Asymptotic inversion, quasi-locality, and the string-charge relations for the fused transfer-matrix hierarchy of nested Bethe-ansatz models

Integrable lattice models solved by the nested Bethe ansatz, such as the fundamental SU(N)-symmetric spin chains with N at least 3, are expected to equilibrate to Generalized Gibbs Ensembles (GGE) built from local and quasi-local conserved charges. Fehér and Pozsgay (SciPost Phys. 8, 034 (2020)) showed that a complete candidate set of charges for these models is generated by the fusion hierarchy of commuting transfer matrices, proved asymptotic inversion and quasi-locality only for the two m=1 cases of SU(3) (fundamental and conjugate auxiliary representations), and derived the string-charge relations - a linear dictionary between the charge eigenvalues and the Bethe-string densities - under the assumption that a designated Young-tableaux term dominates the fused transfer-matrix eigenvalue sums. Their general statements remain unproven: Conjectures 1-5 - asymptotic inversion and quasi-locality for arbitrary fused levels and arbitrary rank, and a uniform SU(N) proof of the string-charge map - together with the dominance of the leading tableaux term for all thermodynamic Bethe macrostates assumed in their derivation (leaving open the existence of exceptional macrostates), and GGE completeness. No proof or refutation of any of these conjectures is known in the literature to date.

Background

One-dimensional integrable quantum systems possess an extensive set of conserved quantities, which prevents ordinary thermalization: after a quench they relax instead to a Generalized Gibbs Ensemble (GGE), a density matrix built from the conserved charges. A GGE is called complete if its charges determine the stationary state uniquely; for Bethe-ansatz-solvable models the natural labels of a stationary state are the thermodynamic densities of Bethe roots, so completeness means that the charge eigenvalues fix all root densities (this is the content of generalized eigenstate thermalization).

The paradigmatic case is the spin-1/2 isotropic Heisenberg (XXX) chain. Its Bethe excitations organize into m-strings - bound-state patterns of $m$ rapidities with imaginary spacings, postulated by the string hypothesis - and it was found that the strictly local charges are insufficient: they do not see the densities of m-strings with $m$ at least 2. The missing quasi-local charges are generated by the fusion hierarchy of transfer matrices. A transfer matrix $t(u)$ is the trace, over an auxiliary space, of a product of R-matrices (solutions of the Yang-Baxter equation) along the chain; it commutes with itself at different spectral parameters $u$, so derivatives of $\log t(u)$ generate conserved charges. Fusing $m$ copies of the fundamental auxiliary representation yields a family of fused transfer matrices labelled by Young diagrams, and the traceless generators $X_m(u) = -i\,\partial_u \log t_m(u)$ are quasi-local: with respect to the Hilbert-Schmidt norm $\|A\|^2_{\mathrm{HS}} = N^{-L}\,\mathrm{Tr}(A^\dagger A)$ on chains of length $L$, their norm squared grows proportionally to $L$ and their overlaps with fixed local operators have a thermodynamic limit. For the XXX chain, quasi-locality of the fused charges is proven analytically only up to spin $s = 3/2$, and the complete GGE and the string-charge duality (a linear dictionary between charge eigenvalues and string densities) were established for SU(2) on this basis.

The present problem concerns nested Bethe-ansatz models: chains with SU(N), $N$ at least 3, symmetry, such as the fundamental SU(3) Lai-Sutherland chain $H = -L + \sum_j P_{j,j+1}$ built from permutation operators on $(\mathbb{C}^3)^{\otimes L}$. "Nested" refers to the multi-level Bethe ansatz needed when the elementary excitations carry internal degrees of freedom: the coordinate-ansatz wavefunction itself must be solved by a further Bethe ansatz, producing $N-1$ distinct species of rapidities (two for SU(3), describing fundamental-type and conjugate-type excitations), each of which can bind into m-strings. Fehér and Pozsgay (2020) proposed the complete candidate set of charges for these models: the generators $X^{(a)}_m(u) = -i\,\partial_u \log t^{(a)}_m(u)$ of the full fusion hierarchy, where $t^{(a)}_m(u)$ are transfer matrices with auxiliary representations labelled by rectangular Young diagrams of shape $a \times m$, with $a = 1,\ldots,N-1$ and $m = 1,2,\ldots$. Under two unproven assumptions they derived the string-charge relations, which fix the densities of m-strings of every rapidity type from the eigenvalues of these charges.

