Physicality of singular Bethe solutions in the remaining open cases: sufficiency for the higher-spin XXZ chain and eigenstate construction for strange repeated-root solutions

The Bethe equations of the periodic XXX and XXZ Heisenberg spin chains admit singular solutions, containing an exact string such as $\{i/2,-i/2\}$ for spin 1/2, for which the naive eigenvalue and Bethe vector are ill-defined. A 2014 twist-regularization analysis (Nepomechie and Wang) derived necessary conditions for such solutions to be physical, but left three gaps: sufficiency of the conditions, an explicit construction of the physical eigenvectors at arbitrary chain length and magnon number, and the treatment of 'strange' singular solutions with repeated roots for spin $s>1/2$. Later work closed most of this program: sufficiency is now proven for the spin-1/2 and spin-$s$ XXX chains and for the spin-1/2 XXZ chain — in the trigonometric case through quasi-polynomiality of the dual solution of the TQ-relation rather than plain polynomiality — via rational Q-systems, and an explicit regularization recipe for non-repeated singular solutions of the higher-spin XXX chain has recently appeared in an unrefereed preprint. This problem records the surviving core: prove or refute sufficiency of the physicality condition for the higher-spin (spin-$s$) XXZ chain, where no Q-system or regularization-based proof exists, and construct the physical eigenstates of strange repeated-root singular solutions at general multiplicity within the algebraic Bethe ansatz, together with a classification of the admissible multiplicity patterns.

Background

Bethe ansatz and singular solutions. The periodic isotropic Heisenberg (XXX) quantum spin chain of $N$ sites is diagonalized by the Bethe ansatz: eigenstates are labeled by sets of complex rapidities (Bethe roots) $\{\lambda_1,\dots,\lambda_M\}$, where $M$ is the number of spin flips (magnons), satisfying the Bethe equations

$$\left(\frac{\lambda_j+\frac{i}{2}}{\lambda_j-\frac{i}{2}}\right)^{N} =\prod_{k\neq j}\frac{\lambda_j-\lambda_k+i}{\lambda_j-\lambda_k-i},\qquad j=1,\dots,M.$$

In the algebraic Bethe ansatz (ABA), an eigenstate is built as $B(\lambda_1)\cdots B(\lambda_M)|0\rangle$, where $B(\lambda)$ is an off-diagonal entry of the monodromy matrix (the ordered product of Lax operators around the chain) and $|0\rangle$ is the fully polarized reference state; the transfer matrix, whose logarithmic derivative gives the Hamiltonian, has an eigenvalue and an energy that are rational functions of the roots. The XXZ chain is the anisotropic deformation with anisotropy parameter $\eta$ ($q=e^{\eta}$), obtained by replacing $\lambda\mapsto\sinh\lambda$ and $i/2\mapsto\eta/2$. Integrable spin-$s$ versions exist for every half-integer $s$; their Bethe equations admit the special values $is,i(s-1),\dots,-is$ as roots.

A Bethe solution is called singular if it contains an exact string: for spin 1/2 the pair $\{i/2,-i/2\}$; for spin $s$ the string of length $2s+1$ centered at the origin. For such solutions both the energy formula (which has poles at the string values) and the Bethe vector (which becomes the null vector) are ill-defined, so a singular solution is not automatically an eigenstate. Singular solutions divide into physical ones, which do correspond to genuine eigenvalues and eigenvectors of the Hamiltonian after proper regularization, and unphysical ones, which do not. The simplest example, due to Avdeev and Vladimirov, is $\{i/2,-i/2\}$: physical when $N$ is even, unphysical when $N$ is odd. Numerics show that most singular solutions are unphysical, so a simple criterion selecting the physical ones is essential for completeness of the Bethe ansatz, i.e. for the statement that regular plus physical singular solutions account for all $2^N$ states of the spin-1/2 chain.

Physicality conditions and the twist regularization. Nepomechie and Wang (2013) regularized the singular pair by displacing the two roots by a small parameter $\epsilon$ together with a correction of order $\epsilon^{N}$, and demanded that the regularized Bethe vector be an eigenvector; this yielded the extra condition

$$\left[-\prod_{j=3}^{M}\frac{\lambda_j+\frac{i}{2}}{\lambda_j-\frac{i}{2}}\right]^{N}=1$$

on the remaining roots, alongside the reduced Bethe equations they satisfy. Hao, Nepomechie and Sommese generalized the condition to spin-$s$ chains and to solutions with repeated roots. Nepomechie and Wang (2014) re-derived these conditions by a twist regularization: periodic boundary conditions are replaced by twisted ones with angle $\beta$, for which no singular solutions exist, and the limit $\beta\to0$ is taken. This produced, for the first time, the corresponding conditions for the XXZ chains, e.g. for spin $s$

$$\left[(-1)^{2s}\prod_{j=2s+2}^{M}\frac{\sinh(\lambda_j+s\eta)}{\sinh(\lambda_j-s\eta)}\right]^{N}=1,$$

but the twist arguments establish only necessity. Sufficiency was attributed to the earlier ABA analyses and to numerics; the explicit eigenvector construction inside the twist framework was carried out only for the smallest case $(N,M)=(4,2)$, where the roots are expanded in powers of $\beta$ up to order $\beta^{N}$, the Bethe vector is renormalized by $1/\beta^{N}$, and the limit $\beta\to0$ is taken. The same paper observed that for $s>1/2$ there also exist physical 'strange' singular solutions with repeated roots, and expected that the twist regularization could be extended to derive conditions for them, without carrying this out.

