Completing the 20-vertex to solid-on-solid correspondence: explicit evaluation of the Boltzmann weight constants

The 20-vertex model — the triangular-lattice analogue of the six-vertex model, with six arrows per site under a three-in/three-out ice rule and $\binom{6}{3}=20$ local configurations — was recently given a candidate solid-on-solid (SOS) counterpart by Rigas (arXiv:2412.04515; Int. J. Mod. Phys. B, 2026), adapting the intertwining-vector method with which Antonenko and Valinevich transformed the rational seven-vertex model into an SOS model and computed its weights as terminating hypergeometric series. In the 20-vertex construction, the SOS Boltzmann weight matrix is obtained only through unevaluated constants $C_1$, $C_2$, $C_3$ and a scalar $\mathscr{C}_1$ defined implicitly via entries of the 20-vertex R-matrix and its factorization into q-exponential mode sums (combinatorial coefficients $u_m$) and K-matrix factors; the solution set $\mathcal{SS}$ of the sixteen-component intertwining system is determined through a general normalized solution whose components are fixed WLOG; the weights' asymptotic dependence on tensor products of basis elements and spectral-parameter shifts is stated qualitatively but not quantified; and neither the face-type Yang-Baxter relation for the weights nor any transfer-matrix spectral, fusion, or thermodynamic property of the resulting SOS model is established. The problem posed here is to complete the decisive step of this vertex-SOS correspondence: evaluate the constants in closed form from the universal R-matrix data, so that the SOS Boltzmann weight matrix and the intertwining vectors become explicit, at the level of explicitness achieved for the rational seven-vertex model. No work evaluating these constants or otherwise completing the correspondence is known.

Background

A vertex model in statistical mechanics assigns Boltzmann weights to local configurations of arrows on lattice edges, constrained by a local rule. In the six-vertex model on the square lattice each vertex has four incident edges and the 'ice rule' (two arrows in, two out) leaves $\binom{4}{2}=6$ admissible configurations; the weights are organized into a $4\times 4$ R-matrix acting on $\mathbb{C}^2\otimes\mathbb{C}^2$, and integrability — commutativity of the row-to-row transfer matrix for all values of a spectral parameter $u$ — follows from the Yang-Baxter equation satisfied by the R-matrix. The 20-vertex model is the triangular-lattice analogue: each vertex has six incident edges, the ice rule requires three arrows in and three out, and $\binom{6}{3}=20$ local configurations are admissible. Its R-matrix acts on $\mathbb{C}^3\otimes\mathbb{C}^3$ and involves three edge directions, so the model is often described as a three-dimensional vertex model; in the zero-field case it is equivalent to edge three-colorings of the square lattice, and its version with domain-wall boundary conditions (all boundary arrows fixed to point either into or out of the domain) has an active combinatorial literature on arctic curves — phase boundaries separating frozen, ordered regions from a disordered central region — and on tiling enumeration (Di Francesco 2021; Debin, Di Francesco, Guitter 2020). Unlike the six-vertex model, no general exact solution of the 20-vertex model is known, so each extension of six-vertex technology to it requires separate justification.

A solid-on-solid (SOS) model, or face model, is the dual formulation in which integer heights live on the faces of the lattice and each local weight depends on the heights surrounding a face. The vertex-SOS (vertex-face) correspondence is a classical construction that transforms a vertex model into a face model through intertwining vectors: vectors $\psi$ depending on a spectral parameter and on adjacent heights, satisfying a linear intertwining relation of the form

$$R(u-v)\,\psi^{a}_{b}(u)\otimes\psi^{b}_{c}(v)=\sum_{b'}\psi^{a}_{b'}(v)\otimes\psi^{b'}_{c}(u)\,W,$$

by matching components on both sides of which the entries of the SOS Boltzmann weight matrix $W$ and the components of the intertwining vectors are determined simultaneously. For the rational seven-vertex model — a rational (XXX-type) R-matrix whose sample space augments the six six-vertex configurations by a seventh — Antonenko and Valinevich carried out this program completely: they derived the seven-vertex universal R-matrix (Valinevich-Antonenko 2021), transformed the model into an SOS model by the intertwining-vector method, computed the resulting SOS Boltzmann weights in closed form as terminating hypergeometric series ${}_9F_8$, and extended them by a fusion procedure (a construction producing higher-spin or composite Yang-Baxter solutions from a given one) and to a further eleven-vertex family (Antonenko-Valinevich 2024).

