Completing the exact solution of the Izergin-Korepin Gaudin model: Bethe-root structure, eigenstates, and completeness

The Izergin-Korepin (IK) Gaudin model - a long-range integrable spin system obtained as the quasi-classical limit of the 19-vertex $A_2^{(2)}$ transfer matrix - was solved at the level of eigenvalues in 2025 by Xu, Sun, Zhang, Cao and Yang (SciPost Phys. Core 8, 041), who derived the eigenvalues of the commuting Gaudin operators and the corresponding Bethe ansatz equations (BAEs) for periodic boundaries and for constrained non-diagonal and diagonal open boundaries. The solution is explicitly incomplete: the authors defer the analysis of the new BAEs (existence of infinite Bethe roots, singular physical solutions, and string-like root patterns), the construction of the eigenstates - obstructed in the non-diagonal open regimes by the absence of $U(1)$ symmetry - and the completeness of the Bethe states together with the explanation of the observed degeneracies. No published result addresses the BAE solution structure, the eigenstates, or the completeness of the IK Gaudin model. The problem - to complete the exact solution of the IK Gaudin model by classifying the BAE solution structure, constructing the eigenstates in all boundary regimes, and establishing completeness - is open.

Background

Gaudin models are a family of exactly solvable quantum spin systems with long-range two-site interactions, introduced by Gaudin in 1976. In such a model a set of $N$ mutually commuting operators $H_1,\dots,H_N$, labeled by site-dependent inhomogeneity parameters $\theta_j$, is diagonalized simultaneously by the Bethe ansatz - a method in which eigenstates are parameterized by complex numbers called Bethe roots, whose values are fixed by Bethe ansatz equations (BAEs). Gaudin models underpin the exactly solvable pairing models of nuclear and condensed-matter physics (the Richardson-Gaudin systems) and provide integral representations of solutions of the Knizhnik-Zamolodchikov equations of two-dimensional conformal field theory.

A systematic construction of Gaudin-type Hamiltonians is the quasi-classical expansion: if an integrable vertex model has an $R$-matrix (a solution of the quantum Yang-Baxter equation that generates a commuting family of transfer matrices) depending on an anisotropy parameter $\eta$, then expanding $R(u)=\sinh u\cdot\mathrm{id}+\eta\,r(u)+o(\eta^2)$ around $\eta=0$ and evaluating the inhomogeneous transfer matrix at the spectral parameter $u=\theta_j$ produces commuting long-range operators built from the classical $r$-matrix; this route was worked out for vertex models with boundaries by Di Lorenzo, Amico, Hikami, Osterloh and Giaquinta. For open (finite, two-sided) systems, integrable boundary conditions are imposed through Sklyanin's reflection equation - the boundary analogue of the Yang-Baxter equation - whose solutions are boundary $K$-matrices.

Most Gaudin models studied so far descend from six-vertex ($A_1^{(1)}$-type) $R$-matrices. The Izergin-Korepin (IK) model is the 19-vertex integrable model associated with the twisted affine algebra $A_2^{(2)}$: each site carries a three-dimensional Hilbert space, and the $9\times 9$ $R$-matrix yields in its quasi-classical limit the classical $A_2^{(2)}$ $r$-matrix. The eigenvalues of the IK transfer matrix are known for periodic boundaries (by the algebraic Bethe ansatz) and for generic non-diagonal open boundaries (by the off-diagonal Bethe ansatz, ODBA, developed by Cao, Yang, Shi and Wang - a method for models without $U(1)$ symmetry, in which the absence of a reference state is circumvented through operator identities between transfer matrices at shifted spectral parameters; the IK open chain was solved this way by Hao, Cao, Li, Yang, Shi and Wang). Here $U(1)$ symmetry means conservation of the total spin $z$-component, whose presence supplies the pseudo-vacuum (reference state) on which the algebraic Bethe ansatz is built.

