Lieb and Wu solved the one-dimensional Hubbard chain of two-flavor fermions by the coordinate Bethe ansatz, and in the continuum the analogous delta-function gas is solvable for any number of internal states and any permutation symmetry (Yang-Sutherland). Choy and Haldane (1982) showed that this program breaks on the lattice: for the lattice gas with hopping −t-t and on-site pair interaction UU, the Lieb-Wu-type ansatz acquires a residual supported exactly on configurations with three or more particles on one site, and no Hermitian operator acting within the multiply-occupied on-site subspace can repair it or restore completeness, so the Bethe-ansatz states do not span the general permutation-symmetry sectors. The open problem is to determine whether this lattice gas admits an exact solution at finite U/tU/t and finite density: an exact solution in the Bethe-ansatz / exact-solvability sense used by Choy and Haldane and the subsequent literature — a complete solution with rapidities fixed by a finite system of Bethe-type equations determined by the model, not mere finite-dimensional diagonalization — whose eigenstates overcome the obstruction, remain valid on multiply-occupied configurations, and form a complete set in the general permutation-symmetry sectors. The problem remains open: every known exact construction excludes, projects out, or works around triple occupancy, while non-integrability is unproven for this family.

Background

The one-dimensional (1D) Hubbard model is the paradigm of an exactly solvable interacting lattice system: spin-1/2 fermions on a chain with nearest-neighbour hopping amplitude −t-t and on-site interaction UU between the two fermions of opposite spin that one site can host. Lieb and Wu solved it with the coordinate Bethe ansatz (BA), the hypothesis that a many-body eigenfunction is a superposition of free plane-wave sectors whose amplitudes are fixed entirely by two-particle scattering whenever two particles meet; consistency requires multi-particle scattering to factorize into two-body processes, the condition encoded in the Yang-Baxter equation, the algebraic backbone of quantum integrability. In the continuum the same program is far more general. Yang solved the two-component delta-function gas, and Sutherland's nested Bethe ansatz (in which the rapidities — the auxiliary quasi-momenta parametrizing one scattering process — become the particles of the next) handles any number of internal states and, in fact, arbitrary permutation symmetry: the many-body wavefunction may transform under any irreducible representation of the symmetric group SNS_N on the particles — fully antisymmetric (spinless fermions), fully symmetric (bosons), or any intermediate symmetry class labelled by a Young tableau (the standard row-and-column diagrams classifying such symmetry classes); the direct sum of all these sectors is the Hilbert space of distinguishable particles. On the lattice, the corresponding Hubbard-type lattice gas lets a site hold any number of particles, each co-resident pair costing energy UU:

H0=−t∑i=1N(Ti+Ti−1)+U2∑j=1Lnj(nj−1),H_0 = -t\sum_{i=1}^{N}\left(T_i + T_i^{-1}\right) + \frac{U}{2}\sum_{j=1}^{L} n_j\left(n_j-1\right),

where TiT_i shifts particle ii to the neighbouring site and njn_j counts the particles at site jj. In first-quantized form this is one Hamiltonian for bosons, distinguishable particles, and multi-flavor fermions (particles carrying one of several internal states), the symmetry class being imposed as a sector condition; its special cases include the boson-Hubbard chain (fully symmetric sector) and the NfN_f-flavor fermionic Hubbard chain with Nf≥3N_f\ge 3, where a site can hold up to NfN_f fermions. Choy (1980) for the degenerate (SU(NfN_f)-flavor) chain and Haldane (1980) for the boson-Hubbard chain proposed Lieb-Wu-type solutions for this gas. Choy and Haldane (1982) then showed the construction fails: for three distinguishable particles the BA state ∣k⟩|k\rangle built from the Lieb-Wu matching conditions obeys

