Rigorous status of the claimed exact solution of the two-dimensional Ising model with next-nearest-neighbor interactions at zero field

The zero-field square-lattice Ising model with two axial nearest-neighbor couplings and two diagonal next-nearest-neighbor couplings is a long-standing open problem of exactly solvable statistical mechanics: with both diagonal bonds present the interaction graph is non-planar, so the transfer-matrix spinor and free-fermion methods behind all known two-dimensional solutions do not apply, and the critical line is known only from series, transfer-matrix scaling, and Monte Carlo data. In 2026, Zhidong Zhang published a claimed exact solution (Phys. Rev. E 113, 064111; arXiv:2601.10749): a closed-form partition function and spontaneous magnetization derived by adapting a Clifford-algebraic/topological approach originally proposed for the three-dimensional Ising model, with a topological Lorentz (gauge) rotation whose parameter $K_5=K_2K_4/K_1$ is fixed by a star-triangle relation and topological phases $\phi=2\pi$, $\phi'=\pi/2$. The paper omits the decisive derivations, and the parent framework was publicly disputed (Wu, McCoy, Fisher and Chayes; Perk) and is quantitatively contradicted for the simple-cubic lattice by high-precision Monte Carlo; the newly claimed critical points also exceed validated approximate values by 6-7%. The open problem is to rigorously settle the status of the published claim by completing the missing mathematics (construction and uniqueness of the trivializing rotation, derivation of the phase assignments, principled branch selection, fully reproducible diagonalization) or by rigorously refuting the claimed results.

Background

The Ising model places spins $s_{i,j}=\pm1$ on a lattice with pairwise exchange couplings $J$ and is the canonical testing ground for exact methods in statistical mechanics. Onsager (1944) computed the partition function of the zero-field square-lattice model; Kaufman (1949) recast the solution using spinor (Clifford-algebra) representations, writing the transfer matrix (the matrix whose powers propagate the system row by row, so that the partition function is its trace) as a product of exponentials of quadratic forms in Clifford generators (matrices obeying the same anticommutation relations as the Dirac gamma matrices); Yang (1952) derived the spontaneous magnetization. Houtappel (1950) and Newell (1950) extended exact solutions to the triangular and honeycomb lattices using the star-triangle relation, a local transformation that replaces the three spins of a triangle by a single spin at the dual-lattice site and relates the partition functions of dual lattices (the same circle of ideas as the Kramers-Wannier duality, which maps the high- and low-temperature expansions of a planar Ising model onto each other and, for the square lattice, defines a dual coupling $K^*=\frac{1}{2}\ln(\coth K)$).

The model at issue here is the square (rectangular) lattice with two axial nearest-neighbor couplings $J_1,J_2$ and both diagonal next-nearest-neighbor couplings $J_3,J_4$, so each spin interacts with its four axial and four diagonal neighbors. Its solvable limits are classical: with both diagonals removed it is Onsager's model; with one diagonal family removed it is the triangular lattice of Houtappel and Newell; with the axial couplings removed it decouples into two independent square lattices. With all four couplings present, both diagonals of every plaquette carry a bond, the interaction graph becomes non-planar (bond edges cross), and the combinatorial determinant/free-fermion techniques (which reduce the partition function to a determinant by mapping the spins to noninteracting fermions) underlying every known two-dimensional solution no longer apply. The 2026 paper cites the 1953 review of Newell and Montroll for the assessment that this model, the three-dimensional Ising model, and the two-dimensional Ising model in a magnetic field are three unsolved problems whose common difficulty is topological in origin. Fan and Wu (1969) obtained exact results only for special coupling cases (ground-state classification, cases with $T_c=0$) and an approximation for the general model. Accordingly, the ferromagnetic critical line is known only numerically: from high-temperature series (Oitmaa, 1981), Monte Carlo simulation of the competing antiferromagnetic case (Yin and Landau, 2009), and the accurate closed-form approximate condition of Zandvliet (2006), refined in Zandvliet (2023),

$e^{-2J_x/k_BT_c}+e^{-2J_y/k_BT_c}+e^{-2(J_x+J_y)/k_BT_c}\left(2-e^{-4J_d/k_BT_c}\right)=e^{4J_d/k_BT_c},$

where $J_x,J_y$ are the axial and $J_d$ the diagonal couplings. This condition reduces to Onsager's exact result at $J_d=0$ and was found to agree excellently with series expansion, transfer-matrix finite-size scaling, and Monte Carlo data, with best accuracy when the axial couplings dominate.

