Generalizing the boundary-limit transfer-matrix solution of the Brascamp-Kunz Ising model: fermion-parity sector projection and partition-function factorization on other lattices and boundary conditions

The square-lattice Ising model with Brascamp-Kunz (B-K) boundary conditions - periodic around a cylinder, with the top row fixed to all $+1$ and the bottom row fixed to an alternating $+-+-$ pattern - is one of the few finite geometries whose exact partition function factorizes into a single double product, which is what makes its Fisher zeros (zeros of the partition function in the complex temperature plane) analytically determinable. Li and Wang (arXiv:2604.16992, 2026) gave the first derivation of this double product inside the Schultz-Mattis-Lieb (SML) fermionic transfer-matrix framework: the B-K cylinder is embedded in a torus whose two extra rows carry couplings $J_3$ and $-J_3$, and the limit $J_3\to+\infty$ projects out the odd fermion-parity sector of the transfer-matrix trace while equalizing the two even subsectors, collapsing Kaufman's toroidal sum of four double products into one. The present problem - posed explicitly at the close of that work and anticipated by Kastening's 2002 twisted-sector analysis - asks whether this boundary-limit technique combined with SML fermionization extends to other lattices (concretely the Kagomé lattice) and other boundary conditions (concretely open boundaries), still producing sector projection and factorized finite-lattice partition functions or a demonstrated obstruction; and what general criteria on boundary interactions determine which fermion-number sectors survive and when factorization results.

Background

The two-dimensional Ising model - spins $s_i=\pm1$ on a lattice with nearest-neighbour couplings and partition function $Z=\sum_{\{s\}}e^{-\beta H}$ - is the founding exactly solvable model of statistical mechanics. Onsager (1944) obtained its thermodynamic free energy by the transfer-matrix method, in which $Z$ is written as a trace of a power of a row-to-row operator $V$ (the transfer matrix), so that diagonalizing $V$ yields the free energy. Kaufman (1949) evaluated the finite torus (periodic boundary conditions in both directions) partition function by spinor analysis: it is a sum of four product terms. Schultz, Mattis, and Lieb (SML, 1964) rederived and physically clarified this result using the Jordan-Wigner transformation - a nonlocal change of variables that rewrites a chain of spin (Pauli) operators as creation and annihilation operators of fermions - in which the transfer matrix becomes a bilinear form in free-fermion operators. In this representation the trace over spin configurations splits into sectors labelled by fermion parity (the operator $U=e^{i\pi\sum_j c_j^\dagger c_j}$, which counts fermions modulo 2) and by periodic versus antiperiodic boundary conditions for the fermions; the four surviving combinations generate Kaufman's four products. A separate, combinatorial line of solution expresses Ising partition functions as Pfaffians (a signed square root of a determinant of an antisymmetric matrix).

Fisher zeros are the zeros of $Z$ in the complex temperature plane; their loci and their accumulation on the real axis encode the phase transition. Under toroidal boundary conditions the four-term sum obstructs an analytic determination of the zeros. Brascamp and Kunz (1974) introduced boundary conditions that avoid this obstruction: an $M\times 2N$ cylinder, periodic in the transverse direction, with the top boundary row fixed to all $+1$ and the bottom boundary row fixed to the alternating pattern $+-\cdots+-$. For these B-K boundary conditions the finite-lattice partition function takes a single double-product form - originally obtained through a duality mapping, later by Pfaffian-type and Grassmann methods - so the Fisher zeros are explicit (in the isotropic case they lie on a circle), and their density (Lu and Wu, 2001) and finite-size corrections (Janke and Kenna; Izmailian, Oganesyan, and Hu; Izmailian, Kenna, and Papoyan) have since been worked out in detail.