The Q-system reformulation and what it settled. A parallel line of work recast the Bethe equations through Baxter's TQ-relation, a second-order difference equation

$$\tau(u)\,Q(u)=(u+is)^{L}\,Q(u-i)+(u-is)^{L}\,Q(u+i)$$

(here $L$ is the chain length) whose polynomial solution $Q$ has the Bethe roots as zeros, and through the rational Q-system, an equivalent set of QQ-relations whose polynomial solutions can be found efficiently. The crucial structural fact is that the TQ-relation has a second, dual solution $P(u)$: when the zeros of $Q$ form a singular string, $P$ need not be a polynomial, and requiring $P$ to be a polynomial imposes exactly the physicality condition above. Granet and Jacobsen (2020) proved for the periodic spin-1/2 XXX chain that physical solutions are exactly the zero-remainder (QQ) ones, including the statement that a regularization of a singular solution whose normalized Bethe-vector limit is a genuine transfer-matrix eigenvector exists if and only if $P$ is a polynomial. Bajnok, Granet, Jacobsen and Nepomechie (2020) formulated Q-systems for the closed XXZ chain (and open chains) and proved, at generic anisotropy, the analogous equivalence chain: quasi-polynomiality of $P$ (a polynomial in $t=e^{u}$ and $t^{-1}$, up to a $\log t$ term), polynomiality of the $Q$-system functions, admissibility of the roots, and the Bethe vector being an eigenvector of the transfer matrix are all equivalent. Tarasov (2018) proved completeness for the periodic isotropic model of arbitrary spin: a polynomial $\tau(u)$ is an eigenvalue of the transfer matrix if and only if the TQ equation has two polynomial solutions $Q,P$ (with $\deg Q<\deg P$), the eigenvector being unique up to proportionality. On this basis Hou, Jiang and Zhu (2024) constructed the rational Q-system for the spin-$s$ XXX chain, proved its completeness, and derived physicality conditions for singular solutions with arbitrary multiplicities of the singular-string roots, which cover the strange configurations; He, Hou, Liu and Tan (2025) extended the analysis to twisted and open boundaries. Two parts of the original 2014 program nevertheless remain unsettled: no Q-system or analogue of these proofs exists for the higher-spin XXZ chain, where sufficiency of the condition displayed above is known only as a necessity result supported by analogy; and the constructive ABA side for repeated roots, i.e. explicitly regularized Bethe states and a classification of which multiplicity patterns of the singular string can occur in physical solutions, is explicitly listed as open in Hou, Jiang and Zhu (2024).

Problem Statement

Complete the physicality theory of singular solutions of Bethe's equations for the periodic integrable Heisenberg chains in the two cases not settled by the Q-system literature, with the common objective of a uniform constructive criterion: new necessary-and-sufficient physicality conditions, together with explicit regularization recipes producing the corresponding physical eigenvectors, are required only for these two unsettled cases, and they must reduce to the known conditions in the already-settled cases (spin-1/2 and spin-$s$ XXX, spin-1/2 XXZ), so that a single criterion covers every chain in the family (XXX and XXZ, spin 1/2 and spin $s$) and every singular solution type (non-repeated and repeated singular roots).

(1) Higher-spin XXZ chain. For the integrable spin-$s$ XXZ chain with periodic boundary conditions and generic anisotropy ($q=e^{\eta}$ not a root of unity), a singular solution contains the exact string $\{s\eta,(s-1)\eta,\dots,-s\eta\}$ followed by remaining roots $\lambda_{2s+2},\dots,\lambda_M$ (all roots assumed distinct, both from the string and among themselves). Prove, or refute, that the condition

$$\left[(-1)^{2s}\prod_{j=2s+2}^{M}\frac{\sinh(\lambda_j+s\eta)}{\sinh(\lambda_j-s\eta)}\right]^{N}=1$$

together with the reduced Bethe equations satisfied by the remaining roots,

$$\left(\frac{\sinh(\lambda_j+s\eta)}{\sinh(\lambda_j-s\eta)}\right)^{N-1}\frac{\sinh(\lambda_j-(s+1)\eta)}{\sinh(\lambda_j+(s+1)\eta)}=\prod_{k=2s+2,\,k\neq j}^{M}\frac{\sinh(\lambda_j-\lambda_k+\eta)}{\sinh(\lambda_j-\lambda_k-\eta)},\qquad j=2s+2,\dots,M,$$

is necessary and sufficient for such a solution to be physical, in the constructive sense that a regularization of the singular roots exists whose renormalized Bethe-vector limit is a genuine eigenvector of the transfer matrix; and exhibit the corresponding construction for arbitrary chain length $N$, magnon number $M$, and spin $s$. In the refutation branch, a completion consists of an explicit example separating physicality from the condition: a solution satisfying the condition and the reduced Bethe equations that admits no such regularized limit, or a physical solution that violates the condition.

