For two-dimensional lattice models, the star-triangle relation — a form of the Yang-Baxter equation (YBE) — is the classical sufficient condition under which a family of row-to-row transfer matrices T(λ)T(\lambda) commutes, which is the hallmark of exact solvability. Whether it is also necessary, i.e. whether [T(λ),T(λ′)]=0[T(\lambda),T(\lambda')]=0 forces the local Boltzmann weights onto a star-triangle manifold, has been repeatedly acknowledged as open (McCoy, Advanced Statistical Mechanics, p. 418). Schmidt (2012) proved necessity for the six-vertex model by promoting the operator coalgebra of the monodromy-matrix entries to a bialgebra through quadratic 'zero-operators', one of which is the transfer-matrix commutator itself: the coproduct recursions close only when the two weight triples share the same anisotropy Δ=(a2+b2−c2)/(2ab)\Delta=(a^2+b^2-c^2)/(2ab). The surviving problem is to establish a necessity theorem, or exhibit a non-degenerate counterexample, for model classes beyond the six-vertex case (eight-vertex and other two-state vertex models, odd eight-vertex, face models). The same source further discusses two auxiliary questions that it does not impose as part of that problem: whether the six-vertex zero-operator bialgebra extends to a Hopf (antipode) or quasi-triangular (R-matrix) structure — a step the source itself calls unnecessary for a complete quadrature — and whether the zero-operator method generalizes from the five-vertex, six-vertex and hexagonal-dimer cases to further models such as the eight-vertex model, including solvability of the unwanted-term-elimination equations of the algebraic Bethe ansatz, which the source calls a separate issue.

Background

A two-dimensional vertex model assigns Boltzmann weights to the allowed configurations of arrows on the four edges meeting at a lattice site. In the six-vertex model each edge carries one of two arrow states and only the six 'ice-rule' configurations (two arrows into the site, two out) are allowed; they receive weights aa, bb, cc (aa and bb for the two arrow-preserving configurations, cc for the two arrow-reversing ones), and the triple λ=(a,b,c)\lambda=(a,b,c) plays the role of a spectral parameter. The row-to-row transfer matrix T(λ)T(\lambda) is built from the 2×22\times 2 weight matrix M(λ)M(\lambda) by an iterated matrix coproduct: tensor product over horizontal indices, matrix product over vertical indices. Writing the resulting monodromy matrix in block form with operator entries AA, BB, CC, DD, one has T=A+DT=A+D, and the coproduct satisfies Δ(A)=A⊗A+B⊗C\Delta(A)=A\otimes A+B\otimes C, Δ(B)=A⊗B+B⊗D\Delta(B)=A\otimes B+B\otimes D, and so on — a coassociative coalgebra structure on the operator algebra (coassociative meaning that the two ways of iterating the coproduct, (Δ⊗id)Δ(\Delta\otimes\mathrm{id})\Delta and (id⊗Δ)Δ(\mathrm{id}\otimes\Delta)\Delta, coincide). A model is (transfer-matrix) integrable if the family T(λ)T(\lambda) commutes, [T(λ),T(λ′)]=0[T(\lambda),T(\lambda')]=0 for different spectral parameters, because commuting transfer matrices can be diagonalized simultaneously, yielding exact thermodynamics and correlation functions. The classical sufficient condition is the star-triangle relation: if an invertible matrix RR intertwines two monodromy matrices, R M(λ)M(λ′)=M(λ′)M(λ) RR\,M(\lambda)M(\lambda')=M(\lambda')M(\lambda)\,R, then taking traces gives commutativity, and the intertwining follows when RR satisfies the Yang-Baxter equation (YBE), the cubic relation R12R13R23=R23R13R12R_{12}R_{13}R_{23}=R_{23}R_{13}R_{12} guaranteeing consistent factorization of many-body processes. For the six-vertex model, eliminating the auxiliary parameter from the YBE collapses the condition on the weights to a single constraint: the anisotropy Δ=(a2+b2−c2)/(2ab)\Delta=(a^2+b^2-c^2)/(2ab) must be the same for both parameter sets. This route underlies the standard RTT construction of quantum groups (RTT after Reshetikhin, Takhtajan and Faddeev; a quantum group being a Hopf-algebraic deformation of a classical symmetry algebra), in which the relations among A,B,C,DA,B,C,D generated by a YBE solution make the operator coalgebra a bialgebra (an algebra whose coproduct is an algebra homomorphism); adjoining an antipode (an anti-homomorphism implementing multiplicative inversion) makes it a Hopf algebra, and a compatible R-matrix makes it quasi-triangular. The algebraic Bethe ansatz (ABA) then diagonalizes TT by acting with BB-operators on a vacuum vector and fixing the Bethe vector's spectral parameters so that the 'unwanted terms' in the action of TT cancel. Whether the star-triangle relation is also necessary for commutativity is a separate matter, acknowledged as open at least since McCoy's monograph (p. 418) and restated by Schmidt's 2012 paper, which approached it without the YBE: requiring the coproduct to be an algebra homomorphism forces a closed set of quadratic 'zero-operators' — operator products annihilating the entire physical state space, preserved by the coproduct so that annihilation at one site propagates to arbitrarily long rows — and one of these zero-operators is the commutator [T,T′][T,T'] itself. For the six-vertex model the recursions close only when Δ=Δ′\Delta=\Delta', which proves necessity for that model; the same framework reproduces the known bialgebra structures of the five-vertex model (the six-vertex descendant with one weight zero, equivalent to the dimer model on the hexagonal lattice, in which diatomic tiles occupy lattice edges so that every vertex is touched by exactly one tile), where MM is singular so no intertwiner of the form R=M(μ)R=M(\mu) exists, and of several smaller Hopf and Yangian-type algebras (the Yangian being the deformation of the algebra of polynomial currents on a symmetry algebra that underlies rational Yang-Baxter solutions). What remains is the general question for other models, plus the algebraic loose ends the paper lists: whether its bialgebra admits an antipode or an R-matrix, and whether the zero-operator method extends to models such as the eight-vertex model, where the solvability of the unwanted-term-elimination equations of the ABA is untested. One scope caveat is essential when reading the question: with no qualifications it admits trivial commuting families — for instance simultaneously diagonalizable weight matrices, for which every transfer matrix is diagonal (and the identity is a trivial intertwiner) — so the substantive question concerns non-degenerate models and must state which target is meant: the weights themselves lying on a star-triangle manifold, or the existence of some gauge-equivalent local intertwiner satisfying the YBE (gauge equivalence meaning equality up to invertible changes of basis in the auxiliary and lattice spaces at each site).

