Zhao, Jiang and Zhu (arXiv:2604.04627, 2026) proposed an effective Bethe ansatz (EBA) that approximates eigenstates of non-integrable quantum spin chains near an integrable point: the algebraic Bethe-ansatz wave function is retained while its rapidities (Bethe roots) are treated as variational parameters optimized by minimizing cost functions. The paper's outlook posed four unresolved issues. The proposed extension to spin-1 chains and deformations of the Takhtajan-Babujian model has since been carried out in a companion benchmark study (Wang and Zhu, arXiv:2604.04640), while three scientifically linked facets remain open. (i) A principled rule for selecting the off-shell Bethe manifold (choice of R-matrix, inhomogeneities, superpositions) and the cost function for a given non-integrable Hamiltonian and target observable, replacing the ad hoc choices tested so far. (ii) Identification of the physical mechanism by which the factorized-scattering picture encoded in the Bethe wave function breaks down in spin chains, whose quantum-field-theory counterpart is particle decay and inelastic production. (iii) The performance of the EBA under time evolution - quench dynamics, long-time evolution, prethermalization, and transport - including whether the fidelity of optimized states survives or decays. The problem is to settle these three facets, which jointly determine when and why the Bethe wave function with renormalized roots describes a non-integrable system.

Background

Bethe ansatz and off-shell states. A one-dimensional quantum many-body system is called integrable when it possesses an extensive set of independent conserved quantities; the archetype is the spin-1/21/2 Heisenberg XXX chain H0=∑n=1Lσ⃗n⋅σ⃗n+1H_0=\sum_{n=1}^{L}\vec{\sigma}_n\cdot\vec{\sigma}_{n+1} (periodic, with σ⃗n\vec{\sigma}_n the Pauli matrices on site nn). Such models are exactly solvable by the Bethe ansatz: eigenfunctions are superpositions of plane waves for quasi-particles (magnons) whose scattering is factorized, meaning many-body scattering reduces completely to two-body phase shifts. In the algebraic Bethe ansatz (ABA), eigenstates are built as ∣{uj}⟩=B(u1)⋯B(uM)∣Ω⟩|\{u_j\}\rangle=\mathbf{B}(u_1)\cdots\mathbf{B}(u_M)|\Omega\rangle, where ∣Ω⟩=∣↑L⟩|\Omega\rangle=|\uparrow^L\rangle is the ferromagnetic reference state and the magnon-creation operators B(u)\mathbf{B}(u) are entries of the monodromy matrix, a product of elementary RR-matrices (the 6-vertex RR-matrix for the XXX chain) acting on a two-dimensional auxiliary space; the transfer matrix (the auxiliary trace of the monodromy) generates all conserved charges. The complex parameters uju_j, called Bethe roots or rapidities, must satisfy the Bethe ansatz equations (BAE) for the state to be an eigenstate. A Bethe state whose roots solve the BAE is called on-shell; the same wave function with arbitrary roots not solving the BAE is called an off-shell Bethe state.

Integrability breaking and the effective Bethe ansatz. Deformations H(λ)=H0+λH1H(\lambda)=H_0+\lambda H_1 generically destroy integrability. The literature distinguishes, qualitatively, weak integrability breaking - perturbations that preserve quasi-conserved quantities to high order in λ\lambda, exemplified by the next-nearest-neighbour coupling of the J1J_1-J2J_2 chain, H1=∑nσ⃗n⋅σ⃗n+2H_1=\sum_n\vec{\sigma}_n\cdot\vec{\sigma}_{n+2} - from strong breaking, exemplified by a staggered magnetic field H1=∑n(−1)nσnzH_1=\sum_n(-1)^n\sigma_n^z, which rapidly produces generic non-integrable behavior such as chaotic level statistics. Zhao, Jiang and Zhu proposed that close to an integrable point the functional form of the Bethe wave function survives while the roots renormalize: treating the off-shell Bethe state as a variational ansatz, the roots (and, in refined versions, additional parameters of a more general 8-vertex RR-matrix (a two-parameter deformation of the 6-vertex RR-matrix whose corner entries flip two spins at once, so the number of down spins is no longer conserved), or site-dependent inhomogeneities {θj}\{\theta_j\} inserted into the transfer matrix) are optimized by minimizing cost functions - for the ground state the Rayleigh quotient H0=⟨{uj}∣H(λ)∣{uj}⟩/⟨{uj}∣{uj}⟩\mathcal{H}_0=\langle\{u_j\}|H(\lambda)|\{u_j\}\rangle/\langle\{u_j\}|\{u_j\}\rangle, and for excited states iterated functionals with large penalty constants KnK_n enforcing orthogonality to previously found states. Accuracy is quantified by the energy error, by the fidelity F=∣⟨ψex∣ψEBA⟩∣2F=|\langle\psi_{\mathrm{ex}}|\psi_{\mathrm{EBA}}\rangle|^2 against the exact eigenstate from exact diagonalization (ED), and by the bipartite entanglement entropy. Empirically, for the J1J_1-J2J_2 chain the ground-state fidelity stays near 11 up to the level crossing at the exactly solvable Majumdar-Ghosh point λ=0.5\lambda=0.5, whereas for the staggered-field chain it drops below 0.90.9 already at λ≈0.1\lambda\approx 0.1; the rate of breakdown therefore serves as a diagnostic of the strength of integrability breaking.

