The Bethe equations of the integrable generalized (graded, multi-leg) Hubbard models constructed by Fomin, Frappat and Ragoucy differ from the Lieb-Wu equations of the one-dimensional Hubbard model by two discrete, state-dependent phase factors e2πiΦe^{2\pi i\Phi} and e2πiΨe^{2\pi i\Psi}. The source paper left open whether Takahashi's string hypothesis survives these phases in the thermodynamic limit: preliminary numerics suggested that, for Φ,Ψ≠0\Phi,\Psi\neq0, the imaginary parts of complex Bethe roots deviate from the standard integer multiples of UU, so that the conjectured Takahashi-type equations for the string centres might hold only in zero-phase subsectors.

Background

The one-dimensional Hubbard model describes spin-1/2 fermions hopping on a chain with an on-site repulsion UU. Lieb and Wu (1968) diagonalized it exactly by the Bethe ansatz: its eigenstates are parameterized by charge momenta kjk_j and spin rapidities Λm\Lambda_m satisfying coupled algebraic Bethe equations. Finite-temperature and excited-state computations rely on Takahashi's string hypothesis (1972): in the thermodynamic limit of chain length L→∞L\to\infty, the complex solutions of these equations organize into regular patterns called strings. Two kinds occur: Λ\Lambda-strings, where nn spin rapidities sit at Λa,jn=Λan+iU(n+1−2j)\Lambda^n_{a,j}=\Lambda^n_a+iU(n+1-2j) around a common real centre Λan\Lambda^n_a (adjacent imaginary parts differing by 2U2U), and kk-Λ\Lambda strings mixing 2n2n momenta with nn rapidities. Substituting these patterns into the Bethe equations and taking logarithms yields discrete Takahashi equations for the real string centres, the starting point of the thermodynamic Bethe ansatz (TBA), i.e. integral equations for root-density functions from which thermodynamics and excitation spectra are computed. Strings in such models are asymptotic objects: at finite LL the roots only approximately follow the pattern, with deviations whose fate as LL grows must be analyzed, not assumed.

The source paper concerns a family of integrable generalizations of the Hubbard model built from Shastry's R-matrix construction: the R-matrix is the two-particle scattering matrix satisfying the Yang-Baxter consistency relation, and it is the object from which the exact solvability of the one-dimensional Hubbard model follows. Shastry's construction was extended to auxiliary spaces carrying representations of the Lie superalgebras gl(n∣m)\mathfrak{gl}(n|m), i.e. Z2\mathbb{Z}_2-graded extensions of the general linear algebra that mix bosonic and fermionic auxiliary degrees of freedom. The resulting models are the universal or generalized Hubbard models, whose electron sectors can be interpreted as integrable multi-leg Hubbard models, i.e. coupled chains (ladders). The motivation was twofold. In condensed matter, multi-leg Hubbard models approximate aspects of the genuinely two-dimensional Hubbard model while remaining exactly solvable in one dimension. In the AdS/CFT correspondence (the duality relating a four-dimensional conformal gauge theory to string theory on a curved higher-dimensional background), the Bethe equations of planar N=4\mathcal{N}=4 super Yang-Mills theory and of the AdS5×S5AdS_5\times S^5 superstring contain a momentum-dependent dressing phase eiθ(uk,uj)e^{i\theta(u_k,u_j)} (Arutyunov-Frolov-Staudacher, 2004), and one sought Hubbard-type lattice models whose Bethe equations reproduce such a phase.

For these generalized models Fomin, Frappat and Ragoucy derived the full Bethe ansatz equations by a coordinate Bethe ansatz (the direct construction of multiparticle wavefunctions from two-body scattering data, without reference to transfer matrices). After the momenta of one class of particles (the π\pi-particles) and certain auxiliary discrete variables blb_l are fixed by quantization conditions, the remaining dynamical momenta kjk_j and rapidities Λm\Lambda_m obey exactly the Lieb-Wu equations modified by two extra phase factors e2πiΦe^{2\pi i\Phi} and e2πiΨe^{2\pi i\Psi} (eqs. (4.4)-(4.5) of the source paper). The phases are not free parameters: Φ\Phi and Ψ\Psi are computable functions of the state's discrete quantum numbers (the occupation numbers of the particle species and integer labels), and in the thermodynamic limit the ratios Φ/L\Phi/L and Ψ/L\Psi/L remain finite and density-dependent, so the phases grow with LL and change from sector to sector. In zero-phase subsectors the model behaves as the usual Hubbard model: at half filling the ground state has the Lieb-Wu root densities (the phases drop out of the differentiated equations). Assuming strings also for nonzero phases, the authors derived the corresponding Takahashi-type discrete equations for the string centres, eqs. (4.26)-(4.28), with quantum-number bounds eqs. (4.30)-(4.32) shifted by the extremal values of Φ\Phi and Ψ\Psi. These equations are explicitly conditional: preliminary numerics reported in the paper indicate that when Φ,Ψ≠0\Phi,\Psi\neq0 the imaginary parts of complex Bethe roots are no longer spaced by integer multiples of UU, so the standard string pattern, and hence the derived Takahashi equations, may fail outside the zero-phase subsectors. Settling this is the prerequisite for any excited-state spectrum or thermodynamics of these models in the general sectors.

