Commuting transfer matrices for the odd eight-vertex model with asymmetric weights: the free-fermion manifold and the staggered two-Lax-operator product

The odd eight-vertex model — the square-lattice vertex model whose allowed configurations carry an odd number of arrows entering each vertex — has an exactly computable thermodynamic free energy under the Fan–Wu free-fermion condition, via Wu and Kunz's equivalence with a staggered even eight-vertex model. Its finite-lattice integrability, however, is understood only for symmetric weights (invariant under arrow reversal), which Martins (2018) showed to share the even model's Baxter manifold through a weight-independent Pauli transformation, uncovering a sheaf of R-matrices on the same elliptic curve. This problem concerns the questions left open for asymmetric weights: (i) within the free-fermion hypersurface, does any algebraic manifold of weights yield commuting transfer matrices on finite lattices — the direct transplant of Krinsky's even-model constraints is already excluded by two-site commutator computations — or is the odd free-fermion model a case of exact thermodynamic solvability without finite-lattice commuting transfer matrices; and (ii) for the staggered odd model, whose partition function is a trace of powers of the product $T_1T_2$ of transfer matrices built from two distinct alternating Lax operators, prove algebraically for arbitrary lattice size that these products commute on the free-fermion plus Krinsky manifolds (currently verified only for $N=1,2$), with the Yang–Baxter equation for the tensor product of the two Lax operators indicated by the source as the contemplated means.

Background

Two-dimensional vertex models place arrows on the edges of a square lattice and assign a Boltzmann weight to each allowed configuration of the four arrows meeting at a vertex. Of the sixteen arrow configurations at a vertex, eight have an even number of arrows pointing into the vertex; the statistical model built on these is the familiar (even) eight-vertex model. The remaining eight configurations have an odd number of incoming arrows, and the model built on them is the odd eight-vertex model, introduced and named by Wu and Kunz. A model is called symmetric when its weights are invariant under simultaneous reversal of all arrows at a vertex ($w_1=w_2$, $w_3=w_4$, $w_5=w_6$, $w_7=w_8$) and asymmetric otherwise; a staggered model assigns different weights to the two checkerboard sublattices of the square lattice.

The row-to-row transfer matrix of a vertex model is the trace, over an auxiliary two-dimensional space, of the ordered tensor product of $4\times 4$ Lax operators, one per site of a row, whose matrix elements are the vertex weights. A family of transfer matrices labeled by the weights commutes precisely when the model is integrable in Baxter's sense, which follows from the Yang–Baxter (intertwiner) relation $R_{12}L_{13}L_{23}=L_{23}L_{13}R_{12}$ for a suitable R-matrix $R$, where the subscripts label which of the three two-dimensional tensor-factor spaces each matrix acts on. For the even model, Baxter proved that with symmetric weights lying on a two-parameter algebraic manifold — the Baxter manifold — the transfer matrices commute, which yields the exact solution. Separately, Fan and Wu showed that when the weights satisfy the free-fermion condition

$$w_1 w_2 + w_3 w_4 - w_5 w_6 - w_7 w_8 = 0$$

the thermodynamic free energy is exactly computable by Pfaffian methods — the determinant-like algebra available when a model maps to noninteracting (free) fermions — even though this condition alone does not produce commuting transfer matrices on finite lattices. Krinsky first found, in the subspace $w_5=w_6$, $w_7=w_8$, additional algebraic constraints that restore finite-lattice commutation; adapted to arbitrary weights they become the three relations $w_6 w_8/(w_5 w_7)=\Delta_1$, $(w_1 w_4+w_2 w_3)/(w_5 w_7)=\Delta_2$ and $(w_1^2+w_4^2-w_2^2-w_3^2)/(w_5 w_7)=\Delta_3$ with fixed parameters $\Delta_1,\Delta_2,\Delta_3$. Galleas and Martins later showed that the asymmetric even model admits exactly three distinct free-fermion integrable branches of this type, one of them previously unknown.

For the odd model, Wu and Kunz established an exact equivalence of partition functions: the uniform odd model equals a staggered even model whose second-sublattice weights are a fixed permutation of the first. Through this equivalence the odd model's free energy is exactly computable in the thermodynamic limit whenever its weights $v_1,\dots,v_8$ satisfy the free-fermion condition $v_1 v_2 + v_3 v_4 - v_5 v_6 - v_7 v_8 = 0$; for positive weights the model exhibits no phase transition. Before 2017 nothing was known about finite-lattice integrability of the odd model: as Martins put it, the possibility of building commuting transfer matrices for finite lattices, either within the free-fermion restriction or on any other manifold, had not been studied in the literature.