(2) Strange repeated-root solutions. For chains with $s>1/2$ (both XXX and XXZ), construct the physical eigenvectors corresponding to strange singular solutions, i.e. those with repeated roots inside the singular string with multiplicities $(m_0,\dots,m_{2s})$ (in the general case the regular part of the solution may also contain repeated roots), by an explicit regularization of the Bethe roots at general multiplicity within the algebraic Bethe ansatz (analogous to the non-repeated recipe $u_j^{\epsilon}=i(s+1-j)+\epsilon+c_j\epsilon^{L}$ for a chain of length $L$), and classify which multiplicity patterns can occur in physical solutions. For the XXZ chain this includes deciding what the correct physicality conditions for repeated-root singular solutions are, since beyond the necessity-only twist result no conditions are known there at all.

References

1. R. I. Nepomechie and C. Wang, "Twisting singular solutions of Bethe's equations", J. Phys. A: Math. Theor. 47 (2014) 505004. DOI: 10.1088/1751-8113-47-50-505004. URL: https://arxiv.org/abs/1409.7382
2. R. I. Nepomechie and C. Wang, "Algebraic Bethe ansatz for singular solutions", J. Phys. A: Math. Theor. 46 (2013) 325002. DOI: 10.1088/1751-8113-46-32-325002. URL: https://arxiv.org/abs/1304.7978
3. W. Hao, R. I. Nepomechie and A. J. Sommese, "Completeness of solutions of Bethe's equations", Phys. Rev. E 88 (2013) 052113. DOI: 10.1103/PhysRevE.88.052113. URL: https://arxiv.org/abs/1308.4645
4. W. Hao, R. I. Nepomechie and A. J. Sommese, "Singular solutions, repeated roots and completeness for higher-spin chains", J. Stat. Mech. (2014) P03024. DOI: 10.1088/1742-5468/2014/03/P03024. URL: https://arxiv.org/abs/1312.2982
5. L. Avdeev and A. Vladimirov, "Exceptional solutions to the Bethe ansatz equations", Theor. Math. Phys. 69 (1986) 1071. DOI: 10.1007/BF01037864. URL: https://doi.org/10.1007/BF01037864
6. A. N. Kirillov and R. Sakamoto, "Singular Solutions to the Bethe Ansatz Equations and Rigged Configurations", J. Phys. A: Math. Theor. 47 (2014) 205207. DOI: 10.1088/1751-8113-47-20-205207. URL: https://arxiv.org/abs/1402.0651
7. A. N. Kirillov and R. Sakamoto, "Some Remarks On Nepomechie-Wang Eigenstates For Spin 1/2 XXX Model", arXiv:1406.1958 (2014). URL: https://arxiv.org/abs/1406.1958
8. R. Sakamoto, "Rigged configurations approach for the spin-1/2 isotropic Heisenberg model", J. Phys. A: Math. Theor. 48 (2015) 165201. DOI: 10.1088/1751-8113-48-16-165201. URL: https://arxiv.org/abs/1501.00532
9. E. Granet and J. L. Jacobsen, "On zero-remainder conditions in the Bethe ansatz", JHEP 2020 (2020) 178. DOI: 10.1007/JHEP03(2020)178. URL: https://arxiv.org/abs/1910.07797
10. Z. Bajnok, E. Granet, J. L. Jacobsen and R. I. Nepomechie, "On generalized Q-systems", JHEP 2020 (2020) 177. DOI: 10.1007/JHEP03(2020)177. URL: https://arxiv.org/abs/1910.07805
11. V. Tarasov, "Completeness of the Bethe Ansatz for the Periodic Isotropic Heisenberg Model", Rev. Math. Phys. 30 (2018) 1840018. DOI: 10.1142/S0129055X18400184. URL: https://doi.org/10.1142/S0129055X18400184
12. J. Hou, Y. Jiang and R.-D. Zhu, "Spin-s rational Q-system", SciPost Phys. 16 (2024) 113. DOI: 10.21468/SciPostPhys.16.4.113. URL: https://arxiv.org/abs/2303.07640
13. Y.-J. He, J. Hou, Y.-C. Liu and Z.-X. Tan, "Spin-s Q-systems: Twist and open boundaries", SciPost Phys. Core 8 (2025) 057. DOI: 10.21468/SciPostPhysCore.8.3.057. URL: https://arxiv.org/abs/2502.15636
14. Y. Hao, Y. Jiang, B.-Q. Yang and D. Zhong, "Quantum Entanglement of Bethe States", arXiv:2606.14140 (2026), preprint. URL: https://arxiv.org/abs/2606.14140
15. R. I. Nepomechie, "Completing the solution for the OSp(1|2) spin chain", Nucl. Phys. B 951 (2020) 114887. DOI: 10.1016/j.nuclphysb.2019.114887. URL: https://arxiv.org/abs/1910.05127