Prove or disprove that the boost condition $[Q_2,Q_3]=0$ is equivalent to Yang–Baxter solvability for non-difference-form and non-diagonalizable nearest-neighbour spin chains

For regular nearest-neighbour quantum spin chains, de Leeuw, Paletta, Pribytok, Retore and Ryan (SciPost Phys. 11, 069 (2021)) observed that every Yang–Baxter solution they classified is characterized by a single Hamiltonian-level condition, $[Q_2,Q_3]=0$, where $Q_2$ is the Hamiltonian and $Q_3$ is the charge generated from it by the boost operator; they posed as a central open question whether this condition is provably equivalent to the full Yang–Baxter equation. The necessity direction (Yang–Baxter solvability implies $[Q_2,Q_3]=0$) is proven in general, including non-difference form. A 2026 theorem of Sanatani, Shiraishi and Ishii (arXiv:2607.29660) settled sufficiency affirmatively for diagonalizable Hamiltonians in difference form, constructively. The surviving open core — stated explicitly both by the source paper and by the recent literature — is sufficiency for genuinely spectral-parameter-dependent (non-difference-form) densities, the regime containing the Hubbard model and the AdS$_5\times$S$^5$ $S$-matrix, and for non-diagonalizable Hamiltonians. The problem is to prove or refute the equivalence in the remaining regimes. (The source separately raises, as related mathematical questions rather than as part of the equivalence question, a possible closed-form $R$-matrix in terms of the Hamiltonian entries — speculatively, in connection with the difference-form case — and an algebraic-geometric description of the $[Q_2,Q_3]=0$ variety.)

Background

The Yang–Baxter equation (YBE) is the master consistency equation of exactly solvable quantum lattice models. Its unknown, the $R$-matrix, is an operator on $V\otimes V$ (with $V$ a finite-dimensional local Hilbert space) depending on spectral parameters, obeying $R_{12}(u_1,u_2)\,R_{13}(u_1,u_3)\,R_{23}(u_2,u_3)=R_{23}(u_2,u_3)\,R_{13}(u_1,u_3)\,R_{12}(u_1,u_2),$ where the subscripts indicate on which of the three factors of $V\otimes V\otimes V$ the operator acts. An $R$-matrix is called regular if it reduces to the permutation operator $P$ (which swaps the two tensor factors) on the diagonal of spectral-parameter space, $R(u,u)=P$; for a difference-form $R$-matrix this says $R(0)=P$. From a regular solution one builds the transfer matrix $t(u)=\mathrm{tr}_0\,R_{0L}\cdots R_{01}$ on a chain of length $L$; the YBE guarantees $[t(u),t(v)]=0$, and the Hamiltonian is recovered from $R$ at the regularity diagonal, $H_{12}(\theta)=P\,\partial_u R_{12}(u,\theta)\big|_{u=\theta}$, so solving the YBE solves the model. If $R$ depends on the spectral parameters only through the difference $u-v$, the model is of difference form. Genuinely two-parameter (non-difference-form) $R$-matrices are far less understood — the second parameter then acts as a model parameter $\theta$, and the local conserved charges arise as Taylor coefficients of $\log t(u,\theta)$ in powers of $u-\theta$ — and this regime contains two of the most prominent exactly-solved systems known: the Hubbard model (the paradigmatic lattice model of correlated electrons) and the worldsheet $S$-matrix of strings propagating on the AdS$_5\times$S$^5$ spacetime (five-dimensional anti-de Sitter space times a five-sphere), the setting of the AdS/CFT correspondence.

Solving the YBE directly is difficult — it is a large system of coupled functional equations — so an alternative Hamiltonian-level criterion has long been sought. Kulish–Sklyanin (1982), Öttinger–Honerkamp (1982) and Jimbo–Miwa (1984) observed that for a translation-invariant nearest-neighbour density $h$ in difference form, the Reshetikhin condition — a single commutation relation $[h_{12}+h_{23},[h_{12},h_{23}]]=X_{23}-X_{12}$ for some operator $X$ — appears to single out exactly the Yang–Baxter-solvable chains. Grabowski and Mathieu (1995) reformulated this conjecture using the boost operator $B=\partial_\theta+\sum_n n\,h_{n,n+1}$, which recursively generates a tower of local charges $Q_{k+1}=[B[Q_2],Q_k]$ from the Hamiltonian $Q_2=\sum_n h_{n,n+1}$: their Conjecture 2 states that integrability is equivalent to the single relation $[Q_2,Q_3]=0$, where $Q_3$ (physically the energy current) is the first boost-generated charge.

