Fjelstad and Månsson (2012) found that the local terms of the three-state chiral Potts quantum chain regroup into two coupled Temperley–Lieb (TL) algebras, which they abstracted into a one-parameter family of algebras An(α)A_n(\alpha) and A~n(α)\tilde{A}_n(\alpha) whose representations generically yield superintegrable Hamiltonians with Onsager-algebra spectra. They left open the structure theory of these algebras — the dimension for general chain length nn and the classification of the values of α\alpha where the algebra is simple, semisimple, or non-semisimple — as well as the construction of spectral-parameter RR-matrices (Baxterization) from their representations and comparison with the chiral Potts RR-matrix, the proof that the charges generated by their conjectured generalized boost operator are conserved and mutually commuting, and a corner-transfer-matrix interpretation of the generalized boost operator. The subsequent literature has constructed the higher-NN coupled TL algebra and given it a planar-parafermion presentation that still leaves the dimension and representation theory open, and has proven, in a general framework, the mutual commutativity of charges generated by parametrized boost-plus-derivation operators of this type, under hypotheses that remain to be verified for these non-Hermitian, staggered models; the surviving open core is the algebraic structure theory of An(α)/A~n(α)A_n(\alpha)/\tilde{A}_n(\alpha), its Baxterization, the boost-commutativity conjecture for these models, and the generalized corner-transfer-matrix question.

Background

The chiral Potts model is the NN-state, ZN\mathbb{Z}_N-symmetric generalization of the Ising model whose Boltzmann weights are parametrized by points of a high-genus algebraic curve; it is a flagship exactly-solved model precisely because, unlike the six-vertex model, its spectral parameter is not additive — the model lacks the so-called difference property. At the superintegrable point of the corresponding quantum chain the Hamiltonian commutes with an infinite family of conserved charges and carries a hidden action of the Onsager algebra, the infinite-dimensional Lie algebra with generators Am,GmA_m, G_m and relations [Am,Al]=4Gm−l[A_m, A_l] = 4G_{m-l}, [Am,Gl]=2(Am−l−Am+l)[A_m, G_l] = 2(A_{m-l} - A_{m+l}), [Gm,Gl]=0[G_m, G_l] = 0, introduced in Onsager's original solution of the Ising model. The Temperley–Lieb (TL) algebra — generated by eie_i with ei2=δeie_i^2 = \delta e_i, eiei±1ei=eie_i e_{i\pm1} e_i = e_i, and eiej=ejeie_i e_j = e_j e_i for ∣i−j∣>1|i-j|>1, with a diagrammatic interpretation as noncrossing arcs — plays the analogous symmetry role for the Potts and XXZ models, and representations of the TL algebra directly produce solutions of the Yang–Baxter equation (the functional consistency condition behind exact solvability) by the procedure known as Baxterization, i.e. attaching a spectral parameter to the algebra generators.

