Rigorous foundations of Baxter's $T-Q$ relation for the generic periodic XYZ spin chain: the parameter $\beta$, completeness, singular solutions, and Bethe vectors

The periodic spin-$\frac{1}{2}$ XYZ chain — the most general anisotropic Heisenberg chain, with three distinct exchange couplings and no conserved magnetization — was parameterized at the level of eigenvalues by Baxter through a homogeneous $T-Q$ relation whose elliptic Bethe-ansatz equations contain an a priori undetermined integer parameter $\beta$ and are supplemented by the sum rule $2\sum_j\lambda_j=k+p\tau$. A 2024 numerical study by Xin Zhang conjectured that $\beta=p$ for every generic (non-root-of-unity) chain with an even number of sites, that the homogeneous $T-Q$ relation and its Bethe-ansatz equations give the complete spectrum, and that the Bethe-ansatz equations admit physical singular solutions of the bound-pair form $\lambda_1=-\lambda_2=\eta/2$; it further conjectured that the determinant formulas for scalar products of Bethe vectors, proven at roots of unity, persist for generic anisotropy, and it left the construction of Bethe vectors for odd-site chains — whose spectrum is given by an inhomogeneous $T-Q$ relation — unresolved. No analytic proof or refutation of any of these statements is known. The problem is to prove or refute these hypotheses at their stated generality, thereby placing the Bethe-ansatz solution of the archetype elliptic integrable model on rigorous footing.

Background

The spin-$\frac{1}{2}$ XYZ chain is the most general nearest-neighbor anisotropic Heisenberg chain: on a ring of $N$ sites it has Hamiltonian

$H=\sum_{n=1}^{N}\left(J_x\,\sigma_n^x\sigma_{n+1}^x+J_y\,\sigma_n^y\sigma_{n+1}^y+J_z\,\sigma_n^z\sigma_{n+1}^z\right),\qquad \vec\sigma_{N+1}\equiv\vec\sigma_1,$

with three distinct exchange couplings $J_x,J_y,J_z$. These are parametrized by two complex numbers — a crossing parameter $\eta$ and a modular parameter $\tau$ with $\mathrm{Im}\,\tau>0$ — through Jacobi theta functions $\theta_\alpha$ (elliptic generalizations of trigonometric functions, quasiperiodic on the torus with half-periods $1$ and $\tau$), for example $J_x=\theta_4(\eta)/\theta_4(0)$. When $\tau$ is purely imaginary and $\eta$ real or purely imaginary the chain is Hermitian; the model also admits non-Hermitian instances at generic complex parameters. Because the three couplings differ, the total magnetization $\sum_n\sigma_n^z$ is not conserved: the chain has no U(1) symmetry (no continuous rotational symmetry about any axis), and this absence is the source of most of its mathematical difficulty.

The chain is integrable through the eight-vertex model, the two-dimensional lattice model whose row-to-row transfer matrix is built from an elliptic R-matrix satisfying the Yang–Baxter equation (the consistency condition guaranteeing a commuting family of transfer matrices). The transfer matrix $t(u)=\mathrm{tr}_0\,R_{0,N}(u)\cdots R_{0,1}(u)$ — the trace, over an auxiliary space, of the ordered product of R-matrices at spectral parameter $u$ — commutes with itself for all $u$ and yields $H$ by a logarithmic derivative. Baxter's original solution (1972) parameterizes the eigenvalues $\Lambda(u)$ of $t(u)$ through a homogeneous $T-Q$ relation: with $Q(u)=\prod_{j=1}^{N/2}\theta_1(u-\lambda_j+\eta/2)$ built from $M=N/2$ Bethe roots $\lambda_j$ (the zeros of the auxiliary Q-function),

$\Lambda(u)=e^{\beta\gamma}\,\frac{\theta_1^{N}(u+\eta)}{\theta_1^{N}(\eta)}\,\frac{Q(u-\eta)}{Q(u)}+e^{-\beta\gamma}\,\frac{\theta_1^{N}(u)}{\theta_1^{N}(\eta)}\,\frac{Q(u+\eta)}{Q(u)},\qquad \gamma=i\pi\eta .$