The claim was published by Zhidong Zhang as "Two-dimensional Ising model with the next nearest interactions" (Phys. Rev. E 113, 064111 (2026); arXiv:2601.10749, v1 January 2026 through v6 May 2026). The derivation writes the transfer matrix as $V=V_4V_3V_2V_1$, a product of exponentials of Clifford generators in the Onsager-Kaufman manner, with the term $V_4$ induced by $J_4$ being nonlinear and nonlocal. The author reinterprets the model as a triangular-lattice Ising model plus an interaction along a fictitious third direction and imports machinery he had earlier proposed for the three-dimensional Ising model: a topological Lorentz transformation, regarded as a local gauge transformation that rotates the transfer operator from a nontrivial to a trivial topological basis, together with topological phase factors attached to eigenvectors and eigenvalues. The rotation parameter is fixed as $K_5=K_2K_4/K_1$ (with $K_l=J_l/k_BT$) by the star-triangle relation $K_1K_1^*=K_1K_2+K_1K_4+K_2K_4$, and the phases are set to $\phi=2\pi$ and $\phi'=\pi/2$. From this the paper states, without exhibiting the intermediate operator algebra, a triple-integral free energy, a Yang-type spontaneous magnetization with critical exponent $1/8$, and critical points such as $1/K_c=4.069257$ for $J_3=J_4=\frac{1}{2}J_1$ and $1/K_c=5.77078017$ (that is, $e^{-2K_c}=2^{-1/2}$) for all four couplings equal.

The claim's provenance is essential for evaluating it. The machinery is adapted from the same author's 2007 conjectures for the three-dimensional Ising model, which he and collaborators later claimed to prove by a Clifford-algebra approach (Zhang, Suzuki and March, 2019) and by Riemann-Hilbert methods (Suzuki and Zhang, 2021). That program was publicly disputed: Wu, McCoy, Fisher and Chayes (2008) state that they "disprove the conjecture and point out the flaws in the arguments leading to the conjectured expressions", and Perk (2009) argues the work "is in error and violates well-known theorems". Independently, the 2007 conjectures place the simple-cubic Ising critical point at $1/K_c=4.15617384$ (the golden-ratio value $e^{-2K_c}=(\sqrt{5}-1)/2$), which is excluded by high-precision Monte Carlo: $K_c=0.2216595\pm0.0000026$ (Ferrenberg and Landau, 1991), i.e. $1/K_c=4.5115$, an 8% discrepancy far beyond the uncertainty. The 2026 paper reasserts the golden-ratio cubic value in its comparison table. For the new two-dimensional claim, the asserted critical points exceed the values implied by Zandvliet's validated approximate condition by about 6-7% ($5.7708$ versus approximately $5.38$ at equal couplings; $4.0693$ versus approximately $3.84$ at $J_3=J_4=\frac{1}{2}J_1$), the same order as the known 3D error. No independent verification or refutation of the claim has been published.

Problem Statement

Determine, with mathematical rigor and complete reproducibility, the status of the claimed exact solution of the zero-field square-lattice Ising model with axial nearest-neighbor couplings $J_1,J_2$ and diagonal next-nearest-neighbor couplings $J_3,J_4$ (ferromagnetic, with the paper's conventions $J_1\geq J_2$, $J_3\geq J_4$ up to permutations), published by Zhang in Physical Review E 113, 064111 (2026). The claim asserts a closed-form partition function and spontaneous magnetization obtained through a topological Lorentz (gauge) rotation with parameter $K_5=K_2K_4/K_1$ fixed by the star-triangle relation $K_1K_1^*=K_1K_2+K_1K_4+K_2K_4$ (where $K_1^*$ is the Kramers-Wannier dual coupling), and topological phases $\phi=2\pi$ and $\phi'=\pi/2$ attached to the eigenvectors and eigenvalues of the rotated transfer operator. An acceptable answer does at least one of the following, with every step explicit enough for independent verification.