Fjelstad and Månsson (J. Phys. A 45, 155208, 2012) found that the local terms of the three-state chiral Potts chain Hamiltonian, written with clock operators X,ZX, Z (XZ=ωZXXZ = \omega ZX, ω=e2πi/3\omega = e^{2\pi i/3}), regroup into two intermeshed families ei,fie_i, f_i: the odd-indexed generators are built from XiX_i and the even-indexed ones from ZiZi+1Z_i Z_{i+1}, each family separately satisfies the TL relations with loop weight 3\sqrt{3}, the generators annihilate each other on the same bond (eifi=0e_i f_i = 0), and further quadratic and cubic cross-relations hold. Abstracting these relations yields a one-parameter family of algebras: imposing the TL relations together with quadratic cross-relations gives A~n(α)\tilde{A}_n(\alpha) (with μ=α−α−1\mu = \alpha - \alpha^{-1}, γ=α+α−1\gamma = \alpha + \alpha^{-1} entering the coefficients), and adding a further cross-relation (quadratic with linear and constant terms) gives the smaller algebra An(α)A_n(\alpha), with open-chain variants obtained by dropping the last generators; the chiral Potts case corresponds to α=e−πi/6\alpha = e^{-\pi i/6}. Without the extra relations the coupled TL algebra is infinite-dimensional for n>2n > 2. The structure is only partially known: An(α)A_n(\alpha) is finite-dimensional as long as α≠±i\alpha \neq \pm i and trivial at α=±1\alpha = \pm 1, dim⁡A2(α)=9\dim A_2(\alpha) = 9 for generic α\alpha, and the open algebra A~no(α)\tilde{A}_n^{o}(\alpha) is finite-dimensional for all nn; but for A~n(α)\tilde{A}_n(\alpha) even finite-dimensionality is only conjectured, no dimension formula for general nn is known, and nothing is known about which α\alpha give simple, semisimple, or non-semisimple algebras. The physical interest is that any representation of An(α)A_n(\alpha) produces a nearest-neighbor spin-chain Hamiltonian which, generically in α\alpha (excluding the degenerate values γ=0\gamma = 0 and γ=±2\gamma = \pm 2 (i.e. α=±1\alpha = \pm 1, where the algebra is trivial), the values α2=2±3\alpha^2 = 2 \pm \sqrt{3}, equivalently γ=±6\gamma = \pm\sqrt{6}, where the coefficient α2+α−2−4\alpha^2 + \alpha^{-2} - 4 vanishes, and eight additional roots of an explicit quartic in α2\alpha^2), satisfies the Dolan–Grady relations [A1,[A1,[A1,A0]]]=16[A1,A0][A_1,[A_1,[A_1,A_0]]] = 16[A_1,A_0] and their A0A_0-analogue, and therefore carries an Onsager-algebra symmetry with an Ising-like spectrum. The superintegrable chiral Potts chain, the staggered isotropic XY chain, and the staggered XXZ chain arise as such representations. Beyond these three cases, further explicit physical realizations of the algebras An(α)A_n(\alpha) as quantum chains are of interest, but finding them is not required for the questions below.

Two further strands frame the open questions. First, in Yang–Baxter-solvable chains with the difference property, the conserved charges are generated by a boost operator D0=∑ii hi,i+1D_0 = \sum_i i\, h_{i,i+1} through Qn+1=[D0,Qn]Q_{n+1} = [D_0, Q_n] (Tetelman; Sogo–Wadati). Fjelstad and Månsson exhibited a generalized derivation D=D0+DED = D_0 + D_E, where DE=β∂φ+βˉ∂φˉD_E = \beta\partial_\varphi + \bar{\beta}\partial_{\bar{\varphi}} is a linear combination of derivatives with respect to the Hamiltonian parameters φ,φˉ\varphi, \bar{\varphi}, such that the first charge Q1=[D,H]Q_1 = [D, H] is conserved on the integrable parameter manifold; they verified [Q2,Q1]=0[Q_2, Q_1] = 0 by computer and conjectured, in parallel with Conjecture 2 of Grabowski and Mathieu (J. Phys. A 28, 4777, 1995), that all such charges are conserved and mutually commute. Second, Baxter's corner transfer matrix (CTM) method — the partition function of a quadrant, whose four-fold product reproduces the full transfer matrix — yields order parameters and spectra in the Ising and XYZ models but is blocked for the chiral Potts model by the lack of the difference property (Baxter, J. Phys. A 40, 12577, 2007).

Later work has clarified the landscape. Adderton, Batchelor and Wedrich (J. Phys. A 53, 36LT01, 2020) generalized the coupled TL algebra to the NN-state superintegrable chiral Potts chain (N−1N-1 types of TL generators) and gave a pictorial presentation involving a pole around which loops entangle. Adderton and Batchelor (J. Phys. A 59, 045201, 2026) recast this as the planar parafermion algebra, with a string Fourier transform resolving diagrammatic rotations and a pictorial description of the Hilbert space; they state that the dimension of the coupled TL algebra remains open, conjecturing dim⁡cTL2L(N)=N2L−1\dim \mathrm{cTL}_{2L}(\sqrt{N}) = N^{2L-1} for the even case of their presentation. On the integrability side, Hokkyo (Phys. Rev. B 114, 055108, 2026) proved the Grabowski–Mathieu conjecture in a form covering parametrized Hamiltonians whose charge generator is [BH(λ),⋅]+v⋅∇[B_{H(\lambda)}, \cdot] + v\cdot\nabla with vv a smooth function of the parameters — the Fjelstad–Månsson type D=D0+DED = D_0 + D_E: if the lowest charge is conserved throughout an open parameter family, then all iterated charges are conserved and mutually commute, for Hermitian, translation-invariant nearest-neighbor chains.