Rigas (2024a, 2024b) developed a quantum inverse scattering method — the algebraic framework in which the row-to-row transfer matrix is obtained as the trace of a monodromy matrix, a product of local L-operator matrices along a lattice row — for the 20-vertex model, introducing higher-dimensional L-operators based on the factorization of the universal R-matrix due to Boos, Göhmann, Klümper, Nirov and Razumov: a decomposition of the universal R-matrix of the XXZ-type quantum group into q-exponential mode sums (with combinatorial coefficients $u_m$) and a K-matrix factor. Rigas (arXiv:2412.04515, v2 2025; published in the International Journal of Modern Physics B, 2026) then adapts the Antonenko-Valinevich construction to the 20-vertex model: the intertwining relation for the 20-vertex R-matrix is reduced to a system of nine equations for each entry of the R-matrix, each of which decomposes into sixteen component relations $(∗)(1),\dots,(∗)(16)$; the solution set of the full system is denoted $\mathcal{SS}\equiv\bigcup_{1\le i\le16}(∗)(i)$; and a general normalized class of intertwining-vector solutions, with components fixed to 1 WLOG, is obtained explicitly. The main theorem asserts that there exist constants $C_1$, $C_2$, $C_3$ and a scalar $\mathscr{C}_1$ in terms of which the entries of the SOS Boltzmann weight matrix $W$ (indexed by height patterns over the triangular lattice) and the normalized intertwining vectors are expressed. However, the constants are defined only implicitly — as entries of the 20-vertex R-matrix and ratios of such entries, to be evaluated through the universal R-matrix factorization — and are never expanded into closed form; the paper states only qualitatively that the Boltzmann weight matrix 'asymptotically depends upon several contributions', including tensor products of basis elements and spectral-parameter shifts over the triangular lattice; and no property of the resulting SOS model (face-type Yang-Baxter/star-triangle relation, transfer-matrix spectrum, fusion relations, thermodynamics) is derived. This is the state of the art at which the present problem is posed.

Problem Statement

Within the framework of the cited source — the 20-vertex intertwining relation, its reduction to a system of nine equations per entry of the R-matrix, the sixteen-component decompositions $(∗)(1),\dots,(∗)(16)$ with solution set $\mathcal{SS}=\bigcup_{1\le i\le16}(∗)(i)$, and the factorization of the universal R-matrix into q-exponential mode sums with combinatorial coefficients $u_m$, K-matrix factors, and spectral parameters — complete the vertex-to-SOS correspondence for the 20-vertex model, at the level of explicitness achieved for the rational seven-vertex model by Antonenko and Valinevich. Concretely: evaluate the constants $C_1$, $C_2$, $C_3$ and the scalar $\mathscr{C}_1$ in closed form as functions of the universal R-matrix data — the q-exponential mode sums with coefficients $u_m$, the K-matrix parameters, and the spectral and lattice parameters — in terms of which the constants are defined in the source only implicitly, through the $R$-matrix entry $R^{(1,1)}$ and ratios of such entries at shifted arguments (e.g. $C_1\equiv R^{(1,1)}(u)+1$, $C_2\equiv R^{(1,1)}(u+l+w)/R^{(1,1)}(l+w)$ and $C_3\equiv R^{(1,1)}(u(1+l+w))/R^{(1,1)}(l+w)$, where $u$ is the spectral parameter and $l$ and $w$ are the parameters entering the shifted arguments, built in the source from the universal R-matrix data). An evaluation must express each constant explicitly in terms of the mode-sum data rather than restate these implicit definitions; alternatively, provide a constructive, terminating algorithm that computes them in explicit form, with proof of correctness, so that the entries of the SOS Boltzmann weight matrix $W$ and of the normalized intertwining vectors are obtained explicitly rather than through unevaluated constants.