De Leeuw, Paletta, Pribytok, Retore and Ryan (SciPost Phys. 11, 069 (2021)) turned this conjecture into a classification engine. For a density $H_{12}(\theta)$ possibly depending on a parameter $\theta$, they impose $[Q_2(\theta),Q_3(\theta)]=0$ — a system of ordinary differential equations in $\theta$, reducing to cubic polynomial equations in the difference-form case — and then solve the Sutherland equations (the recursive equations for the Taylor coefficients of $R$ obtained by expanding the YBE around the regularity diagonal) with regularity boundary conditions. Every solution they obtained this way, including all regular $4\times4$ $R$-matrices with $\mathfrak{su}(2)\oplus\mathfrak{su}(2)$ symmetry (invariance under two independent copies of the spin-rotation algebra $\mathfrak{su}(2)$), indeed solved the YBE. They proved necessity in general, including non-difference form (their Appendix A): any regular nearest-neighbour $R$-matrix solution of the YBE yields charges with $[Q_2,Q_3]=0$. That all their solutions were characterized by the single condition led them to write that it is unclear why this is the case and that it would be very important to understand and prove these relations.

Recent work has closed part of the conjecture. Zhang (SciPost Phys. 20, 102 (2026)) reconstructed $R$ order-by-order from the Hamiltonian but left sufficiency open. Hokkyo (Phys. Rev. B 114, 055108 (2026)) proved that the Reshetikhin condition implies conservation and mutual commutativity of all standard-boost-generated charges — Grabowski–Mathieu Conjecture 2 in its charge-conservation form — without constructing an $R$-matrix. Shiraishi–Yamaguchi (Phys. Rev. B 113, L241111 (2026)) proved a dichotomy theorem for isotropic chains: the existence of a single nontrivial local conserved charge forces an infinite commuting family. Most decisively, Sanatani, Shiraishi and Ishii (arXiv:2607.29660) proved that for a translation-invariant nearest-neighbour chain with finite-dimensional local space and diagonalizable Hamiltonian, the difference-form Reshetikhin condition is sufficient: it yields, constructively, a regular analytic $R$-matrix solving the difference-form YBE and reproducing the Hamiltonian as the logarithmic derivative of the transfer matrix, and the Sutherland equations are equivalent to the YBE. They explicitly leave open the non-difference-form setting — where models need not satisfy the Reshetikhin condition and charges are generated by generalized boost operators — and the non-diagonalizable case, and they give an algorithm rather than a closed-form $R(H)$.

Problem Statement

Prove or disprove, at the generality stated by the source paper, that for every translation-invariant nearest-neighbour spin chain with finite-dimensional local Hilbert space and density $H_{12}(\theta)$ the single boost condition $[Q_2(\theta),Q_3(\theta)]=0,\qquad Q_2=\sum_n H_{n,n+1}(\theta),\quad Q_3=[B[Q_2],Q_2],\quad B=\partial_\theta+\sum_n n\,H_{n,n+1}(\theta),$ is equivalent to regular Yang–Baxter solvability: the existence of a regular $R$-matrix $R(u,v)$ — an operator on $V\otimes V$ with $\dim V<\infty$, analytic in a neighbourhood of the regularity diagonal and satisfying $R(u,u)=P$ — that satisfies the (generally two-parameter) Yang–Baxter equation and reproduces the Hamiltonian through $H_{12}(\theta)=P\,\partial_u R_{12}(u,\theta)\big|_{u=\theta}$, i.e. as the logarithmic derivative of its transfer matrix. The necessity direction is already proven in general; the open sufficiency directions are: (i) densities $H_{12}(\theta)$ with genuine $\theta$-dependence (non-difference form), where the boost constraint is a system of ODEs in $\theta$ rather than cubic polynomial equations and the target is a two-parameter $R$-matrix — the regime containing the Hubbard model and the AdS$_5\times$S$^5$ $S$-matrix; and (ii) non-diagonalizable Hamiltonians, such as the non-Hermitian models found empirically to be integrable. The difference-form, diagonalizable case has been resolved affirmatively (Sanatani–Shiraishi–Ishii, arXiv:2607.29660) and is not part of what remains to be settled. As related mathematical questions raised by the source — not part of the equivalence to be proved or refuted — it additionally notes the interest of a closed-form expression for $R$ in terms of the Hamiltonian entries (speculatively, in connection with the difference-form cubic equations) and of an algebraic-geometric characterization of the variety cut out by $[Q_2,Q_3]=0$ (cubic polynomial constraints in projective space in the difference-form case).