⟨x∣(H0−E(k))∣k⟩=U24t f(k) ⟨x∣P3∣k⟩,\langle x | \left(H_0 - E(k)\right) |k\rangle = \frac{U^2}{4t}\, f(k)\, \langle x | P_3 |k\rangle,

where E(k)=−2t(cos⁡k1+cos⁡k2+cos⁡k3)E(k)=-2t(\cos k_1+\cos k_2+\cos k_3) is the free dispersion, P3P_3 projects onto configurations xx with all three particles on one site, and f(k)={cos⁡[(k1+k2)/2] cos⁡[(k2+k3)/2] cos⁡[(k3+k1)/2]}−1f(k)=\{\cos[(k_1+k_2)/2]\,\cos[(k_2+k_3)/2]\,\cos[(k_3+k_1)/2]\}^{-1}. The residual is supported exactly on triply-occupied sites. Worse, no Hermitian operator acting only within the multiply-occupied on-site subspace can cancel it, because such a correction would require f(k)f(k) to depend only on the conserved total momentum, which it does not; consequently the BA states do not form a complete orthogonal basis in the general permutation-symmetry sectors. The construction is exact only where triple occupancy has vanishing weight: the dilute (low-density) limit, which maps onto the Yang-Sutherland continuum gas, and the U→∞U\to\infty limit with at most one particle per site, which maps onto Sutherland's SU(NfN_f) generalization of the Heisenberg spin chain. This negative result remains the accepted status: the review of 1D Fermi gases by Guan, Batchelor and Lee states that the 1D multi-component Hubbard model is 'no longer BA solvable' because of it. Everything exactly solvable found since sidesteps multi-occupancy — projected degenerate chains, hard-core integrable cousins, continuum models, partial-integrability subsectors — so the question the 1982 paper left open, whether the general permutation-symmetry lattice gas admits any complete exact solution at finite UU and finite density, is still live.

Problem Statement

Consider the one-dimensional lattice gas of NN particles on an LL-site ring with periodic boundary conditions and Hamiltonian

H0=−t∑i=1N(Ti+Ti−1)+U2∑j=1Lnj(nj−1),H_0 = -t\sum_{i=1}^{N}\left(T_i + T_i^{-1}\right) + \frac{U}{2}\sum_{j=1}^{L} n_j\left(n_j-1\right),

where TiT_i shifts particle ii to the neighbouring site, njn_j counts the particles at site jj, and each co-resident pair at a site costs interaction energy UU; take the particle density n=N/Ln=N/L finite and nonzero and the ratio U/tU/t finite and nonzero. The many-body wavefunction is taken in an arbitrary permutation-symmetry sector — an irreducible representation of the symmetric group SNS_N on the particles — which covers the fully symmetric (bosonic) sector and the Nf≥3N_f\ge 3-flavor fermionic sectors as special cases. Choy and Haldane (1982) proved that the Lieb-Wu-type Bethe-Ansatz states for this model fail to be eigenstates precisely on configurations with three or more particles on a single site, and that no Hermitian operator supported on the multiply-occupied on-site subspace can repair them or make them a complete orthogonal basis for these sectors. As background, the failed construction is exact only where triple occupancy has vanishing weight: the dilute (low-density) limit, which maps onto the Yang-Sutherland solutions of the continuum delta-function gas, the U→∞U\to\infty limit with at most one particle per site, which maps onto Sutherland's SU(NfN_f) spin chain, and the two-flavor fermionic sector, which is solved by Lieb and Wu; these regimes lie outside the obstruction. Determine whether this model admits an exact solution at finite UU and finite density in the general permutation-symmetry sectors: an exact solution in the exact-solvability sense used by Choy and Haldane and the subsequent literature, that is, a complete solution — eigenstates that overcome the Choy-Haldane obstruction, are valid on all configuration classes, including multiply-occupied sites, form a complete set in each permutation-symmetry sector concerned, and have rapidities fixed by a finite system of Bethe-type equations determined by the model — rather than mere finite-dimensional exact diagonalization of H0H_0. The determination is required for the general case: every permutation-symmetry sector in which triple occupancy is possible must be covered, so a result valid for one sector alone — for instance the fully symmetric/bosonic sector — does not by itself settle the question. A rigorous proof that no solution of this class exists also settles the question in the negative, whereas a weaker obstruction that does not exclude the class is only a partial result.