Martins settled the symmetric-weight case: the odd Lax operator (which has zero diagonal blocks, reflecting the odd arrow parity) is related to the even one by a weight-independent product of Pauli matrices, so the odd model with symmetric weights has commuting transfer matrices on exactly the same Baxter manifold as the even model. The R-matrix governing the odd model nevertheless has the even (block-diagonal) form, and the four even/odd combinations of R-matrices form a sheaf satisfying a colored-type Yang–Baxter equation — each factor carries an even or odd label playing the role of a color — on Baxter's elliptic curve, the curve parametrizing the weights of the symmetric even model by Jacobi theta functions. Two concrete questions about the odd model with asymmetric weights were left open, and they are the subject of this problem: the existence of a commuting manifold within the free-fermion hypersurface of the uniform asymmetric model, and the algebraic status of the numerically observed commutation of staggered transfer-matrix products. Both rest on the same Wu–Kunz equivalence, since the uniform odd model is a staggered even model and, conversely, the uniform even model is a staggered odd model at the level of partition functions.

Problem Statement

The problem is to determine the finite-lattice integrable structure of the odd eight-vertex model beyond the solved symmetric-weight case: to decide whether its exact thermodynamic solvability is accompanied by commuting transfer matrices and, where it is, to identify the underlying R-matrix/intertwiner algebra. Two linked questions, posed in the concluding remarks of Martins (2018), instantiate this single objective.

(i) The asymmetric free-fermion manifold. Let $L(v_1,\dots,v_8)$ denote the odd-model Lax operator with weights $v_1,\dots,v_8$ (the $4\times 4$ matrix with zero diagonal blocks, given explicitly in part (ii)), and let $T(v_1,\dots,v_8)=\mathrm{Tr}_0[L_{01}\cdots L_{0N}]$ be the corresponding transfer matrix of an $N$-site row. Within the free-fermion hypersurface $v_1 v_2+v_3 v_4-v_5 v_6-v_7 v_8=0$ (weights otherwise arbitrary), decide whether there exists an algebraic manifold of weights of positive dimension — a set defined by polynomial relations on the $v_i$ and containing at least two distinct weight sets, possibly the hypersurface itself, possibly a proper submanifold, and not contained in the symmetric subspace $v_1=v_2$, $v_3=v_4$, $v_5=v_6$, $v_7=v_8$ already solved by Martins — on which the transfer matrices commute pairwise, meaning $T(v)T(v')=T(v')T(v)$ for any two weight sets $v$ and $v'$ on the manifold, for every finite lattice size $N$. The direct transplant of Krinsky's constraints — the three relations quoted in the Background with each $w_i$ replaced by $v_i$, namely $v_6 v_8/(v_5 v_7)=\Delta_1$, $(v_1 v_4+v_2 v_3)/(v_5 v_7)=\Delta_2$ and $(v_1^2+v_4^2-v_2^2-v_3^2)/(v_5 v_7)=\Delta_3$ — is already excluded by explicit two-site commutator computations, but no other candidate manifold has been tested or excluded. An answer may take either form: (a) exhibit such a manifold together with an algebraic proof of commutation valid for arbitrary $N$, ideally through explicit Yang–Baxter intertwiner relations for the odd Lax operators; or (b) prove that no such manifold exists, which would establish that the free-fermion odd eight-vertex model is a model whose free energy is exactly computable in the thermodynamic limit without commuting transfer matrices on finite lattices.

(ii) The staggered two-Lax-operator product. For the staggered odd model with partition function $Z=\mathrm{Tr}[(T_1T_2)^N]$, the transfer matrices of a row of $2N$ sites are built by alternating the two distinct Lax operators

$$L(v_1,\dots,v_8)=\begin{pmatrix}0&v_1&v_7&0\\ v_3&0&0&v_6\\ v_5&0&0&v_4\\ 0&v_8&v_2&0\end{pmatrix},\qquad \mathbf{L}(v_1,\dots,v_8)=\begin{pmatrix}0&v_3&v_6&0\\ v_1&0&0&v_7\\ v_8&0&0&v_2\\ 0&v_5&v_4&0\end{pmatrix},$$

as $T_1=\mathrm{Tr}_0[L_{01}\,\mathbf{L}_{02}\,L_{03}\,\mathbf{L}_{04}\cdots L_{0,2N-1}\,\mathbf{L}_{0,2N}]$ and $T_2=\mathrm{Tr}_0[\mathbf{L}_{01}\,L_{02}\,\mathbf{L}_{03}\,L_{04}\cdots \mathbf{L}_{0,2N-1}\,L_{0,2N}]$, where each subscript pairs the two-dimensional auxiliary space $0$ with one physical site. Establish algebraically, for arbitrary lattice size $N$, that these products commute pairwise — $T_1(v)T_2(v)\,T_1(v')T_2(v')=T_1(v')T_2(v')\,T_1(v)T_2(v)$ for any two weight sets $v$ and $v'$ on the same manifold — when the weights lie on the free-fermion condition $v_1 v_2+v_3 v_4-v_5 v_6-v_7 v_8=0$ combined with the Krinsky-type constraints $v_6 v_8/(v_5 v_7)=\Delta_1$, $(v_1 v_4+v_2 v_3)/(v_5 v_7)=\Delta_2$ and $(v_1^2+v_4^2-v_2^2-v_3^2)/(v_5 v_7)=\Delta_3$, with fixed parameters $\Delta_1,\Delta_2,\Delta_3$. This commutation is currently verified only numerically for $N=1,2$, while $T_1$ and $T_2$ individually do not commute for distinct weights on these manifolds. Martins (2018) proposes to confirm it by considering the respective Yang–Baxter equation for the tensor product of the two types of Lax operators; the explicit intertwiner relations are the source-suggested route rather than a separate mandatory deliverable, and any algebraic proof of the arbitrary-$N$ commutation answers the question.