Problem Statement

Let M(λ)M(\lambda) be the local weight matrix of a two-dimensional vertex model and T(λ)T(\lambda) its row-to-row transfer matrix, built from M(λ)M(\lambda) by the iterated matrix coproduct. The star-triangle (Yang-Baxter) relation is the classical sufficient condition for the commuting family [T(λ),T(λ′)]=0[T(\lambda),T(\lambda')]=0 (Baxter), and Schmidt (2012) proved it necessary for the six-vertex model: the commutator [T(λ),T(λ′)][T(\lambda),T(\lambda')] belongs to the closed set of quadratic zero-operators of the operator coalgebra exactly when the two weight triples share the same anisotropy Δ=(a2+b2−c2)/(2ab)\Delta=(a^2+b^2-c^2)/(2ab). The open problem, as posed in the source literature, is to settle this question beyond the six-vertex case: for other two-dimensional lattice models — beginning with the eight-vertex model and other two-state vertex models, the odd eight-vertex model (the variant whose allowed vertex configurations carry an odd number of arrows into each site, breaking the even parity of the ice rule), and interaction-round-a-face models (face models, in which Boltzmann weights are assigned to the spin configurations around a square plaquette rather than to arrow configurations at a vertex) — determine whether commutativity of the transfer-matrix family for all lattice sizes forces the local weights to satisfy a star-triangle (Yang-Baxter) relation, either directly on the weights or through the existence of a local intertwiner R-matrix satisfying the Yang-Baxter equation up to gauge equivalence. An acceptable resolution is either a rigorous necessity theorem for one of the model classes listed above, with explicit hypotheses excluding degenerate cases, or an explicit non-degenerate counterexample for one of those classes: a commuting family whose weights admit no such local Yang-Baxter structure. Here a resolution for any one of the listed classes settles the task while a class narrower than every listed one does not, and a case is degenerate — hence excluded from both branches — when commutativity is trivial, that is, when the weight matrices are simultaneously diagonalizable or gauge-equivalent to simultaneously diagonalizable weights, or when every transfer matrix T(λ)T(\lambda) is a scalar multiple of the identity. The same source additionally discusses two auxiliary algebraic questions that it does not impose as part of this open problem: whether the six-vertex operator bialgebra produced by the zero-operator construction extends to a Hopf algebra (antipode) or a quasi-triangular structure (R-matrix) — steps the source itself describes as not needed for a complete quadrature, noting that the standard RTT realization carries these structures after localization at the quantum determinant (the central grouplike element playing the role of the determinant of the monodromy matrix, whose inversion supplies the antipode) while the singular-MM five-vertex and dimer bialgebras were reported to possess no antipode — and whether the zero-operator method extends beyond the five-vertex, six-vertex and hexagonal-dimer cases to further models such as the eight-vertex model, where the source calls the solvability of the algebraic-Bethe-ansatz unwanted-term-elimination equations a separate issue. Work on these auxiliary questions is relevant context, but the open problem is the necessity question stated above, and it may be attacked by any mathematically rigorous means.

