. We find that the shortest electron lifetime occurs near optimal doping where the superconducting critical temperature is maximal. The mean field theory provides a simple physical picture of this effect. In the underdoped regime, the charge carriers propagate coherently among spin singlets, formed by the superexchange interaction. In the overdoped, large carrier concentration regime, the Kondo effect dominates resulting in spin-charge composite quasiparticles which are also coherent. Separating these two Fermi liquid regimes, there is a critical doping where the superexchange and the Kondo interaction balance each other. At this point, the normal phase is highly incoherent and the optical conductivity exhibits power law behavior at intermediate frequencies. The onset of superconductivity restores coherence, causing the appearance of a resonance in the spin channel.

DOI
10.1103/physrevb.76.092503

Related papers

Strongly correlated superconductivity: A plaquette dynamical mean-field theory study

Kristjan Haule, Gabriel Kotliar

We use cluster dynamical mean-field theory to study the simplest models of correlated electrons, the Hubbard model and the $t\text{\ensuremath{-}}J$ model. We use a plaquette embedded in a medium as a reference frame to compute and interpret the physical properties of these models. We study various observables such as electronic lifetimes, one electron spectra, optical conductivities, superconducting stiffness, and the spin response in both the normal and the superconducting state in terms of correlation functions of the embedded cluster. We find that the shortest electron lifetime occurs near optimal doping where the superconducting critical temperature is maximal. A second critical doping connected to the change of topology of the Fermi surface is also identified. The mean-field theory provides a simple physical picture of three doping regimes, the underdoped, the overdoped, and the optimally doped regime, in terms of the physics of the quantum plaquette impurity model. We compare the plaquette dynamical mean-field theory results with earlier resonating valence bond mean-field theories, noting the improved description of the momentum space anisotropy of the normal state properties and the doping dependence of the coefficient of the linear temperature dependence of the superfluid density in the superconducting state.

arXiv:0709.0019•Physical Review B•Sep 13, 2007•193 citations

Antagonistic effects of nearest-neighbor repulsion on the superconducting pairing dynamics in the doped Mott insulator regime

Alexis Reymbaut, M. Charlebois, Marco Fellous-Asiani +4 more

The nearest-neighbor superexchange-mediated mechanism for ${d}_{{x}^{2}\ensuremath{-}{y}^{2}}$ superconductivity in the one-band Hubbard model faces the challenge that nearest-neighbor Coulomb repulsion can be larger than superexchange. To answer this question, we use cellular dynamical mean-field theory (CDMFT) with a continuous-time quantum Monte Carlo solver to determine the superconducting phase diagram as a function of temperature and doping for on-site repulsion $U=9t$ and nearest-neighbor repulsion $V=0,2t,4t$. In the underdoped regime, $V$ increases the CDMFT superconducting transition temperature ${T}_{c}^{d}$ even though it decreases the superconducting order parameter at low temperature for all dopings. However, in the overdoped regime $V$ decreases ${T}_{c}^{d}$. We gain insight into these paradoxical results through a detailed study of the frequency dependence of the anomalous spectral function, extracted at finite temperature via the MaxEntAux method for analytic continuation. A systematic study of dynamical positive and negative contributions to pairing reveals that even though $V$ has a high-frequency depairing contribution, it also has a low frequency pairing contribution since it can reinforce superexchange through $J=4{t}^{2}/(U\ensuremath{-}V)$. Retardation is thus crucial to understanding pairing in doped Mott insulators, as suggested by previous zero-temperature studies. We also comment on the tendency to charge order for large $V$ and on the persistence of $d$-wave superconductivity over extended-$s$ or $s+d$ wave.

arXiv:1607.02517•Physical review. B./Physical review. B•Oct 27, 2016•28 citations

Energetic balance of the superconducting transition across the BCS—Bose Einstein crossover in the attractive Hubbard model