Problem Statement

For the coupled Temperley–Lieb algebras defined by Fjelstad and Månsson — the unital associative algebras generated by ei,fie_i, f_i (i=1,…,ni = 1, \dots, n; in the closed type the indices are read modulo nn) subject, with μ=α−α−1\mu = \alpha - \alpha^{-1} and γ=α+α−1\gamma = \alpha + \alpha^{-1} for a parameter α∈C∖{0}\alpha \in \mathbb{C} \setminus \{0\}, to the relations

ei2=γei,fi2=γfi,eiei±1ei=ei,fifi±1fi=fi,eifi±1ei=ei,fiei±1fi=fi,eifi=0=fiei,e_i^2 = \gamma e_i, \quad f_i^2 = \gamma f_i, \quad e_i e_{i\pm1} e_i = e_i, \quad f_i f_{i\pm1} f_i = f_i, \quad e_i f_{i\pm1} e_i = e_i, \quad f_i e_{i\pm1} f_i = f_i, \quad e_i f_i = 0 = f_i e_i,

together with commutativity of any two generators whose indices differ by more than one; the cross-relations

∓μ[ei−fi,ei±1+fi±1]+γ{ei+fi,ei±1−fi±1}+4(fi±1−ei±1)=0,\mp\mu[e_i - f_i, e_{i\pm1} + f_{i\pm1}] + \gamma\{e_i + f_i, e_{i\pm1} - f_{i\pm1}\} + 4(f_{i\pm1} - e_{i\pm1}) = 0, μ{ei−fi,ei−1−fi−1}−γ[ei−1+fi−1,ei+fi]=0,\mu\{e_i - f_i, e_{i-1} - f_{i-1}\} - \gamma[e_{i-1} + f_{i-1}, e_i + f_i] = 0,

(their eqs. (29)–(30)), which define A~n(α)\tilde{A}_n(\alpha); and the additional cross-relation

k1{ei+fi,ei−1+fi−1}+k2[ei−1−fi−1,ei−fi]+k3(ei+fi+ei−1+fi−1)+k4=0,k_1\{e_i + f_i, e_{i-1} + f_{i-1}\} + k_2[e_{i-1} - f_{i-1}, e_i - f_i] + k_3(e_i + f_i + e_{i-1} + f_{i-1}) + k_4 = 0,

with k1=γ(γ2−6)k_1 = \gamma(\gamma^2 - 6), k2=−μ(μ2+2)k_2 = -\mu(\mu^2 + 2), k3=4γ2k_3 = 4\gamma^2 and k4=−8γ(γ2−2)k_4 = -8\gamma(\gamma^2 - 2) (their eq. (31)), which defines An(α)A_n(\alpha); and the open-chain variants of all of these, obtained by dropping the generators en,fne_n, f_n and every relation involving them — determine the structure theory and connect it to standard RR-matrix integrability. Throughout, [a,b]=ab−ba[a, b] = ab - ba denotes the commutator and {a,b}=ab+ba\{a, b\} = ab + ba the anticommutator. Concretely: (1) Structure theory: determine the dimension of An(α)A_n(\alpha) and A~n(α)\tilde{A}_n(\alpha) for general chain length nn, describe their representation theory — irreducible modules and a diagrammatic or combinatorial presentation of the Hilbert-space sectors in the spirit of the TL theory — and classify the parameter values α\alpha for which the algebras are simple, semisimple, or non-semisimple, deciding whether the exceptional values already identified (α=±1\alpha = \pm 1, α=±i\alpha = \pm i, α2=2±3\alpha^2 = 2 \pm \sqrt{3}, and the eight roots of the quartic arising in the Dolan–Grady analysis) delimit genuine changes of algebraic type. (2) Baxterization: construct, from representations of An(α)A_n(\alpha), spectral-parameter RR-matrices satisfying the Yang–Baxter equation and compare them with the known chiral Potts RR-matrix (gauge equivalence, inequivalence, or specialization), thereby connecting the coupled-TL framework to mainstream Yang–Baxter integrability. (3) Corner transfer matrix: decide whether the generalized boost derivation D=D0+DED = D_0 + D_E — if it is established as a suitable generalization of the boost operator, i.e. one whose adjoint action generates the conserved charges — is analogously related to a suitable generalization of Baxter's corner transfer matrix for models lacking the difference property such as the chiral Potts model, the analogy being fixed by Baxter's defining properties: exhibit a quadrant operator whose product over the four quadrants reproduces the transfer matrix (or partition function) of the chain, with the generalized boost derivation realized as the corresponding logarithmic (spectral-parameter) generator and usable to compute order parameters or spectral information — or prove that the difference-property obstruction excludes any such generalization. (4) Boost commutativity: prove or refute that the charges generated by the generalized boost D=D0+DED = D_0 + D_E are all conserved and mutually commute for these models. These four components share one objective — establishing the coupled TL algebras as a usable symmetry framework for the chiral Potts model: the structure theory supplies the representations from which Hamiltonians and, via Baxterization, RR-matrices are built, the boost-commutativity conjecture asks whether the generalized boost derivation genuinely generates the commuting charge hierarchy of the model, and the CTM question asks whether the generalized boost derivation fits into an analogous generalization of the corner-transfer-matrix method that the difference-property obstruction denies the standard technique. All four components are required for a complete resolution: each is an independently open sub-question — as are the dimension, the representation theory, and the simple/semisimple classification listed within (1) — so a result on any one of them settles only that part and leaves the rest open. The conjecture that originally motivated the construction — that the charges generated by the generalized boost D=D0+DED = D_0 + D_E are all conserved and mutually commute — is stated as item (4) above and remains part of the open problem for these models: it has been proven in a general framework by Hokkyo (2026) for parametrized generators of exactly this algebraic form, but under assumptions (Hermiticity, one-site translational invariance, conservation of the lowest charge throughout an open parameter domain) that are not established for the non-Hermitian, staggered chiral Potts families, where for α≠e−πi/6\alpha \neq e^{-\pi i/6} the integrable family lies on a quartic-defined constraint manifold; carrying out this transfer has not been written down in the literature.

