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:1207.0185 · Physical Review B

Universality class of the Mott transition in two dimensions

S. Moukouri, Eitan Eidelstein

Published Oct 8, 2012• 16 citations
arXivHTMLPDF

Abstract

We use the two-step density-matrix renormalization group method to elucidate the long-standing issue of the universality class of the Mott transition in the Hubbard model in two dimensions. We studied a spatially anisotropic two-dimensional Hubbard model with a nonperfectly nested Fermi surface at half-filling. We find that unlike the pure one-dimensional case where there is no metallic phase, the quasi-one-dimensional model displays a genuine metal-insulator transition at a finite value of the interaction. The critical exponent of the correlation length is found to be $\ensuremath{\nu}\ensuremath{\approx}1.0$. This implies that the fermionic Mott transition belongs to the universality class of the 2D Ising model.

DOI
10.1103/physrevb.86.155112

Related papers

Mott transition in quasi-one-dimensional systems

S. Moukouri, Eitan Eidelstein

We report the application of the density-matrix renormalization-group method to a spatially anisotropic two-dimensional Hubbard model at half filling. We find a deconfinement transition induced by the transverse hopping parameter ${t}_{y}$ from an insulator to a metal. Therefore, if ${t}_{y}$ is fixed in the metallic phase, increasing the interaction $U$ leads to a metal-to-insulator transition at a finite critical $U$. This is in contrast to the weak-coupling Hartree-Fock theory which predicts a nesting-induced antiferromagnetic insulator for any $U>0$.

arXiv:1105.0150•Physical Review B•Nov 16, 2011•12 citations

Mott transition and dimerization in the one-dimensional SU(n) Hubbard model

K. Buchta, Örs Legeza, E. Szirmai +1 more

The one-dimensional $\mathrm{SU}(n)$ Hubbard model is investigated numerically for $n=2$, 3, 4, and 5 at half-filling and $1∕n$ filling using the density-matrix renormalization-group method. The energy gaps and various quantum information entropies are calculated. In the half-filled case, finite spin and charge gaps are found for arbitrary positive $U$ if $n>2$. Furthermore, it is shown that the transition to the gapped phase at ${U}_{c}=0$ is of Kosterlitz-Thouless type and is accompanied by a bond dimerization both for even and odd $n$. In the $1∕n$-filled case, the transition has similar features as the metal-insulator transition in the half-filled SU(2) Hubbard model. The charge gap opens exponentially slowly for $U>{U}_{c}=0$, the spin sector remains gapless, and the ground state is nondimerized.

arXiv:cond-mat/0607374•Physical Review B•Apr 19, 2007•48 citations

The Mott-Anderson transition in the disordered one-dimensional Hubbard model

Ramesh V. Pai, Alexander Punnoose, Rudolf A. Römer

We use the density matrix renormalization group to study the quantum transitions that occur in the half-filled one-dimensional fermionic Hubbard model with onsite potential disorder. We find a transition from the gapped Mott phase with algebraic spin correlations to a gapless spin-disordered phase beyond a critical strength of the disorder $Δ_c \approx U/2$. Both the transitions in the charge and spin sectors are shown to be coincident. We also establish the finite-size corrections to the charge gap and the spin-spin correlation length in the presence of disorder and using a finite-size-scaling analysis we obtain the zero temperature phase diagram of the various quantum phase transitions that occur in the disorder-interaction plane.

arXiv:cond-mat/9704027•Apr 3, 1997•1 citation

Tracing the Mott-Hubbard transition in one-dimensional Hubbard models without Umklapp scattering