The first assumption is the family of asymptotic inversion relations $\bar{t}^{(a)}_m(-u)\,t^{(a)}_m(u) \approx 1$, where $\bar{t}$ is the space-reflected transfer matrix and the relation is meant as convergence in Hilbert-Schmidt norm as $L$ grows, for $u$ in the physical strip $|\mathrm{Im}\,u| < 1/2$. Inversion is what converts the formally non-local generators into sums of quasi-local densities. It was proven by Fehér and Pozsgay only for the two $m = 1$ families of SU(3) ($a = 1$ and $a = 2$), by computing Hilbert-Schmidt norms through finite-site quantum transfer matrices (QTMs) - transfer matrices of an auxiliary two-dimensional lattice model obtained by viewing the norm computation in the crossed channel - and using their spectral gaps. For general fused levels this route requires diagonalizing two-site and four-site QTMs whose auxiliary spaces carry higher-dimensional fused representations, with a gap uniform in the chain length. The second assumption concerns eigenvalues: $t^{(a)}_m(u)$ acts on Bethe states with eigenvalues given by finite sums over semi-standard Young tableaux (fillings increasing down columns, non-decreasing along rows) of products of Baxter Q-functions (the Q-function of a rapidity species is a function of the spectral parameter whose zeros are the Bethe roots of that species, so each product carries one factor per Bethe root). The string-charge relations follow only if a designated tableau term dominates each sum for every thermodynamic macrostate; the authors proved this selection near infinite-temperature root distributions and verified one explicit quench, but could not exclude exceptional macrostates selecting a different term. The general SU(N) string-charge relations were then conjectured by analogy with the T-system structure (the functional relations satisfied by the fused transfer matrices); these are their Conjectures 1-5.

Problem Statement

Prove or disprove the conjectures left open by Fehér and Pozsgay (SciPost Phys. 8, 034 (2020)) on the conserved charges of nested Bethe-ansatz models, taking the fundamental SU(3)-symmetric spin chain as the primary concrete setting and the fundamental SU(N)-symmetric chain, $N$ at least 3, as the target generality. The common objective is to establish the string-charge relations and the completeness of the Generalized Gibbs Ensemble for these models as theorems; the requested results are scientifically linked steps of one derivation chain, and the problem is settled only when these components are resolved together, each either proven in its stated generality or refuted along the branch given for it below; for component 4, which combines components 1-3, a refutation of any of those components invalidates the string-charge relations and GGE completeness as stated and constitutes the corresponding negative outcome.

1. Asymptotic inversion (Conjectures 1 and 3 of the source): show that for every fused level $m = 1, 2, \ldots$ and every nesting index $a = 1, \ldots, N-1$ the space-reflected fused transfer matrices satisfy

$$\bar{t}^{(a)}_m(-u)\,t^{(a)}_m(u) \approx 1$$

in Hilbert-Schmidt norm as the chain length $L$ grows, for spectral parameters $u$ in the physical strip $|\mathrm{Im}\,u| < 1/2$; or disprove it by exhibiting a fused level and a point of the strip where the convergence fails.

2. Quasi-locality (Conjectures 2 and 4): show that the traceless generators $X^{(a)}_m(u) = -i\,\partial_u \log t^{(a)}_m(u)$ are quasi-local for all fused levels $m$, all nesting indices $a$, and all $u$ in the physical strip, in the sense of the source paper (translational invariance, Hilbert-Schmidt norm squared growing proportionally to $L$, and thermodynamic-limit overlaps with local operators).

3. Dominance and exceptional macrostates: establish for all thermodynamic Bethe macrostates that the designated Young-tableaux term dominates the eigenvalue sums of the fused transfer matrices - the step that converts asymptotic inversion into the string-charge relations - or else characterize the exceptional Bethe root distributions, if any exist, for which a different term dominates. A single exceptional macrostate for which a different term dominates is already the negative outcome resolving this component; a full characterization of the exceptional macrostates, when any exist, is its stronger positive form.

4. SU(N) string-charge relations and completeness (Conjecture 5): combining 1-3, prove the string-charge relations of the SU(N)-symmetric chain - the linear dictionary by which the densities of m-strings of rapidity type $a$ are expressed through the eigenvalues of the quasi-local charges $X^{(a)}_m(u)$ - and hence that the fused-transfer-matrix charges form a complete GGE, so that generalized eigenstate thermalization holds for these nested models.

The concrete technical bottleneck identified by the source is the analytic diagonalization of the finite-site (two-site and four-site) quantum transfer matrices whose auxiliary spaces carry fused representations, together with spectral-gap estimates uniform in the chain length across the physical strip; a solution may proceed by these means or by any alternative technique that establishes the statements above.