In this framework Xu, Sun, Zhang, Cao and Yang defined and solved the IK Gaudin model - the Gaudin descendant of the IK vertex model - for periodic and open boundaries. The Gaudin operators are the $O(\eta)$ coefficients of the transfer matrix at $u=\theta_j$. In the periodic case the eigenvalues are

$E_j^{(p)}=\sum_{k\neq j}^{N}\frac{1-5\cosh(\theta_j-\theta_k)}{\sinh(\theta_j-\theta_k)}+2\sum_{k=1}^{M}\coth\frac{\theta_j-\mu_k}{2},$

where the Bethe roots $\{\mu_1,\dots,\mu_M\}$, with $0\le M\le N$ and $M$ interpreted as the number of spinons, satisfy the BAEs

$\sum_{l\neq j}^{M}\left[-2\coth\frac{\mu_j-\mu_l}{2}+\tanh\frac{\mu_j-\mu_l}{2}\right]+\sum_{l=1}^{N}\coth\frac{\mu_j-\theta_l}{2}=0,\qquad j=1,\dots,M.$

For open boundaries the published version solves two regimes. For constrained non-diagonal $K$-matrices - where the boundary parameters obey $e^{\sigma-\sigma'}=e^{-4k\eta}$, $k\in\mathbb{Z}$, so that the inhomogeneous term of the $T$-$Q$ relation vanishes and a homogeneous $T$-$Q$ relation exists - the eigenvalues are given by Eq. (6.1) in terms of Bethe roots $\{\tilde\mu_j\}$ satisfying the BAEs (6.2), and $U(1)$ symmetry remains broken. For diagonal $K$-matrices ($\epsilon\to+\infty$) the eigenvalues are given by Eq. (6.5) in terms of roots $\{\bar\mu_j\}$ satisfying the BAEs (6.6), and $U(1)$ symmetry is recovered; in both regimes the Gaudin roots are $\tilde\mu_j,\bar\mu_j=\lim_{\eta\to0}\cosh\lambda_j$ of the transfer-matrix Bethe roots. The first arXiv version of the paper additionally derived Gaudin-level BAEs for fully generic non-diagonal boundaries, where the Bethe roots carry an $\eta$-expansion $\lambda_j=\lambda_j^{(0)}+\eta\lambda_j^{(1)}+\cdots$ and the BAEs become coupled algebraic systems for the coefficients $\lambda_j^{(0)}$ and $\lambda_j^{(1)}$ (Eqs. (6.7)-(6.8) of arXiv:2412.04761v1); the published version omits these and treats the generic regime only at the transfer-matrix level (inhomogeneous $T$-$Q$ relation, Eq. (4.14), and BAEs, Eq. (4.18), of arXiv:2412.04761v2). The authors verified the eigenvalue formulas against exact diagonalization for $N=3$ and recorded the degeneracy structure - in the periodic case all eigenstates in the $M\neq N$ sectors are degenerate due to a $Z_2$ symmetry.

Two standard phenomena of Bethe-ansatz solutions delimit what remains open. First, besides regular finite root sets, BAEs can admit nonstandard solutions: string solutions, in which roots form regular patterns in the complex plane, and infinite (singular) Bethe roots, which diverge to infinity yet may still label physical eigenstates; such solutions are often essential for completeness, i.e. for the statement that the Bethe eigenstates span the full $3^N$-dimensional Hilbert space. Completeness of Bethe ansatz solutions is established only case by case - for instance for the open XXZ chain with non-diagonal boundary terms and for the rational spin-1/2 Richardson-Gaudin system - and no result covers trigonometric non-A-type Gaudin BAEs. Second, the eigenstates themselves: with $U(1)$ symmetry they are constructed as Bethe vectors by the algebraic Bethe ansatz (carried out for the rational $sl(2)$ Gaudin model with boundary, and for the trigonometric $sl(2)$ Gaudin model with triangular boundary $K$-matrices, which still preserve $U(1)$), whereas in regimes whose boundaries break $U(1)$ symmetry one needs separation of variables - Sklyanin's construction of eigenstates from eigenvalues of commuting transfer matrices rather than from creation operators acting on a vacuum - or a generalization of the Bethe ansatz, neither of which has been developed for any trigonometric Gaudin model in such regimes.