References

  1. T. C. Choy and F. D. M. Haldane, Failure of Bethe-Ansatz solutions of generalisations of the Hubbard chain to arbitrary permutation symmetry, Physics Letters A 90 (1-2), 83-84 (1982). DOI: 10.1016/0375-9601(82)90057-3, https://doi.org/10.1016/0375-9601(82)90057-3
  2. E. H. Lieb and F. Y. Wu, Absence of Mott Transition in an Exact Solution of the Short-Range, One-Band Model in One Dimension, Physical Review Letters 20, 1445-1448 (1968). DOI: 10.1103/PhysRevLett.20.1445, https://doi.org/10.1103/PhysRevLett.20.1445
  3. C. N. Yang, Some Exact Results for the Many-Body Problem in one Dimension with Repulsive Delta-Function Interaction, Physical Review Letters 19, 1312-1315 (1967). DOI: 10.1103/PhysRevLett.19.1312, https://doi.org/10.1103/PhysRevLett.19.1312
  4. B. Sutherland, Further Results for the Many-Body Problem in One Dimension, Physical Review Letters 20, 98-100 (1968). DOI: 10.1103/PhysRevLett.20.98, https://doi.org/10.1103/PhysRevLett.20.98
  5. B. Sutherland, Model for a multicomponent quantum system, Physical Review B 12, 3795-3805 (1975). DOI: 10.1103/PhysRevB.12.3795, https://doi.org/10.1103/PhysRevB.12.3795
  6. T. C. Choy, Some exact results for a degenerate Hubbard model in one dimension, Physics Letters A 80, 49-52 (1980). DOI: 10.1016/0375-9601(80)90451-X, https://doi.org/10.1016/0375-9601(80)90451-X
  7. F. D. M. Haldane, 'Solidification' in a soluble model of bosons on a one-dimensional lattice: The 'Boson-Hubbard chain', Physics Letters A 80, 281-283 (1980). DOI: 10.1016/0375-9601(80)90022-5, https://doi.org/10.1016/0375-9601(80)90022-5
  8. K.-J.-B. Lee and P. Schlottmann, Ground state properties of the degenerate Hubbard chain: 1/N-expansion of the Bethe-Ansatz solution for a lattice system, Physica B: Condensed Matter 163, 398-400 (1990). DOI: 10.1016/0921-4526(90)90223-H, https://doi.org/10.1016/0921-4526(90)90223-H
  9. H. Frahm and A. Schadschneider, Critical exponents of the degenerate Hubbard model, Journal of Physics A: Mathematical and General 26, 1463-1480 (1993). arXiv:cond-mat/9207026, https://arxiv.org/abs/cond-mat/9207026 (DOI: 10.1088/0305-4470/26/7/009, https://doi.org/10.1088/0305-4470/26/7/009)
  10. W. Krauth, Bethe ansatz for the one-dimensional boson Hubbard model, Physical Review B 44, 9772-9775 (1991). DOI: 10.1103/PhysRevB.44.9772, https://doi.org/10.1103/PhysRevB.44.9772
  11. Z. Maassarani, The su(N) Hubbard model, Physics Letters A 239, 187-190 (1998). arXiv:cond-mat/9709252, https://arxiv.org/abs/cond-mat/9709252 (DOI: 10.1016/S0375-9601(97)00977-8, https://doi.org/10.1016/S0375-9601(97)00977-8)
  12. Z. Maassarani, Exact Integrability of the su(n) Hubbard model, Modern Physics Letters B 12, 51-56 (1998). arXiv:cond-mat/9710083, https://arxiv.org/abs/cond-mat/9710083 (DOI: 10.1142/S0217984998000081, https://doi.org/10.1142/S0217984998000081)