Problem Statement

For the generalized (graded, multi-leg) Hubbard models of Fomin-Frappat-Ragoucy, whose dynamical Bethe equations for the momenta kjk_j and spin rapidities Λm\Lambda_m are the Lieb-Wu equations supplemented by the discrete, state-dependent phases e2πiΦe^{2\pi i\Phi} and e2πiΨe^{2\pi i\Psi} (eqs. (4.4)-(4.5) of the source paper), determine the structure of the complex Bethe roots in the thermodynamic limit at generic nonzero Φ\Phi and Ψ\Psi. The single objective is the thermodynamic-limit solution structure of these phased Bethe equations, pursued through three jointly required deliverables: (1) decide whether the Hubbard string hypothesis (real kjk_j, Λ\Lambda-strings with spacings that are integer multiples of UU, and kk-Λ\Lambda strings) remains valid in the nonzero-phase sectors, whether it is modified as the source's preliminary numerics suggested, or whether no asymptotic string structure survives; (2) if the standard strings are deformed, determine the correct pattern of complex roots; and (3) if a string-like pattern (standard or deformed) survives, derive the corresponding Takahashi-type discrete equations for the real string centres, validating or correcting eqs. (4.26)-(4.28) together with the phase-shifted quantum-number bounds (4.30)-(4.32), taking into account that Φ\Phi and Ψ\Psi depend on the state's discrete quantum numbers and are of order LL in the thermodynamic limit, so they act as state-dependent boundary twists rather than fixed small phases. This structure is the prerequisite for the excited-state spectrum and thermodynamics of these models; a resolution consists of a demonstrated root structure (the surviving string pattern, or a demonstration that no asymptotic string structure remains) together with the discrete equations any surviving pattern implies.

References

  1. V. Fomin, L. Frappat, E. Ragoucy, "Including a phase in the Bethe equations of the Hubbard model", J. High Energy Phys. 2010 (2010) 062, DOI: 10.1007/JHEP04(2010)062, arXiv:1002.1806, https://arxiv.org/abs/1002.1806 (source paper)
  2. E. H. Lieb, F. Y. Wu, "Absence of Mott Transition in an Exact Solution of the Short-Range, One-Band Model in One Dimension", Phys. Rev. Lett. 20 (1968) 1445, DOI: 10.1103/PhysRevLett.20.1445, https://doi.org/10.1103/PhysRevLett.20.1445
  3. M. Takahashi, "One-Dimensional Hubbard Model at Finite Temperature", Prog. Theor. Phys. 47 (1972) 69, DOI: 10.1143/PTP.47.69, https://doi.org/10.1143/PTP.47.69
  4. B. S. Shastry, "Exact Integrability of the One-Dimensional Hubbard Model", Phys. Rev. Lett. 56 (1986) 2453, DOI: 10.1103/PhysRevLett.56.2453, https://doi.org/10.1103/PhysRevLett.56.2453
  5. V. Fomin, L. Frappat, E. Ragoucy, "Bethe equations for generalized Hubbard models", J. High Energy Phys. 2009 (2009) 055, DOI: 10.1088/1126-6708/2009-09-055, arXiv:0906.4512, https://arxiv.org/abs/0906.4512
  6. G. Feverati, L. Frappat, E. Ragoucy, "Universal Hubbard models with arbitrary symmetry", J. Stat. Mech. (2009) P04014, DOI: 10.1088/1742-5468/2009/04/P04014, arXiv:0903.0190, https://arxiv.org/abs/0903.0190
  7. V. Fomin, "Modèles de Hubbard unidimensionels généralisés" ("One-dimensional integrable Hubbard model and its generalizations"), PhD thesis, LAPTH Annecy (defended 20 September 2010), HAL: tel-00608340, https://theses.hal.science/tel-00608340
  8. G. Arutyunov, S. Frolov, M. Staudacher, "Bethe Ansatz for Quantum Strings", J. High Energy Phys. 2004 (2004) 016, DOI: 10.1088/1126-6708/2004-10-016, arXiv:hep-th/0406256, https://arxiv.org/abs/hep-th/0406256
  9. A. A. Ziolkowska, F. H. L. Essler, "Yang-Baxter integrable Lindblad equations", SciPost Phys. 8 (2020) 044, DOI: 10.21468/SciPostPhys.8.3.044, arXiv:1911.04883, https://arxiv.org/abs/1911.04883
  10. M. V. Medvedyeva, F. H. L. Essler, T. Prosen, "Exact Bethe Ansatz Spectrum of a Tight-Binding Chain with Dephasing Noise", Phys. Rev. Lett. 117 (2016) 137202, DOI: 10.1103/PhysRevLett.117.137202, https://doi.org/10.1103/PhysRevLett.117.137202
  11. M. de Leeuw, C. Paletta, B. Pozsgay, "A range three elliptic deformation of the Hubbard model", SciPost Phys. 15 (2023) 071, DOI: 10.21468/SciPostPhys.15.2.071, arXiv:2301.01612, https://arxiv.org/abs/2301.01612