Problem Statement

The problem is to complete the exact solution of the Izergin-Korepin (IK) Gaudin model constructed by Xu, Sun, Zhang, Cao and Yang (SciPost Phys. Core 8, 041 (2025)) - one objective whose parts jointly finish a spectrum-only result - by resolving, for each boundary regime in which that work derives Gaudin-level eigenvalues and BAEs (periodic; constrained non-diagonal open; diagonal open; and the generic non-diagonal open regime, whose Gaudin-level eigenvalues and BAEs appear only in the first arXiv version, Eqs. (6.6)-(6.8) of arXiv:2412.04761v1):

(1) Solution structure of the new Gaudin BAEs. Classify the solutions of the periodic BAEs (Eq. (3.22) of arXiv:2412.04761v2), of the constrained-boundary BAEs (Eq. (6.2) of the published version) and of the diagonal-boundary BAEs (Eq. (6.6)), and - for the generic non-diagonal regime - of the coupled systems for the root-expansion coefficients $\lambda_j^{(0)}$ and $\lambda_j^{(1)}$ (Eqs. (6.7)-(6.8) of arXiv:2412.04761v1): determine which solution types exist - in particular nonstandard types such as infinite Bethe roots, singular physical solutions, and string-like patterns of roots - and what role they play in the spectrum.

(2) Eigenstates. Construct the explicit eigenstates (wavefunctions) of the commuting IK Gaudin operators that realize the eigenvalue formulas of the source (arXiv:2412.04761v2 Eqs. (3.21), (6.1) and (6.5), and arXiv:2412.04761v1 Eq. (6.6) for the generic non-diagonal regime). In regimes retaining $U(1)$ symmetry (for example periodic and diagonal-open boundaries) this is expected to be achievable by the algebraic Bethe ansatz; for boundary regimes where $U(1)$ symmetry is broken and no pseudo-vacuum (reference state) exists - the constrained non-diagonal open case of the published version, and the generic non-diagonal case - the construction requires adapting or generalizing existing methods (for example separation of variables or a modified Bethe ansatz) to this model.

(3) Completeness and degeneracy. Determine whether the Bethe states admitted by the BAEs - including any nonstandard solutions found in (1) - span the full $3^N$-dimensional Hilbert space in each boundary regime (counting the Bethe states over the root-number sectors admissible in each regime: as in the source, $M$ is fixed by the boundary parameters in the constrained non-diagonal regime and adjustable, $M=0,1,\dots,2N$, in the diagonal regime), i.e. whether the quasi-classical construction captures all eigenstates of the Gaudin operators; and account for the degeneracy pattern observed in exact diagonalization, in particular the $Z_2$-related degeneracy of the $M\neq N$ sectors in the periodic case.

These parts form one program with a single completion boundary: (1) supplies the admissible root configurations, (2) supplies the states attached to them, and (3) certifies that spectrum and states together exhaust the model.