References

1. G. Zs. Fehér and B. Pozsgay, "Generalized Gibbs Ensemble and string-charge relations in nested Bethe Ansatz", SciPost Phys. 8, 034 (2020), DOI: 10.21468/scipostphys.8.3.034, arXiv:1909.04470, https://arxiv.org/abs/1909.04470
2. E. Ilievski, M. Medenjak and T. Prosen, "Quasilocal Conserved Operators in the Isotropic Heisenberg Spin-1/2 Chain", Phys. Rev. Lett. 115, 120601 (2015), DOI: 10.1103/PhysRevLett.115.120601, arXiv:1506.05049, https://arxiv.org/abs/1506.05049
3. E. Ilievski, M. Medenjak, T. Prosen and L. Zadnik, "Quasilocal charges in integrable lattice systems", J. Stat. Mech. (2016) 064008, DOI: 10.1088/1742-5468/2016-06-064008, arXiv:1603.00440, https://arxiv.org/abs/1603.00440
4. E. Ilievski, J. De Nardis, B. Wouters, J.-S. Caux, F. H. L. Essler and T. Prosen, "Complete Generalized Gibbs Ensembles in an Interacting Theory", Phys. Rev. Lett. 115, 157201 (2015), DOI: 10.1103/PhysRevLett.115.157201, arXiv:1507.02993, https://arxiv.org/abs/1507.02993
5. E. Ilievski, E. Quinn, J. De Nardis and M. Brockmann, "String-charge duality in integrable lattice models", J. Stat. Mech. (2016) 063101, DOI: 10.1088/1742-5468/2016-06-063101, arXiv:1512.04454, https://arxiv.org/abs/1512.04454
6. C. K. Lai, "Lattice gas with nearest-neighbor interaction in one dimension with arbitrary statistics", J. Math. Phys. 15, 1675 (1974), DOI: 10.1063/1.1666522, https://doi.org/10.1063/1.1666522
7. B. Sutherland, "Model for a multicomponent quantum system", Phys. Rev. B 12, 3795 (1975), DOI: 10.1103/PhysRevB.12.3795, https://doi.org/10.1103/PhysRevB.12.3795
8. M. Borsi, B. Pozsgay and L. Pristyák, "Current operators in integrable models: a review", J. Stat. Mech. (2021) 094001, DOI: 10.1088/1742-5468/ac0f6b, arXiv:2103.12160, https://arxiv.org/abs/2103.12160
9. V. B. Bulchandani, S. Gopalakrishnan and E. Ilievski, "Superdiffusion in spin chains", J. Stat. Mech. (2021) 084001, DOI: 10.1088/1742-5468/ac12c7, arXiv:2103.01976, https://arxiv.org/abs/2103.01976
10. E. Ilievski, J. De Nardis, S. Gopalakrishnan, R. Vasseur and B. Ware, "Superuniversality of Superdiffusion", Phys. Rev. X 11, 031023 (2021), DOI: 10.1103/PhysRevX.11.031023, arXiv:2009.08425, https://arxiv.org/abs/2009.08425
11. A. Hutsalyuk and B. Pozsgay, "Integrability breaking in the one-dimensional Bose gas: Atomic losses and energy loss", Phys. Rev. E 103, 042121 (2021), DOI: 10.1103/PhysRevE.103.042121, arXiv:2012.15640, https://arxiv.org/abs/2012.15640
12. E. Ibarra-García-Padilla and S. Choudhury, "Many-body physics of ultracold alkaline-earth atoms with SU(N)-symmetric interactions", J. Phys.: Condens. Matter 37, 083003 (2024), DOI: 10.1088/1361-648X/ad9658, arXiv:2408.05359, https://arxiv.org/abs/2408.05359
13. F. Møller, B. C. Nagy, M. Kormos and G. Takács, "Anomalous charge transport in the sine-Gordon model", Phys. Rev. B 111, 115121 (2025), DOI: 10.1103/PhysRevB.111.115121, arXiv:2411.11473, https://arxiv.org/abs/2411.11473
14. T. Prosen, "Local and quasilocal conservation laws of three-state IRF cellular automata and their quantum deformations" (2026), arXiv:2608.13080, https://arxiv.org/abs/2608.13080
15. T. Ishiyama, K. Fujimoto and T. Sasamoto, "Integral formula for the propagator of the one-dimensional Hubbard model" (2026), arXiv:2602.21541, https://arxiv.org/abs/2602.21541