Requiring $\Lambda(u)$ to be an entire function of $u$ gives the elliptic Bethe-ansatz equations (BAEs) for the roots,

$e^{2\beta\gamma}\left[\frac{\theta_1(\lambda_j+\eta/2)}{\theta_1(\lambda_j-\eta/2)}\right]^{N}\prod_{k\neq j}^{M}\frac{\theta_1(\lambda_j-\lambda_k-\eta)}{\theta_1(\lambda_j-\lambda_k+\eta)}=1,\qquad j=1,2,\dots,M,\qquad M=\frac{N}{2},$

and a solution is physical only if the roots also satisfy Baxter's sum rule $2\sum_{j=1}^{M}\lambda_j=k+p\tau$ with integers $k,p$. Two structural features distinguish this from the XXZ (six-vertex) case. First, the integer $\beta$ entering the exponentials is a priori free: it is fixed by no local equation, only by global quasiperiodicity of the elliptic $Q$-function on the torus — shifting a root by $\tau$ maps a solution to an equivalent one and sends $\beta\to\beta+2$. Second, the Bethe vector (the eigenvector) cannot be constructed by the ordinary algebraic Bethe ansatz, because without U(1) symmetry there is no reference state on which to build it; the generalized algebraic Bethe ansatz (GABA) instead expresses the Bethe vector as an infinite series whose convergence is uncontrolled. "Generic" throughout means that $\eta$ is not a root of unity, i.e. $\eta\neq(2K+2L\tau)/P$ for integers $P,K,L$; at roots of unity the Bethe vector truncates to finitely many terms and much more becomes provable.

Three strands of prior work delimit what is known. (i) Baxter (2002) settled completeness of the Bethe ansatz — the statement that the Bethe equations yield all $2^N$ eigenstates — for the six-vertex model, resolving the "beyond the equator", infinite-momenta and exact-complete-string problems, and presented the zero-field eight-vertex coordinate Bethe-ansatz equations as expected, but not proven, to be complete; the parameter $\beta$ of the homogeneous elliptic $T-Q$ relation has never been determined analytically in the generic case. (ii) At roots of unity, Slavnov, Zabrodin and Zotov (2020) proved determinant representations for scalar products (overlaps of on-shell and off-shell Bethe vectors) in the inhomogeneous eight-vertex model, and Kulkarni and Slavnov (2023) developed the GABA framework for the XYZ chain, computing monodromy-matrix actions and scalar products explicitly in the free-fermion case; nothing comparable exists for generic $\eta$. (iii) For odd $N$, where the above homogeneous relation is unavailable, the off-diagonal Bethe ansatz (ODBA — a method that builds $T-Q$ relations from functional identities of the transfer matrix at shifted arguments, without any reference state) yields an inhomogeneous $T-Q$ relation giving the exact spectrum (Cao, Cui, Yang, Shi and Wang, 2014), but no construction of the corresponding eigenvectors. Singular Bethe-ansatz solutions — solutions on which the naive Bethe vector vanishes and must be rescaled — are a well-studied phenomenon in the XXX and XXZ chains (Nepomechie and Wang; Hao, Nepomechie and Sommese, both 2013).

In 2024, Xin Zhang performed extensive numerical experiments (small system sizes, Hermitian and non-Hermitian parameters) on Baxter's $T-Q$ relation for the generic periodic chain with even $N$, and on this evidence proposed the hypotheses listed in the Problem Statement: $\beta=p$ (verified analytically only for $N=2$); completeness of the homogeneous $T-Q$ relation and its BAEs; existence of physical singular solutions with the bound pair $\lambda_1=\eta/2$, $\lambda_2=-\eta/2$ (1, 4 and 11 such solutions counted at $N=4,6,8$); the conjecture that the root-of-unity scalar-product formulas persist for generic $\eta$; and the observation that in the regime $(N-2M_0)\eta=2K_1+2L_1\tau$ with an integer $M_0\neq N/2$, where a homogeneous $T-Q$ relation with $M_0$ roots exists (as shown by Zhang, Klümper and Popkov), the sum rule disappears and the hypotheses may fail. The rigorous proofs are explicitly left by the author as "the objective in our forthcoming projects".