A. Toschi, Massimo Capone, Claudio Castellani

We investigate by means of Dynamical mean-field theory the crossover from BCS superconductivity to Bose-Einstein (BE) condensation of preformed pairs which occurs in the attractive Hubbard model by increasing the attraction strength. Fixing the density to a generic value $n=0.75$, we follow the evolution of the two energy scales underlying the superconducting phenomenon, the gap ${\ensuremath{\Delta}}_{0}$, and the superfluid stiffness ${D}_{S}$, which controls the phase coherence. The BCS--BE crossover is clearly mirrored in a change in the hierarchy of these two scales, the smallest of the two controlling the critical temperature. In the whole intermediate-to-strong-coupling region ${T}_{c}$ scales with ${D}_{S}$, while ${T}_{C}$ is proportional to ${\ensuremath{\Delta}}_{0}$ only in the BCS regime. This evolution as a function of the interaction qualitatively resembles what happens in the cuprates when the doping is decreased towards the Mott insulator. This continuous change reflects also in the energetic balance at the superconducting transition. While, as it is well known, superconductivity is stabilized by a potential energy gain in the BCS regime, the strong-coupling superconductivity is made stable by a reduction of kinetic energy. Interestingly the intermediate-coupling region, where the maximum ${T}_{c}$ is achieved, behaves similarly to the strong-coupling regime, and its gain in kinetic energy is the largest as a function of the coupling. Since the integral of the optical conductivity is proportional to the kinetic energy, the above finding implies that the attractive Hubbard model can account qualitatively for the anomalous behavior of optical spectra around ${T}_{c}$, where an increase of spectral weight is observed in under and optimally doped cuprates, while the overdoped samples have a more standard behavior. This qualitative agreement is lost in the normal phase, specifically at strong coupling, calling for the inclusion of strong correlation effects in the theoretical description.

arXiv:cond-mat/0509188•Physical Review B•Dec 30, 2005•104 citations

Dichotomy in underdoped high $T_c$ superconductors and spinon-dopon approach to $t$-$t'$-$t''$-$J$ model

Ying Ran, Xiao-Gang Wen

We studied underdoped high $T_c$ superconductors using a spinon-dopon approach (or doped-carrier approach) to $t$-$t'$-$t''$-$J$ model, where spinon carries spin and dopon carries both spin and charge. In this approach, the mixing of spinon and dopon describes superconductivity. We found that a nonuniform mixing in $k$-space is most effective in lowering the $t'$ and $t''$ hopping energy. We showed that at mean-field level, the mixing is proportional to quasiparticle spectral weight $Z_-$. We also found a simple monte-carlo algorithm to calculate $Z_{-}$ from the projected spinon-dopon wavefunction, which confirms the mean-field result. Thus the non-uniform mixing caused by $t'$ and $t"$ explains the different electron spectral weights near the nodal and anti-nodal points ({\it i.e.} the dichotomy) observed in underdoped high $T_c$ superconductors. For hole-doped sample, we found that $Z$ is enhanced in the nodal region and suppressed in the anti-nodal region. For electron doped sample, the same approach leads to a suppressed $Z$ in the nodal region and enhanced in the anti-nodal region, in agreement with experimental observations.

arXiv:cond-mat/0611034•Nov 2, 2006•9 citations

Quantum Confinement Transition and Cuprate Criticality

T. Senthil, Matthew P. A. Fisher

Theoretical attempts to explain the origin of high temperature superconductivity are challenged by the complexity of the normal state, which exhibits three regimes with increasing hole doping: a pseudo-gap regime when underdoped, strange power laws near optimal doping and more conventional metallic behavior when heavily overdoped. We suggest that the origin of this behavior is linked to a zero temperature quantum phase transition separating the overdoped Fermi liquid from a spin-charge separated underdoped phase. Central to our analysis is a new $Z_2$ gauge theory formulation, which supports topological vortex excitations - dubbed visons. The visons are gapped in the underdoped phase, splitting the electron's charge and Fermi statistics into two separate excitations. Superconductivity occurs when the resulting charge $e$ boson condenses. The visons are condensed in the overdoped phase, thereby confining the charge and statistics of the electron leading to a Fermi liquid phase. Right at the quantum confinement transition the visons are in a critical state, leading to power law behavior for both charge and spin.

arXiv:cond-mat/9912380•Dec 21, 1999•17 citations

Interplay between superconductivity and flux phase in the t−J model

E. Cappelluti, R. Zeyher

We study the phase diagram of the $t\ensuremath{-}J$ model using a mean field type approximation within the Baym-Kadanoff perturbation expansion for Hubbard X operators. The line separating the normal state from a d-wave flux or bond-order state starts near optimal doping at $T=0$ and rises quickly with decreasing doping. The transition temperature ${T}_{c}$ for d-wave superconductivity increases monotonically in the overdoped region towards optimal doping. Near optimal doping a strong competition between the two d-wave order parameters sets in leading to a strong suppression of ${T}_{c}$ in the underdoped region. Treating for simplicity the flux phase as commensurate the superconducting and flux phases coexist in the underdoped region below ${T}_{c},$ whereas a pure flux phase exists above ${T}_{c}$ with a pseudogap of d-wave symmetry in the excitation spectrum. We also find that incommensurate charge-density-wave ground states due to Coulomb interactions do not modify strongly the above phase diagram near the superconducting phase, at least, as long as the latter exists at all.