The open questions. The outlook section of the source paper lists four unresolved issues: the choice of the optimal off-shell Bethe state and cost function for a given non-integrable Hamiltonian; the extension to broader model classes, notably spin-1 chains and deformations of the integrable Takhtajan-Babujian spin-1 chain (the exactly solved arbitrary-spin Heisenberg chain of Takhtajan and Babujian; its deformations include the bilinear-biquadratic family that contains the AKLT point, the exactly solvable Affleck-Kennedy-Lieb-Tasaki valence-bond-solid chain); the physical mechanism behind the breakdown of the factorized-scattering picture in spin chains - in quantum field theory away from integrability, stable particles decay and particle production renders scattering inelastic, and related spin-chain phenomena such as quasiparticle confinement and meson formation after quenches to non-integrable chains are known; and the behavior under real-time evolution, including quench dynamics, prethermalization (the long-lived quasi-stationary state preceding thermalization in nearly integrable systems), and transport. The spin-1 extension was in fact carried out in a companion benchmark study (Wang and Zhu, arXiv:2604.04640), which validated the EBA on the spin-1 bilinear-biquadratic chain (a nearest-neighbour spin-1 chain with both bilinear S⃗n⋅S⃗n+1\vec{S}_n\cdot\vec{S}_{n+1} and biquadratic (S⃗n⋅S⃗n+1)2(\vec{S}_n\cdot\vec{S}_{n+1})^2 exchange terms) from both integrable endpoints (the Takhtajan-Babujian point and the Lai-Sutherland point, the second integrable point of that family); the other three facets remain open and constitute the problem.

Problem Statement

Determine the domain of validity of the effective Bethe ansatz - off-shell Bethe states whose rapidities and auxiliary algebraic parameters are fixed by variational optimization - for the eigenstates and dynamics of non-integrable quantum many-body systems near an integrable point. For Hamiltonians of the form H(λ)=H0+λH1H(\lambda)=H_0+\lambda H_1 with H0H_0 solvable by the Bethe ansatz (starting from, but not limited to, the deformations already studied: the J1J_1-J2J_2 and staggered-field perturbations of the XXX chain, and the spin-1 bilinear-biquadratic chain), three linked questions must be settled:

  1. Optimal ansatz and cost function. Given H(λ)H(\lambda) and a specified target - a particular eigenstate, or accuracy over a specified set of observables - establish a principled rule for choosing (a) the off-shell Bethe manifold (which RR-matrix: 6-vertex, 8-vertex, or inhomogeneous transfer matrix; whether to superpose states from different magnon sectors) and (b) the cost function (energy-based, fidelity-based, or observable-based). The rule must come with a precise optimality criterion (for example, best achievable fidelity per variational parameter, or uniform accuracy over the specified observable set) and benchmark evidence, on the models above, that it matches or outperforms the ad hoc choices already tested.

  2. Mechanism of breakdown. Identify the physical mechanism by which the factorized-scattering picture underlying the Bethe wave function fails as integrability-breaking grows in spin chains. In relativistic quantum field theory, integrability breaking allows stable particles to decay and allows production and annihilation processes that render scattering inelastic; determine whether an analogous mechanism (quasiparticle decay channels, confinement and meson formation, or another identified process) or a different one governs the observed rapid loss of fidelity in spin chains.

  3. Real-time performance. Determine how effective-Bethe-ansatz states behave under time evolution by H(λ)H(\lambda): whether the fidelity of an optimized state (or a superposition of optimized states) to the corresponding exact eigenstate, or of an EBA-prepared initial state to the exactly time-evolved state, remains robust or decays, and on what timescale; and whether EBA-based approximations reproduce quench dynamics, prethermalization plateaus, and transport observables of the non-integrable chain.

The three questions jointly answer one overarching question: when, how well, and why does the Bethe wave function with renormalized roots describe a non-integrable system.

References

  1. W. Zhao, Y. Jiang, R.-D. Zhu, "The Roaming Bethe Roots: An Effective Bethe Ansatz Beyond Integrability," arXiv:2604.04627 (2026), doi:10.48550/arXiv.2604.04627, https://arxiv.org/abs/2604.04627
  2. Z. Wang, R.-D. Zhu, "Effective Bethe Ansatz for Spin-1 Non-integrable Models," Journal of Physics A: Mathematical and Theoretical 59 (2026) 305001, doi:10.1088/1751-8121/ae880e, arXiv:2604.04640, https://arxiv.org/abs/2604.04640
  3. Y. Hao, Y. Jiang, B.-Q. Yang, D.-l. Zhong, "Quantum Entanglement of Bethe States," arXiv:2606.14140 (2026), doi:10.48550/arXiv.2606.14140, https://arxiv.org/abs/2606.14140
  4. L. A. Takhtajan, "The picture of low-lying excitations in the isotropic Heisenberg chain of arbitrary spins," Physics Letters A 87 (1982) 479-482, doi:10.1016/0375-9601(82)90764-2, https://doi.org/10.1016/0375-9601(82)90764-2
  5. H. M. Babujian, "Exact solution of the one-dimensional isotropic Heisenberg chain with arbitrary spins S," Physics Letters A 90 (1982) 479-482, doi:10.1016/0375-9601(82)90403-0, https://doi.org/10.1016/0375-9601(82)90403-0
  6. F. M. Surace, O. Motrunich, "Weak integrability breaking perturbations of integrable models," Physical Review Research 5 (2023) 043019, doi:10.1103/PhysRevResearch.5.043019, arXiv:2302.12804, https://arxiv.org/abs/2302.12804
  7. M. Kormos, M. Collura, G. Takacs, P. Calabrese, "Real-time confinement following a quantum quench to a non-integrable model," Nature Physics 13 (2017) 246-249, doi:10.1038/nphys3934, arXiv:1604.03571, https://arxiv.org/abs/1604.03571
  8. D. Raveh, R. I. Nepomechie, "Estimating Bethe roots with VQE," Journal of Physics A: Mathematical and Theoretical 57 (2024) 355303, doi:10.1088/1751-8121/ad6db2, arXiv:2404.18244, https://arxiv.org/abs/2404.18244