References

  1. Jens Fjelstad, Teresia Månsson, 'New symmetries of the chiral Potts model', J. Phys. A: Math. Theor. 45, 155208 (2012). DOI: 10.1088/1751-8113/45/15/155208; arXiv:1109.6503; https://arxiv.org/abs/1109.6503
  2. M. P. Grabowski, P. Mathieu, 'Integrability test for spin chains', J. Phys. A: Math. Gen. 28, 4777–4798 (1995). DOI: 10.1088/0305-4470/28/17/013; https://doi.org/10.1088/0305-4470/28/17/013
  3. Remy Adderton, Murray T. Batchelor, Paul Wedrich, 'A coupled Temperley–Lieb algebra for the superintegrable chiral Potts chain', J. Phys. A: Math. Theor. 53, 36LT01 (2020). DOI: 10.1088/1751-8121/aba143; arXiv:2004.10392; https://arxiv.org/abs/2004.10392
  4. Remy Adderton, Murray T. Batchelor, 'The planar parafermion algebra: the ZN\mathbb{Z}_N clock model and the coupled Temperley–Lieb algebra', J. Phys. A: Math. Theor. 59, 045201 (2026). DOI: 10.1088/1751-8121/ae36d5; arXiv:2508.01672; https://arxiv.org/abs/2508.01672
  5. Akihiro Hokkyo, 'Integrability from a single conservation law in quantum spin chains', Phys. Rev. B 114, 055108 (2026). DOI: 10.1103/y66t-t8zw; arXiv:2508.20713; https://arxiv.org/abs/2508.20713
  6. R. J. Baxter, 'Corner transfer matrices in statistical mechanics', J. Phys. A: Math. Theor. 40, 12577–12588 (2007). DOI: 10.1088/1751-8113/40/42/s05; https://doi.org/10.1088/1751-8113/40/42/s05
  7. Kazuhiko Minami, 'Onsager algebra and algebraic generalization of Jordan–Wigner transformation', Nucl. Phys. B 973, 115599 (2021). DOI: 10.1016/j.nuclphysb.2021.115599; arXiv:2108.03811; https://arxiv.org/abs/2108.03811
  8. Nick Jones, Abhishodh Prakash, Paul Fendley, 'Pivoting through the chiral-clock family', SciPost Phys. 18, 094 (2025). DOI: 10.21468/SciPostPhys.18.3.094; arXiv:2406.01680; https://arxiv.org/abs/2406.01680
  9. Imogen Camp, Nick G. Jones, 'Matrix-Product State Skeletons in Onsager-Integrable Quantum Chains', J. Stat. Phys. 193, 73 (2026). DOI: 10.1007/s10955-026-03637-8; arXiv:2511.07212; https://arxiv.org/abs/2511.07212