References

  1. Jeffrey R. Schmidt, 'From Coalgebra to Bialgebra for the Six-Vertex Model: The Star-Triangle Relation as a Necessary Condition for Commuting Transfer Matrices', Axioms 1(2):186-200, 2012, DOI 10.3390/axioms1020186, https://doi.org/10.3390/axioms1020186
  2. Barry M. McCoy, 'Advanced Statistical Mechanics', Oxford University Press, 2010 (online edition 2009), DOI 10.1093/acprof:oso/9780199556632.001.0001, https://doi.org/10.1093/acprof:oso/9780199556632.001.0001 (the open problem is attributed to p. 418 of this book by Schmidt, 2012)
  3. Rodney J. Baxter, 'Exactly Solved Models in Statistical Mechanics', Academic Press, London, 1982 (Dover reprint, ISBN 9780486462714), https://store.doverpublications.com/products/9780486462714 (classical sufficiency of the star-triangle relation for commuting transfer matrices)
  4. Jeffrey R. Schmidt, 'The algebraic Bethe Ansatz without the Yang-Baxter equation', Canadian Journal of Physics 86(10):1177-1193, 2008, DOI 10.1139/p08-061, https://doi.org/10.1139/p08-061
  5. Jeffrey R. Schmidt, 'Spectrum-generating bialgebra of the hexagonal-lattice dimer model', Canadian Journal of Physics 87(10):1099-1125, 2009, DOI 10.1139/P09-064, https://doi.org/10.1139/P09-064
  6. Mizuki Sanatani, Naoto Shiraishi, Fuga Ishii, 'A Simple Necessary and Sufficient Condition for Yang-Baxter Integrability', arXiv:2607.29660, 2026, https://arxiv.org/abs/2607.29660
  7. Marius de Leeuw, Chiara Paletta, Anton Pribytok, Ana L. Retore, Paul Ryan, 'Yang-Baxter and the Boost: splitting the difference', SciPost Physics 11(3):069, 2021, DOI 10.21468/scipostphys.11.3.069, https://doi.org/10.21468/scipostphys.11.3.069
  8. Shahane A. Khachatryan, 'Integrability in [d+1] dimensions: combined local equations and commutativity of the transfer matrices', arXiv:2210.16753, 2022, https://arxiv.org/abs/2210.16753
  9. Pramod Padmanabhan, Somnath Maity, Vladimir Korepin, 'Scalene Yang-Baxter triples as a source of hidden symmetries beyond the ordinary Yang-Baxter equation', arXiv:2608.09081, 2026, https://arxiv.org/abs/2608.09081
  10. Paul Fendley, Sascha Gehrmann, Eric Vernier, Frank Verstraete, 'XYZ integrability the easy way', arXiv:2511.04674, 2025, https://arxiv.org/abs/2511.04674
  11. M. J. Martins, 'Integrability of the odd eight-vertex model with symmetric weights', Journal of Physics A: Mathematical and Theoretical 51(22):225002, 2018, DOI 10.1088/1751-8121/aab9cc, https://doi.org/10.1088/1751-8121/aab9cc
  12. Florin Nichita, 'Yang-Baxter Systems, Algebra Factorizations and Braided Categories', Axioms 2(3):437-442, 2013, DOI 10.3390/axioms2030437, https://doi.org/10.3390/axioms2030437
  13. Valentin Buciumas, 'Quantum groups obtained from solutions to the parametrized Yang-Baxter equation', arXiv:1602.04262, 2016, https://arxiv.org/abs/1602.04262