Li and Wang (2026) gave the first derivation of the B-K double product within the SML transfer-matrix framework. Their construction embeds the B-K cylinder in an $(M+2)\times 2N$ torus whose two extra rows carry horizontal couplings $J_3$ and $-J_3$; in the limit $J_3\to+\infty$ (with normalization $e^{-4N\beta J_3}$) only four symmetry-equivalent fixed boundary patterns survive, so $Z_{\text{B-K}}=\frac14\lim_{J_3\to+\infty}e^{-4N\beta J_3}Z(J_3)$, where $Z(J_3)$ is the partition function of the embedded toroidal system. In the fermionic representation the transfer-matrix product decomposes into even and odd fermion-parity parts; the limit annihilates the odd part and makes the two even subparts contribute equally, collapsing the four-product toroidal structure into the single double product $Z_{\text{B-K}}=2^{2MN}\prod_{l=1}^{N}\prod_{j=1}^{M}(\cosh 2K_1\cosh 2K_2-\sinh 2K_1\cos\frac{(2l-1)\pi}{2N}-\sinh 2K_2\cos\frac{j\pi}{M+1})$ with $K_i=\beta J_i$. Relatedly, Kastening (2002) - generalizing a simplified transfer-matrix treatment to the four periodic/antiperiodic sector combinations on the torus - had earlier found an exact relation between a twisted-sector combination and the square of the B-K partition function, and explicitly asked for an understanding of that relation without resorting to the explicit solutions. In parallel, Li and Wang (2025) showed, via mappings to a free-fermion eight-vertex model (a vertex model whose weights satisfy a condition making it equivalent to free fermions), that B-K-type boundary conditions with product-form partition functions exist for the triangular and honeycomb lattices as well - though not by the transfer-matrix route, and not for the Kagomé lattice (the corner-sharing-triangles lattice exactly solved in the 1950s through star-triangle decorated-lattice duality mappings by Kanô and Naya).

What remains unresolved is precisely the generality of this mechanism: whether the boundary-limit-plus-SML technique extends to other lattices - concretely the Kagomé lattice - and to other boundary conditions - concretely open (free) boundaries, whose thermodynamic free energies are known from Kaufman-type spinor calculations (Baxter, 2017) - and, more fundamentally, what general criteria on boundary interactions determine which fermion-number sectors survive the transfer-matrix trace and when factorization of the partition function results.

Problem Statement

The problem is to extend the boundary-limit transfer-matrix solution of the Brascamp-Kunz (B-K) Ising model beyond the square-lattice case and to characterize its mechanism, in two connected parts.

1. Generalization of the technique. Apply the construction that realizes the B-K boundary conditions - in which the $M\times 2N$ cylinder (periodic around its transverse direction of circumference $2N$, with one boundary row fixed to all $+1$ spins and the other fixed to the alternating $+-\cdots+$- pattern) is embedded in an $(M+2)\times 2N$ torus whose two added rows carry the internal couplings $J_3$ (ferromagnetic) and $-J_3$ (antiferromagnetic), and the limit $J_3\to+\infty$ is taken at inverse temperature $\beta$, so that only four symmetry-equivalent fixed boundary patterns survive and $Z_{\text{B-K}}=\frac14\lim_{J_3\to+\infty}e^{-4N\beta J_3}Z(J_3)$, where $Z(J_3)$ is the partition function of the embedded toroidal system - together with the Schultz-Mattis-Lieb (SML) fermionic transfer-matrix method, to other lattices and other boundary conditions, in particular the Kagomé-lattice Ising model and the square-lattice Ising model with open (free) boundary conditions; for part 1 it suffices to treat these two named cases. For each of them, determine whether the limit again projects out particular fermion-number sectors of the transfer-matrix trace and whether the finite-lattice partition function again factorizes into product form (analogous to the B-K double product), or establish why the projection or the factorization fails.

2. Mechanism and criteria. Characterize the general mechanism by which boundary interactions in the transfer-matrix product suppress particular fermion-number sectors in the trace, and formulate general criteria - in terms of the placement, signs, and limiting strengths of the boundary perturbations relative to the fermionization - that determine which sectors survive and when double-product factorization of the partition function results.

The two parts share one objective: a general transfer-matrix (SML) account of boundary-driven fermion-sector selection and partition-function factorization. The known anchor cases - the toroidal sum of four double products under uniform interactions (Kaufman), the B-K single double product with the odd sector projected out (Li and Wang), and the exact relation between twisted-sector combinations and the B-K partition function (Kastening) - are the reference points against which any proposed criterion is to be checked. For each case named in part 1, either an exact factorized partition function (where the theory permits one) or a demonstrated obstruction is an admissible outcome.