Florian Gebhard, Örs Legeza

We apply the density-matrix renormalization group (DMRG) method to a one-dimensional Hubbard model that lacks Umklapp scattering and thus provides an ideal case to study the Mott-Hubbard transition analytically and numerically. The model has a linear dispersion and displays a metal-to-insulator transition when the Hubbard interaction $U$ equals the bandwidth ${U}_{\mathrm{c}}=W$, where the single-particle gap opens linearly, $\mathrm{\ensuremath{\Delta}}(U\ensuremath{\ge}W)=U\ensuremath{-}W$. The simple nature of the elementary excitations permits us to determine numerically with high accuracy the critical interaction strength and the gap function in the thermodynamic limit. The jump discontinuity of the momentum distribution ${n}_{k}$ at the Fermi wave number ${k}_{\mathrm{F}}=0$ cannot be used to locate accurately ${U}_{\mathrm{c}}$ from finite-size systems. However, the slope of ${n}_{k}$ at the band edges ${k}_{\mathrm{B}}=\ifmmode\pm\else\textpm\fi{}\ensuremath{\pi}$, reveals the formation of a single-particle bound state which can be used to determine ${U}_{\mathrm{c}}$ reliably from ${n}_{k}$ using accurate finite-size data.

arXiv:2109.01423•Physical review. B./Physical review. B•Dec 13, 2021•6 citations

Correlation driven Mott-insulator-to-metal transition

Rubem Mondaini, Thereza Paiva

We study transport properties of the half-filled two-dimensional (2D) Hubbard model with spatially varying interactions, where a pattern of interacting and non-interacting sites is formed. We use Determinantal Quantum Monte Carlo method to calculate the double occupation, effective hopping and Drude weight. These data point to two phase transitions, driven by fermionic correlations. The first is the expected metal to a Mott insulating state. The second one, is an exotic transition from a Mott insulating state to a highly anisotropic metal, that takes place at large values of the fermion-fermion interaction. This second transition occurs when the layers formed by the spatially varying interactions decouple due to the suppression of the hopping between interacting and non-interacting sites, leading to fermionic transport along the non-interacting ones.

arXiv:1303.7369•Mar 29, 2013•0 citations

Phase Diagram of the Two-Dimensional Anderson-Hubbard Model

Jaichul Yi Lizeng Zhang, Geoffrey Canright

We study the ground state of the two-dimensional Anderson-Hubbard model using a quantum real space renormalization group method. We obtain the phase diagram near half filling. The system is always insulating with disorder. At half filling, the system undergoes a transition from a gapless (Anderson) insulator to an incompressible (Mott-Hubbard) insulator as the interaction $U$ reaches a critical value $U_c$. Away from half filling, the insulating phase is always gapless and is found to be controlled by the Anderson fixed point at half filling. This result is similar to that obtained in the corresponding one dimensional system and suggests strongly the importance of the electron-electron correlation in this gapless insulating phase.