Problem Statement

The problem is to replace finite-size numerical evidence by rigorous analytic theory for Baxter's homogeneous $T-Q$ relation of the generic periodic spin-$\frac{1}{2}$ XYZ chain. All components below jointly serve this single objective — establishing that the $T-Q$ relation, its Bethe-ansatz equations, and the associated Bethe vectors fully and exactly describe the model — and a complete answer must address each component at the generality of the source (generic non-root-of-unity anisotropy $\eta$, Hermitian and non-Hermitian parameter values alike, for the site numbers and parity to which each component refers):

1. (Parameter $\beta$) Prove, or disprove, that for every generic periodic XYZ chain with an even number of sites $N$ the integer $\beta$ in the Bethe-ansatz equations equals the integer $p$ of the sum rule $2\sum_{j=1}^{N/2}\lambda_j=k+p\tau$, so that the sum rule can be written $\sin\!\big[\big(2\sum_{j}\lambda_j-\beta\tau\big)\pi\big]=0$ and the $M+1$ unknowns $\{\lambda_1,\dots,\lambda_M,\beta\}$, $M=N/2$, are determined by the $M$ Bethe-ansatz equations together with this one additional condition. The only analytic case established so far is $N=2$.

2. (Completeness) Prove, or disprove, that the homogeneous $T-Q$ relation together with its Bethe-ansatz equations and the sum rule yields the complete spectrum of the transfer matrix $t(u)$ and of the Hamiltonian $H$ — every eigenvalue, with no missing states.

3. (Singular solutions) Prove, or disprove, that for every $N\geq4$ the Bethe-ansatz equations admit physical singular solutions in which two roots form the bound pair $\lambda_1=\eta/2$, $\lambda_2=-\eta/2$ while the remaining $M-2$ roots $\nu_j\equiv\lambda_{j+2}$ satisfy the reduced Bethe-ansatz equations

$e^{2\beta\gamma}\left[\frac{\theta_1(\nu_j+\eta/2)}{\theta_1(\nu_j-\eta/2)}\right]^{N-1}\frac{\theta_1(\nu_j-3\eta/2)}{\theta_1(\nu_j+3\eta/2)}\prod_{k\neq j}^{M-2}\frac{\theta_1(\nu_j-\nu_k-\eta)}{\theta_1(\nu_j-\nu_k+\eta)}=1,\qquad j=1,2,\dots,M-2,$

together with the selection rule $\sin\!\big[\big(2\sum_{j=1}^{M-2}\nu_j-\beta\tau\big)\pi\big]=0$ (the source counts 1, 4 and 11 such solutions at $N=4,6,8$, and notes that the naive Bethe vector vanishes on them and must be rescaled).

4. (Scalar products) Prove, or disprove, that the determinant formulas for scalar products of Bethe vectors, established at roots of unity, persist for generic anisotropy when the Bethe roots and $\beta$ satisfy the Bethe-ansatz equations and the sum rule (the source conjectures this through the root-of-unity limit $P\to\infty$).

5. (Odd-$N$ Bethe vectors) Construct the Bethe vectors of the generic periodic XYZ chain with an odd number of sites, whose transfer-matrix eigenvalues are already given by the inhomogeneous $T-Q$ relation of the off-diagonal Bethe ansatz.

A refutation of an individual conjecture by an explicit finite counterexample — for instance, a specific chain and an eigenvalue of $t(u)$ not reproduced by any Bethe-ansatz solution satisfying the Bethe-ansatz equations and the sum rule, or a physical solution with $\beta\neq p$ — is an admissible answer for that component and settles it negatively, but does not by itself resolve the remaining components. The problem is completely resolved when each of components 1–4 is proved or refuted and the odd-$N$ Bethe vectors of component 5 are constructed. The source's closing remark that fast, reliable numerical solvers for the elliptic Bethe-ansatz equations would facilitate exploration is a methodological suggestion, not a required deliverable of this problem.