References

1. De-Zhang Li and Xin Wang, "Solution of the Ising model with Brascamp-Kunz boundary conditions by the transfer matrix method", arXiv:2604.16992 (2026); published version: APS Open Science, DOI 10.1103/1dby-z8np (2026). https://arxiv.org/abs/2604.16992
2. H. J. Brascamp and H. Kunz, "Zeroes of the partition function for the Ising model in the complex temperature plane", Journal of Mathematical Physics 15, 65 (1974). DOI 10.1063/1.1666505. https://doi.org/10.1063/1.1666505
3. L. Onsager, "Crystal Statistics. I. A Two-Dimensional Model with an Order-Disorder Transition", Physical Review 65, 117 (1944). DOI 10.1103/PhysRev.65.117. https://doi.org/10.1103/PhysRev.65.117
4. B. Kaufman, "Crystal Statistics. II. Partition Function Evaluated by Spinor Analysis", Physical Review 76, 1232 (1949). DOI 10.1103/PhysRev.76.1232. https://doi.org/10.1103/PhysRev.76.1232
5. T. D. Schultz, D. C. Mattis, and E. H. Lieb, "Two-Dimensional Ising Model as a Soluble Problem of Many Fermions", Reviews of Modern Physics 36, 856 (1964). DOI 10.1103/RevModPhys.36.856. https://doi.org/10.1103/RevModPhys.36.856
6. K. Kanô and S. Naya, "Antiferromagnetism. The Kagomé Ising Net", Progress of Theoretical Physics 10, 158 (1953). DOI 10.1143/ptp/10.2.158. https://doi.org/10.1143/ptp/10.2.158
7. B. Kastening, "Simplified transfer matrix approach in the two-dimensional Ising model with various boundary conditions", Physical Review E 66, 057103 (2002). arXiv:cond-mat/0209544; DOI 10.1103/PhysRevE.66.057103. https://arxiv.org/abs/cond-mat/0209544
8. M.-C. Wu and C.-K. Hu, "Exact partition functions of the Ising model on M×N planar lattices with periodic-aperiodic boundary conditions", Journal of Physics A: Mathematical and General 35, 5189 (2002). arXiv:cond-mat/0204217; DOI 10.1088/0305-4470/35/25/304. https://arxiv.org/abs/cond-mat/0204217
9. W. T. Lu and F. Y. Wu, "Density of the Fisher Zeroes for the Ising Model", Journal of Statistical Physics 102, 953 (2001). DOI 10.1023/A:1004863322373. https://doi.org/10.1023/A:1004863322373
10. W. Janke and R. Kenna, "Finite-size scaling and corrections in the Ising model with Brascamp-Kunz boundary conditions", Physical Review B 65, 064110 (2002). DOI 10.1103/PhysRevB.65.064110. https://doi.org/10.1103/PhysRevB.65.064110
11. N. Sh. Izmailian, K. B. Oganesyan, and C.-K. Hu, "Exact finite-size corrections for the square-lattice Ising model with Brascamp-Kunz boundary conditions", Physical Review E 65, 056132 (2002). DOI 10.1103/PhysRevE.65.056132. https://doi.org/10.1103/PhysRevE.65.056132
12. N. Izmailian, R. Kenna, and Vl. V. Papoyan, "Exact coefficients of finite-size corrections in the Ising model with Brascamp-Kunz boundary conditions and their relationships for strip and cylindrical geometries", Journal of Physics A: Mathematical and Theoretical 56, 435007 (2023). arXiv:2303.03484; DOI 10.1088/1751-8121/acf96b. https://arxiv.org/abs/2303.03484
13. I. Lyberg, "The Ising lattice with Brascamp-Kunz boundary conditions and an external magnetic field", arXiv:0805.2497 (2008). https://arxiv.org/abs/0805.2497
14. I. Lyberg, "Free energy of the anisotropic Ising lattice with Brascamp-Kunz boundary conditions", Physical Review E 87, 062141 (2013). arXiv:0810.4456 (preprint version, titled "The free energy of the non-isotropic Ising lattice with Brascamp-Kunz boundary conditions"); DOI 10.1103/PhysRevE.87.062141. https://arxiv.org/abs/0810.4456
15. R. J. Baxter, "The bulk, surface and corner free energies of the square lattice Ising model", Journal of Physics A: Mathematical and Theoretical 50, 014001 (2017). arXiv:1606.02029; DOI 10.1088/1751-8113/50/1/014001. https://arxiv.org/abs/1606.02029
16. De-Zhang Li, Xin Wang, and Xiao-Bao Yang, "Free-Fermion Models and Two-Dimensional Ising Models Under Zero Field and Imaginary Field $i(\pi/2)k_BT$", Entropy 27, 799 (2025). arXiv:2406.00232; DOI 10.3390/e27080799. https://arxiv.org/abs/2406.00232
17. De-Zhang Li and Xin Wang, "Free-fermion approach to the partition function zeros: Special boundary conditions and product form of solution", Physical Review Research 7, 043258 (2025). arXiv:2507.21943; DOI 10.1103/b6d1-6sk5. https://arxiv.org/abs/2507.21943