References

1. Xiaotian Xu, Pei Sun, Xin Zhang, Junpeng Cao, Tao Yang, 'Exact solution of the Izergin-Korepin Gaudin model with periodic and open boundaries', SciPost Phys. Core 8, 041 (2025), DOI: 10.21468/SciPostPhysCore.8.2.041, arXiv:2412.04761, https://arxiv.org/abs/2412.04761
2. M. Gaudin, 'Diagonalisation d'une classe d'hamiltoniens de spin', Journal de Physique 37, 1087 (1976), DOI: 10.1051/jphys:0197600370100108700, https://doi.org/10.1051/jphys:0197600370100108700
3. A. G. Izergin, V. E. Korepin, 'The inverse scattering method approach to the quantum Shabat-Mikhailov model', Communications in Mathematical Physics 79, 303 (1981), DOI: 10.1007/BF01208496, https://doi.org/10.1007/BF01208496
4. E. K. Sklyanin, 'Boundary conditions for integrable quantum systems', Journal of Physics A 21, 2375 (1988), DOI: 10.1088/0305-4470/21/10/015, https://doi.org/10.1088/0305-4470/21/10/015
5. J. Cao, W.-L. Yang, K. Shi, Y. Wang, 'Off-Diagonal Bethe Ansatz and Exact Solution of a Topological Spin Ring', Physical Review Letters 111, 137201 (2013), DOI: 10.1103/PhysRevLett.111.137201, https://doi.org/10.1103/PhysRevLett.111.137201
6. K. Hao, J. Cao, G.-L. Li, W.-L. Yang, K. Shi, Y. Wang, 'Exact solution of the Izergin-Korepin model with general non-diagonal boundary terms', Journal of High Energy Physics 2014, 128 (2014), DOI: 10.1007/JHEP06(2014)128, https://doi.org/10.1007/JHEP06(2014)128
7. Kun Hao, Junpeng Cao, Tao Yang, Wen-Li Yang, 'Exact solution of the XXX Gaudin model with generic open boundaries', Annals of Physics 354, 401 (2015), DOI: 10.1016/j.aop.2015.01.007, https://doi.org/10.1016/j.aop.2015.01.007
8. A. Di Lorenzo, L. Amico, K. Hikami, A. Osterloh, G. Giaquinta, 'Quasi-classical descendants of disordered vertex models with boundaries', Nuclear Physics B 644, 409 (2002), DOI: 10.1016/S0550-3213(02)00811-8, https://doi.org/10.1016/S0550-3213(02)00811-8
9. R. I. Nepomechie, F. Ravanini, 'Completeness of the Bethe Ansatz solution of the open XXZ chain with nondiagonal boundary terms', Journal of Physics A 36, 11391 (2003), DOI: 10.1088/0305-4470/36/45/003, https://doi.org/10.1088/0305-4470/36/45/003
10. Jon Links, 'Completeness of the Bethe states for the rational, spin-1/2 Richardson-Gaudin system', SciPost Phys. 3, 007 (2017), DOI: 10.21468/SciPostPhys.3.1.007, arXiv:1603.03542, https://arxiv.org/abs/1603.03542
11. N. Cirilo António, N. Manojlović, E. Ragoucy, I. Salom, 'Algebraic Bethe ansatz for the sl(2) Gaudin model with boundary', Nuclear Physics B 893, 305 (2015), DOI: 10.1016/j.nuclphysb.2015.02.011, arXiv:1412.1396, https://arxiv.org/abs/1412.1396
12. Pascal Baseilhac, Azat M. Gainutdinov, Guillaume Lemarthe, 'Universal TT- and TQ-relations via centrally extended q-Onsager algebra', Nuclear Physics B 1026, 117427 (2026), DOI: 10.1016/j.nuclphysb.2026.117427, arXiv:2511.15876, https://arxiv.org/abs/2511.15876
13. N. Manojlović, I. Salom, 'Algebraic Bethe ansatz for the trigonometric sl(2) Gaudin model with triangular boundary', Symmetry 12, 352 (2020), DOI: 10.3390/sym12030352, arXiv:1709.06419, https://arxiv.org/abs/1709.06419
14. I. Salom, N. Manojlović, 'Bethe states and Knizhnik-Zamolodchikov equations of the trigonometric Gaudin model with triangular boundary', Nuclear Physics B 969, 115462 (2021), DOI: 10.1016/j.nuclphysb.2021.115462, https://doi.org/10.1016/j.nuclphysb.2021.115462