Further directions that the source discusses or suggests but does not pose as open questions — classifying solutions of the sixteen-component system beyond the general normalized solution claimed in the source, quantifying the asymptotic multi-contribution dependence of the weight matrix, verifying (or refuting) the face-type Yang-Baxter (star-triangle) relation for the constructed weights, deriving transfer-matrix spectral, fusion, or thermodynamic properties of the resulting SOS model, and generalizing the method to broader classes of higher-dimensional or higher-rank vertex models — are natural extensions but do not belong to the completion boundary of this problem.

References

1. Pete Rigas: Intertwining vectors, and Boltzmann weight matrices, of a Solid-on-Solid model from the 20-vertex model. International Journal of Modern Physics B (published online 2026-08-14), DOI 10.1142/S0217979226502413; preprint arXiv:2412.04515 (v2, 2025), https://arxiv.org/abs/2412.04515
2. Pavel V. Antonenko, Pavel A. Valinevich: Boltzmann weights and fusion procedure for the rational seven-vertex SOS model. Journal of Mathematical Sciences 284, 582-603 (2024), DOI 10.1007/s10958-024-07374-9; arXiv:2308.10323, https://arxiv.org/abs/2308.10323
3. P. A. Valinevich, P. V. Antonenko: Universal R-Matrix for the Rational Seven-Vertex Model. Journal of Mathematical Sciences 257, 491-494 (2021), DOI 10.1007/s10958-021-05497-x, https://doi.org/10.1007/s10958-021-05497-x
4. Pete Rigas: Quantum inverse scattering for the 20-vertex model up to Dynkin automorphism: 3D Poisson structure, triangular height functions, weak integrability. arXiv:2407.11066 (2024), https://arxiv.org/abs/2407.11066
5. Pete Rigas: The emptiness formation probability, and representations for nonlocal correlation functions, of the 20-vertex model. arXiv:2409.05309 (2024), https://arxiv.org/abs/2409.05309
6. Pete Rigas: Exact solvability of an Ising-type model, and exact solvability of the 6-vertex, and 8-vertex, models. arXiv:2501.01996 (2025), https://arxiv.org/abs/2501.01996
7. Herman Boos, Frank Göhmann, Andreas Klümper, Khazret S. Nirov, Alexander V. Razumov: Universal R-matrix and functional relations. Reviews in Mathematical Physics 26, 1430005 (2014), DOI 10.1142/S0129055X14300052, https://doi.org/10.1142/S0129055X14300052
8. Philippe Di Francesco: Twenty Vertex Model and Domino Tilings of the Aztec Triangle. The Electronic Journal of Combinatorics 28(4) (2021), DOI 10.37236/10227; arXiv:2102.02920, https://arxiv.org/abs/2102.02920
9. Bryan Debin, Philippe Di Francesco, Emmanuel Guitter: Arctic curves of the twenty-vertex model with domain wall boundaries. Journal of Statistical Physics 179, 33-89 (2020), DOI 10.1007/s10955-020-02518-y, https://doi.org/10.1007/s10955-020-02518-y; preprint arXiv:1910.06833 (2019), https://arxiv.org/abs/1910.06833
10. K. Atalikov, A. Zotov: Higher rank generalization of 11-vertex rational R-matrix: IRF-Vertex relations and associative Yang-Baxter equation. Theoretical and Mathematical Physics 216(2), 1083-1103 (2023), DOI 10.1134/S0040577923080019; arXiv:2303.02391, https://arxiv.org/abs/2303.02391