Both questions serve the one objective of locating the odd eight-vertex model inside (or outside) the Yang–Baxter integrability framework: a complete answer addresses both, each by one of its admissible routes. A full classification of integrable manifolds and R-matrices for the asymmetric odd model is outside the scope of this problem.

References

1. M. J. Martins, 'Integrability of the odd eight-vertex model with symmetric weights', Journal of Physics A: Mathematical and Theoretical 51, 225002 (2018). DOI: 10.1088/1751-8121/aab9cc; arXiv:1711.03131. https://arxiv.org/abs/1711.03131
2. F. Y. Wu and H. Kunz, 'The Odd Eight-Vertex Model', Journal of Statistical Physics 116, 67-78 (2004). DOI: 10.1023/B:JOSS.0000037206.47155.58; arXiv:cond-mat/0303303. https://arxiv.org/abs/cond-mat/0303303
3. W. Galleas and M. J. Martins, 'Yang-Baxter equation for the asymmetric eight-vertex model', Physical Review E 66, 047103 (2002). DOI: 10.1103/PhysRevE.66.047103; arXiv:nlin/0303036. https://arxiv.org/abs/nlin/0303036
4. S. Krinsky, 'Equivalence of the free fermion model to the ground state of the linear XY model', Physics Letters A 39, 169-170 (1972). DOI: 10.1016/0375-9601(72)90691-3. https://doi.org/10.1016/0375-9601(72)90691-3
5. C. Fan and F. Y. Wu, 'General Lattice Model of Phase Transitions', Physical Review B 2, 723-733 (1970). DOI: 10.1103/PhysRevB.2.723. https://doi.org/10.1103/PhysRevB.2.723
6. R. J. Baxter, 'Partition function of the Eight-Vertex lattice model', Annals of Physics 70, 193-228 (1972). DOI: 10.1016/0003-4916(72)90335-1. https://doi.org/10.1016/0003-4916(72)90335-1
7. C. S. Hsue, K. Y. Lin and F. Y. Wu, 'Staggered eight-vertex model', Physical Review B 12, 429-437 (1975). DOI: 10.1103/PhysRevB.12.429. https://doi.org/10.1103/PhysRevB.12.429
8. Michael Assis, 'The 16-vertex model and its even and odd 8-vertex subcases on the square lattice', Journal of Physics A: Mathematical and Theoretical 50, 395001 (2017). DOI: 10.1088/1751-8121/aa842c; arXiv:1702.02110. https://arxiv.org/abs/1702.02110
9. Paul Melotti, 'The Free-Fermion Eight-Vertex Model: Couplings, Bipartite Dimers and Z-Invariance', Communications in Mathematical Physics 381, 33-82 (2020). DOI: 10.1007/s00220-020-03901-2; arXiv:1811.02026. https://arxiv.org/abs/1811.02026
10. R. S. Vieira, 'Solving and classifying the solutions of the Yang-Baxter equation through a differential approach. Two-state systems', Journal of High Energy Physics 2018, 110 (2018). DOI: 10.1007/JHEP10(2018)110; arXiv:1712.02341. https://arxiv.org/abs/1712.02341
11. Shi-kun Wang, 'Classification of eight-vertex solutions of the colored Yang-Baxter equation', arXiv:q-alg/9712020 (1997). https://arxiv.org/abs/q-alg/9712020
12. Pete Rigas, 'Operator formalism for discretely holomorphic parafermions of the two-color Ashkin-Teller, loop O(1), staggered eight-vertex, odd eight-vertex, and abelian sandpile models', arXiv:2310.08212 (2023). https://arxiv.org/abs/2310.08212
13. De-Zhang Li and Xin Wang, 'Free-fermion approach to the partition function zeros: Special boundary conditions and product form of solution', arXiv:2507.21943 (2025). https://arxiv.org/abs/2507.21943
14. De-Zhang Li, Xin Wang and Xiao-Bao Yang, 'Free-Fermion Models and Two-Dimensional Ising Models Under Zero Field and Imaginary Field i(pi/2)kBT', Entropy 27, 799 (2025). DOI: 10.3390/e27080799; arXiv:2406.00232. https://arxiv.org/abs/2406.00232