References

1. M. de Leeuw, C. Paletta, A. Pribytok, A. L. Retore, P. Ryan, 'Yang-Baxter and the Boost: splitting the difference', SciPost Phys. 11, 069 (2021), DOI 10.21468/SciPostPhys.11.3.069, https://arxiv.org/abs/2010.11231
2. M. P. Grabowski, P. Mathieu, 'Integrability test for spin chains', J. Phys. A 28, 4777 (1995), DOI 10.1088/0305-4470/28/17/013, https://doi.org/10.1088/0305-4470/28/17/013
3. P. P. Kulish, E. K. Sklyanin, 'Quantum spectral transform method recent developments', Lecture Notes in Physics 151, 61 (1982), DOI 10.1007/3-540-11190-5_8, https://doi.org/10.1007/3-540-11190-5_8
4. H. C. Öttinger, J. Honerkamp, 'Note on the Yang-Baxter equations for generalized Baxter models', Phys. Lett. A 88, 339 (1982), DOI 10.1016/0375-9601(82)90609-0, https://doi.org/10.1016/0375-9601(82)90609-0
5. M. Jimbo, T. Miwa, 'Some remarks on the differential approach to the star-triangle relation', Lett. Math. Phys. 8, 529 (1984), DOI 10.1007/BF00400983, https://doi.org/10.1007/BF00400983
6. M. Sanatani, N. Shiraishi, F. Ishii, 'A Simple Necessary and Sufficient Condition for Yang–Baxter Integrability' (2026), arXiv:2607.29660, https://arxiv.org/abs/2607.29660
7. A. Hokkyo, 'Integrability from a single conservation law in quantum spin chains', Phys. Rev. B 114, 055108 (2026), DOI 10.1103/y66t-t8zw, https://arxiv.org/abs/2508.20713
8. N. Shiraishi, M. Yamaguchi, 'Dichotomy theorem separating complete integrability and nonintegrability of isotropic spin chains', Phys. Rev. B 113, L241111 (2026), DOI 10.1103/lcn4-tbcm, https://arxiv.org/abs/2504.14315
9. Z. Zhang, 'Bootstrapping the R-matrix', SciPost Phys. 20, 102 (2026), DOI 10.21468/SciPostPhys.20.4.102, https://arxiv.org/abs/2504.17773
10. L. Corcoran, M. de Leeuw, 'All regular 4 x 4 solutions of the Yang-Baxter equation', SciPost Phys. Core 7, 045 (2024), DOI 10.21468/SciPostPhysCore.7.3.045, https://arxiv.org/abs/2306.10423
11. M. de Leeuw, V. Posch, 'All 4 x 4 solutions of the quantum Yang-Baxter equation' (2024), arXiv:2411.18685, https://arxiv.org/abs/2411.18685
12. M. de Leeuw, A. Pribytok, A. L. Retore, P. Ryan, 'New integrable 1D models of superconductivity', J. Phys. A 53, 385201 (2020), DOI 10.1088/1751-8121/aba860, arXiv:1911.01439, https://arxiv.org/abs/1911.01439
13. P. Padmanabhan, S. Maity, V. Korepin, 'Scalene Yang–Baxter triples as a source of hidden symmetries beyond the ordinary Yang–Baxter equation' (2026), arXiv:2608.09081, https://arxiv.org/abs/2608.09081