arXiv:cond-mat/9403002•Mar 1, 1994•0 citations
We use the two-step density-matrix renormalization group method to elucidate the long-standing issue of the universality class of the Mott transition in the Hubbard model in two dimensions. We studied a spatially anisotropic two-dimensional Hubbard model with a nonperfectly nested Fermi surface at half-filling. We find that unlike the pure one-dimensional case where there is no metallic phase, the quasi-one-dimensional model displays a genuine metal-insulator transition at a finite value of the interaction. The critical exponent of the correlation length is found to be \ensuremathν\ensuremath≈1.0\ensuremath{\nu}\ensuremath{\approx}1.0\ensuremathν\ensuremath≈1.0. This implies that the fermionic Mott transition belongs to the universality class of the 2D Ising model.
We report the application of the density-matrix renormalization-group method to a spatially anisotropic two-dimensional Hubbard model at half filling. We find a deconfinement transition induced by the transverse hopping parameter ty{t}_{y}ty​ from an insulator to a metal. Therefore, if ty{t}_{y}ty​ is fixed in the metallic phase, increasing the interaction UUU leads to a metal-to-insulator transition at a finite critical UUU. This is in contrast to the weak-coupling Hartree-Fock theory which predicts a nesting-induced antiferromagnetic insulator for any U>0U>0U>0.
The one-dimensional SU(n)\mathrm{SU}(n)SU(n) Hubbard model is investigated numerically for n=2n=2n=2, 3, 4, and 5 at half-filling and 1∕n1∕n1∕n filling using the density-matrix renormalization-group method. The energy gaps and various quantum information entropies are calculated. In the half-filled case, finite spin and charge gaps are found for arbitrary positive UUU if n>2n>2n>2. Furthermore, it is shown that the transition to the gapped phase at Uc=0{U}_{c}=0Uc​=0 is of Kosterlitz-Thouless type and is accompanied by a bond dimerization both for even and odd nnn. In the 1∕n1∕n1∕n-filled case, the transition has similar features as the metal-insulator transition in the half-filled SU(2) Hubbard model. The charge gap opens exponentially slowly for U>Uc=0U>{U}_{c}=0U>Uc​=0, the spin sector remains gapless, and the ground state is nondimerized.
We use the density matrix renormalization group to study the quantum transitions that occur in the half-filled one-dimensional fermionic Hubbard model with onsite potential disorder. We find a transition from the gapped Mott phase with algebraic spin correlations to a gapless spin-disordered phase beyond a critical strength of the disorder Δc≈U/2Δ_c \approx U/2Δc​≈U/2. Both the transitions in the charge and spin sectors are shown to be coincident. We also establish the finite-size corrections to the charge gap and the spin-spin correlation length in the presence of disorder and using a finite-size-scaling analysis we obtain the zero temperature phase diagram of the various quantum phase transitions that occur in the disorder-interaction plane.
We apply the density-matrix renormalization group (DMRG) method to a one-dimensional Hubbard model that lacks Umklapp scattering and thus provides an ideal case to study the Mott-Hubbard transition analytically and numerically. The model has a linear dispersion and displays a metal-to-insulator transition when the Hubbard interaction UUU equals the bandwidth Uc=W{U}_{\mathrm{c}}=WUc​=W, where the single-particle gap opens linearly, $\mathrm{\ensuremath{\Delta}}(U\ensuremath{\ge}W)=U\ensuremath{-}W$. The simple nature of the elementary excitations permits us to determine numerically with high accuracy the critical interaction strength and the gap function in the thermodynamic limit. The jump discontinuity of the momentum distribution nk{n}_{k}nk​ at the Fermi wave number kF=0{k}_{\mathrm{F}}=0kF​=0 cannot be used to locate accurately Uc{U}_{\mathrm{c}}Uc​ from finite-size systems. However, the slope of nk{n}_{k}nk​ at the band edges ${k}_{\mathrm{B}}=\ifmmode\pm\else\textpm\fi{}\ensuremath{\pi}$, reveals the formation of a single-particle bound state which can be used to determine Uc{U}_{\mathrm{c}}Uc​ reliably from nk{n}_{k}nk​ using accurate finite-size data.
We study transport properties of the half-filled two-dimensional (2D) Hubbard model with spatially varying interactions, where a pattern of interacting and non-interacting sites is formed. We use Determinantal Quantum Monte Carlo method to calculate the double occupation, effective hopping and Drude weight. These data point to two phase transitions, driven by fermionic correlations. The first is the expected metal to a Mott insulating state. The second one, is an exotic transition from a Mott insulating state to a highly anisotropic metal, that takes place at large values of the fermion-fermion interaction. This second transition occurs when the layers formed by the spatially varying interactions decouple due to the suppression of the hopping between interacting and non-interacting sites, leading to fermionic transport along the non-interacting ones.
We study the ground state of the two-dimensional Anderson-Hubbard model using a quantum real space renormalization group method. We obtain the phase diagram near half filling. The system is always insulating with disorder. At half filling, the system undergoes a transition from a gapless (Anderson) insulator to an incompressible (Mott-Hubbard) insulator as the interaction UUU reaches a critical value UcU_cUc​. Away from half filling, the insulating phase is always gapless and is found to be controlled by the Anderson fixed point at half filling. This result is similar to that obtained in the corresponding one dimensional system and suggests strongly the importance of the electron-electron correlation in this gapless insulating phase.