  13. B. Y. Hou, D. T. Peng and R.-H. Yue, The exact solution of the SU(3) Hubbard model, Nuclear Physics B 575, 561-578 (2000). arXiv:cond-mat/9910293, https://arxiv.org/abs/cond-mat/9910293 (DOI: 10.1016/S0550-3213(00)00094-8, https://doi.org/10.1016/S0550-3213(00)00094-8)
  14. D.-t. Peng and R.-h. Yue, Yang-Baxter equation for the R-matrix of 1-D SU(n) Hubbard model, Journal of Physics A: Mathematical and General 35, 6985-6993 (2002). arXiv:cond-mat/0109029, https://arxiv.org/abs/cond-mat/0109029 (DOI: 10.1088/0305-4470/35/33/302, https://doi.org/10.1088/0305-4470/35/33/302)
  15. Z.-J. Ying, Y.-Q. Li and S.-J. Gu, Multi-particle scattering at the same position in the Bethe ansatz for two-band electrons with spin-exchange interaction, Journal of Physics A: Mathematical and General 34, 3939-3946 (2001). DOI: 10.1088/0305-4470/34/18/318, https://doi.org/10.1088/0305-4470/34/18/318
  16. Z.-J. Ying, Y.-Q. Li, S.-J. Gu and B. Chen, Extension of Bethe ansatz to multiple occupancies for one-dimensional SU(4) fermions with delta-function interaction, Journal of Mathematical Physics 43, 5977-5986 (2002). DOI: 10.1063/1.1515380, https://doi.org/10.1063/1.1515380
  17. Z.-J. Ying, Y.-Q. Li and S.-J. Gu, Symmetries in the Hubbard model with n-fold orbital degeneracy, Journal of Mathematical Physics 42, 4865-4872 (2001). DOI: 10.1063/1.1402956, https://doi.org/10.1063/1.1402956
  18. X.-W. Guan, M. T. Batchelor and C. Lee, Fermi gases in one dimension: From Bethe ansatz to experiments, Reviews of Modern Physics 85, 1633-1691 (2013). arXiv:1301.6446, https://arxiv.org/abs/1301.6446 (DOI: 10.1103/RevModPhys.85.1633, https://doi.org/10.1103/RevModPhys.85.1633)
  19. A. Hosseinzadeh and S. A. Jafari, Generalization of Lieb-Wu wave function inspired by one-dimensional ionic Hubbard model, Annals of Physics 414, 168075 (2020). arXiv:1906.07793, https://arxiv.org/abs/1906.07793 (DOI: 10.1016/j.aop.2020.168075, https://doi.org/10.1016/j.aop.2020.168075)
  20. M. Nakagawa, H. Katsura and M. Ueda, Exact eigenstates of multicomponent Hubbard models: SU(N) magnetic eta pairing, weak ergodicity breaking, and partial integrability, Physical Review Research 6, 043259 (2024). arXiv:2205.07235, https://arxiv.org/abs/2205.07235 (DOI: 10.1103/PhysRevResearch.6.043259, https://doi.org/10.1103/PhysRevResearch.6.043259)
  21. M. Bonkhoff, G. Gurney, F. Theel, P. Schmelcher, N. L. Harshman and T. Posske, Symmetry and integrability in the anyon-Hubbard model (2026). arXiv:2605.28956, https://arxiv.org/abs/2605.28956
  22. F. Ishii and M. Sanatani, Proof of the absence of local conserved quantities in the Holstein model (2026). arXiv:2605.19606, https://arxiv.org/abs/2605.19606
  23. M. Futami, Absence of nontrivial local conserved quantities in the Hubbard model on the two or higher dimensional hypercubic lattice (2025). arXiv:2507.20106, https://arxiv.org/abs/2507.20106