References

1. Xin Zhang, "Hypotheses regarding Baxter's $T-Q$ relation for the periodic XYZ chain", arXiv:2404.12615 (2024), DOI 10.48550/arXiv.2404.12615, https://arxiv.org/abs/2404.12615
2. Rodney J. Baxter, "One-dimensional anisotropic Heisenberg chain", Annals of Physics 70, 323-337 (1972), DOI 10.1016/0003-4916(72)90270-9, https://doi.org/10.1016/0003-4916(72)90270-9
3. R. J. Baxter, "Completeness of the Bethe ansatz for the six and eight vertex models", Journal of Statistical Physics 108, 1-48 (2002), arXiv:cond-mat/0111188, DOI 10.1023/A:1015437118218, https://arxiv.org/abs/cond-mat/0111188
4. Xin Zhang, Andreas Klümper, Vladislav Popkov, "Pedestrian's way to Baxter's Bethe ansatz for the periodic XYZ chain", Physical Review B 109, 115411 (2024), arXiv:2312.00161, DOI 10.1103/PhysRevB.109.115411, https://arxiv.org/abs/2312.00161
5. N. Slavnov, A. Zabrodin, A. Zotov, "Scalar products of Bethe vectors in the 8-vertex model", Journal of High Energy Physics 2020(6), 123 (2020), arXiv:2005.11224, DOI 10.1007/JHEP06(2020)123, https://arxiv.org/abs/2005.11224
6. G. Kulkarni, N. A. Slavnov, "Scalar products of Bethe vectors in the generalized algebraic Bethe ansatz", Theoretical and Mathematical Physics 217, 1574-1594 (2023), arXiv:2306.12932, DOI 10.1134/S0040577923100100, https://arxiv.org/abs/2306.12932
7. G. Kulkarni, N. A. Slavnov, "Action of the monodromy matrix entries in the generalized algebraic Bethe ansatz", arXiv:2303.02439 (2023), https://arxiv.org/abs/2303.02439
8. Junpeng Cao, Shuai Cui, Wen-Li Yang, Kangjie Shi, Yupeng Wang, "Spin-1/2 XYZ model revisit: general solutions via off-diagonal Bethe ansatz", Nuclear Physics B 886, 185-201 (2014), arXiv:1307.0280, DOI 10.1016/j.nuclphysb.2014.06.026, https://arxiv.org/abs/1307.0280
9. Zhirong Xin, Junpeng Cao, Wen-Li Yang, Yupeng Wang, "Exact physical quantities of the XYZ spin chain in the thermodynamic limit", Journal of High Energy Physics 2025(2), 086 (2025), arXiv:2411.12200, DOI 10.1007/JHEP02(2025)086, https://arxiv.org/abs/2411.12200
10. Rafael I. Nepomechie, Chunguang Wang, "Algebraic Bethe ansatz for singular solutions", Journal of Physics A: Mathematical and Theoretical 46, 325002 (2013), DOI 10.1088/1751-8113/46/32/325002, https://doi.org/10.1088/1751-8113/46/32/325002
11. Wenrui Hao, Rafael I. Nepomechie, Andrew J. Sommese, "Completeness of solutions of Bethe's equations", Physical Review E 88, 052113 (2013), DOI 10.1103/PhysRevE.88.052113, https://doi.org/10.1103/PhysRevE.88.052113
12. G. Niccoli, V. Terras, "The open XYZ spin 1/2 chain: Separation of Variables and scalar products for boundary fields related by a constraint", arXiv:2402.04112 (2024), https://arxiv.org/abs/2402.04112
13. Paul Fendley, Sascha Gehrmann, Eric Vernier, Frank Verstraete, "XYZ Integrability the Easy Way", arXiv:2511.04674 (2025), https://arxiv.org/abs/2511.04674
14. Xin Zhang, Tomaz Prosen, Vladislav Popkov, "Exact nonequilibrium steady states of boundary driven circuit with XYZ gates", arXiv:2605.17018 (2026), https://arxiv.org/abs/2605.17018
15. Xin Qian, Xin Zhang, "Generalized Integrable Boundary States in XXZ and XYZ Spin Chains", arXiv:2601.16605 (2026), https://arxiv.org/abs/2601.16605
16. Shizhang Wu, Shu Chen, Junpeng Cao, Xin Zhang, "Bethe-root configurations and spectral degeneracy in the open XXZ chain with degenerate boundaries", arXiv:2608.20815 (2026), https://arxiv.org/abs/2608.20815