arXiv:cond-mat/9808201•Physical review. B, Condensed matter•Mar 1, 1999•102 citations
We study the crossover from the underdoped to the overdoped regime of the t\ensuremath-Jt\text{\ensuremath{-}}Jt\ensuremath-J model within a plaquette dynamical mean field approach. We find that the shortest electron lifetime occurs near optimal doping where the superconducting critical temperature is maximal. The mean field theory provides a simple physical picture of this effect. In the underdoped regime, the charge carriers propagate coherently among spin singlets, formed by the superexchange interaction. In the overdoped, large carrier concentration regime, the Kondo effect dominates resulting in spin-charge composite quasiparticles which are also coherent. Separating these two Fermi liquid regimes, there is a critical doping where the superexchange and the Kondo interaction balance each other. At this point, the normal phase is highly incoherent and the optical conductivity exhibits power law behavior at intermediate frequencies. The onset of superconductivity restores coherence, causing the appearance of a resonance in the spin channel.
We use cluster dynamical mean-field theory to study the simplest models of correlated electrons, the Hubbard model and the $t\text{\ensuremath{-}}J$ model. We use a plaquette embedded in a medium as a reference frame to compute and interpret the physical properties of these models. We study various observables such as electronic lifetimes, one electron spectra, optical conductivities, superconducting stiffness, and the spin response in both the normal and the superconducting state in terms of correlation functions of the embedded cluster. We find that the shortest electron lifetime occurs near optimal doping where the superconducting critical temperature is maximal. A second critical doping connected to the change of topology of the Fermi surface is also identified. The mean-field theory provides a simple physical picture of three doping regimes, the underdoped, the overdoped, and the optimally doped regime, in terms of the physics of the quantum plaquette impurity model. We compare the plaquette dynamical mean-field theory results with earlier resonating valence bond mean-field theories, noting the improved description of the momentum space anisotropy of the normal state properties and the doping dependence of the coefficient of the linear temperature dependence of the superfluid density in the superconducting state.
The nearest-neighbor superexchange-mediated mechanism for ${d}_{{x}^{2}\ensuremath{-}{y}^{2}}$ superconductivity in the one-band Hubbard model faces the challenge that nearest-neighbor Coulomb repulsion can be larger than superexchange. To answer this question, we use cellular dynamical mean-field theory (CDMFT) with a continuous-time quantum Monte Carlo solver to determine the superconducting phase diagram as a function of temperature and doping for on-site repulsion U=9tU=9tU=9t and nearest-neighbor repulsion V=0,2t,4tV=0,2t,4tV=0,2t,4t. In the underdoped regime, VVV increases the CDMFT superconducting transition temperature Tcd{T}_{c}^{d}Tcd​ even though it decreases the superconducting order parameter at low temperature for all dopings. However, in the overdoped regime VVV decreases Tcd{T}_{c}^{d}Tcd​. We gain insight into these paradoxical results through a detailed study of the frequency dependence of the anomalous spectral function, extracted at finite temperature via the MaxEntAux method for analytic continuation. A systematic study of dynamical positive and negative contributions to pairing reveals that even though VVV has a high-frequency depairing contribution, it also has a low frequency pairing contribution since it can reinforce superexchange through $J=4{t}^{2}/(U\ensuremath{-}V)$. Retardation is thus crucial to understanding pairing in doped Mott insulators, as suggested by previous zero-temperature studies. We also comment on the tendency to charge order for large VVV and on the persistence of ddd-wave superconductivity over extended-sss or s+ds+ds+d wave.
We investigate by means of Dynamical mean-field theory the crossover from BCS superconductivity to Bose-Einstein (BE) condensation of preformed pairs which occurs in the attractive Hubbard model by increasing the attraction strength. Fixing the density to a generic value n=0.75n=0.75n=0.75, we follow the evolution of the two energy scales underlying the superconducting phenomenon, the gap ${\ensuremath{\Delta}}_{0}$, and the superfluid stiffness DS{D}_{S}DS​, which controls the phase coherence. The BCS--BE crossover is clearly mirrored in a change in the hierarchy of these two scales, the smallest of the two controlling the critical temperature. In the whole intermediate-to-strong-coupling region Tc{T}_{c}Tc​ scales with DS{D}_{S}DS​, while TC{T}_{C}TC​ is proportional to ${\ensuremath{\Delta}}_{0}$ only in the BCS regime. This evolution as a function of the interaction qualitatively resembles what happens in the cuprates when the doping is decreased towards the Mott insulator. This continuous change reflects also in the energetic balance at the superconducting transition. While, as it is well known, superconductivity is stabilized by a potential energy gain in the BCS regime, the strong-coupling superconductivity is made stable by a reduction of kinetic energy. Interestingly the intermediate-coupling region, where the maximum Tc{T}_{c}Tc​ is achieved, behaves similarly to the strong-coupling regime, and its gain in kinetic energy is the largest as a function of the coupling. Since the integral of the optical conductivity is proportional to the kinetic energy, the above finding implies that the attractive Hubbard model can account qualitatively for the anomalous behavior of optical spectra around Tc{T}_{c}Tc​, where an increase of spectral weight is observed in under and optimally doped cuprates, while the overdoped samples have a more standard behavior. This qualitative agreement is lost in the normal phase, specifically at strong coupling, calling for the inclusion of strong correlation effects in the theoretical description.
Dichotomy in underdoped high TcT_cTc​ superconductors and spinon-dopon approach to ttt-t′t't′-t′′t''t′′-JJJ model
We studied underdoped high TcT_cTc​ superconductors using a spinon-dopon approach (or doped-carrier approach) to ttt-t′t't′-t′′t''t′′-JJJ model, where spinon carries spin and dopon carries both spin and charge. In this approach, the mixing of spinon and dopon describes superconductivity. We found that a nonuniform mixing in kkk-space is most effective in lowering the t′t't′ and t′′t''t′′ hopping energy. We showed that at mean-field level, the mixing is proportional to quasiparticle spectral weight Z−Z_-Z−​. We also found a simple monte-carlo algorithm to calculate Z−Z_{-}Z−​ from the projected spinon-dopon wavefunction, which confirms the mean-field result. Thus the non-uniform mixing caused by t′t't′ and t"t"t" explains the different electron spectral weights near the nodal and anti-nodal points ({\it i.e.} the dichotomy) observed in underdoped high TcT_cTc​ superconductors. For hole-doped sample, we found that ZZZ is enhanced in the nodal region and suppressed in the anti-nodal region. For electron doped sample, the same approach leads to a suppressed ZZZ in the nodal region and enhanced in the anti-nodal region, in agreement with experimental observations.
Theoretical attempts to explain the origin of high temperature superconductivity are challenged by the complexity of the normal state, which exhibits three regimes with increasing hole doping: a pseudo-gap regime when underdoped, strange power laws near optimal doping and more conventional metallic behavior when heavily overdoped. We suggest that the origin of this behavior is linked to a zero temperature quantum phase transition separating the overdoped Fermi liquid from a spin-charge separated underdoped phase. Central to our analysis is a new Z2Z_2Z2​ gauge theory formulation, which supports topological vortex excitations - dubbed visons. The visons are gapped in the underdoped phase, splitting the electron's charge and Fermi statistics into two separate excitations. Superconductivity occurs when the resulting charge eee boson condenses. The visons are condensed in the overdoped phase, thereby confining the charge and statistics of the electron leading to a Fermi liquid phase. Right at the quantum confinement transition the visons are in a critical state, leading to power law behavior for both charge and spin.
We study the phase diagram of the $t\ensuremath{-}J$ model using a mean field type approximation within the Baym-Kadanoff perturbation expansion for Hubbard X operators. The line separating the normal state from a d-wave flux or bond-order state starts near optimal doping at T=0T=0T=0 and rises quickly with decreasing doping. The transition temperature Tc{T}_{c}Tc​ for d-wave superconductivity increases monotonically in the overdoped region towards optimal doping. Near optimal doping a strong competition between the two d-wave order parameters sets in leading to a strong suppression of Tc{T}_{c}Tc​ in the underdoped region. Treating for simplicity the flux phase as commensurate the superconducting and flux phases coexist in the underdoped region below Tc,{T}_{c},Tc​, whereas a pure flux phase exists above Tc{T}_{c}Tc​ with a pseudogap of d-wave symmetry in the excitation spectrum. We also find that incommensurate charge-density-wave ground states due to Coulomb interactions do not modify strongly the above phase diagram near the superconducting phase, at least, as long as the latter exists at all.