(1) Verification or correction: supply the missing mathematics of the claim, namely (a) a complete construction of the trivializing rotation together with a proof of its existence and uniqueness (or a precise characterization of the admissible rotation set) over the stated parameter regime, replacing the schematic outline; (b) a rigorous derivation of the phase assignments $\phi=2\pi$, $\phi'=\pi/2$, or corrected values derived from the submitted construction; (c) a principled classification of the branches of the star-triangle relation with a proof of which root is physical, replacing the heuristic rule of choosing the lowest critical point; (d) a step-by-step algebraic diagonalization with all intermediate operator identities, including the treatment of the nonlinear, nonlocal factors induced by $J_4$. From these, derive the partition function, the spontaneous magnetization, and the critical points (for instance $1/K_c=4.069257$ at $J_3=J_4=\frac{1}{2}J_1$ and $1/K_c=5.77078017$ at equal couplings, or their corrected values), demonstrate the correct solvable limits ($J_3=J_4=0$ reducing to Onsager's square-lattice solution; $J_4=0$ reducing to Houtappel's triangular-lattice solution), and establish consistency with convergent numerical data for the four-coupling model.

(2) Refutation: demonstrate rigorously that the claimed solution is incorrect, either by proving that a load-bearing step of the derivation fails (exhibiting the specific operator identity or transformation that does not hold, for instance within the finite-dimensional Clifford-generator algebra, and showing that the claimed final formulas cannot be sustained without it), or by a rigorous argument, such as certified bounds obtained from exact finite-lattice computations with controlled finite-size estimates, that the claimed free energy, spontaneous magnetization, or critical points are incompatible with the true thermodynamics of the model.

Both branches settle the same scientific objective: the mathematically established status of the published claim. A refutation does not by itself solve the pre-existing open problem of finding the model's exact solution, which then remains open, and extending the claim beyond ferromagnetic couplings or to nonzero magnetic field is not required; those are stated by the source itself as further generalizations. The two branches are not mutually exclusive: an answer that both exhibits a specific failure of the published derivation and supplies a corrected construction is acceptable, as is an answer that rigorously establishes that the published construction is too incomplete or ill-defined to be verified or refuted as stated.

References

1. Zhidong Zhang, "Two-dimensional Ising model with the next nearest interactions", Physical Review E 113, 064111 (2026). DOI: 10.1103/c3qf-wvjr. https://doi.org/10.1103/c3qf-wvjr Preprint: Zhidong Zhang, "Exact solution of a two-dimensional (2D) Ising model with the next nearest interactions", arXiv:2601.10749 (v1 2026-01-13; v6 2026-05-28). https://arxiv.org/abs/2601.10749
2. Lars Onsager, "Crystal Statistics. I. A Two-Dimensional Model with an Order-Disorder Transition", Physical Review 65, 117-149 (1944). DOI: 10.1103/PhysRev.65.117. https://doi.org/10.1103/PhysRev.65.117
3. Bruria Kaufman, "Crystal Statistics. II. Partition Function Evaluated by Spinor Analysis", Physical Review 76, 1232-1243 (1949). DOI: 10.1103/PhysRev.76.1232. https://doi.org/10.1103/PhysRev.76.1232
4. C. N. Yang, "The Spontaneous Magnetization of a Two-Dimensional Ising Model", Physical Review 85, 808-816 (1952). DOI: 10.1103/PhysRev.85.808. https://doi.org/10.1103/PhysRev.85.808
5. R. M. F. Houtappel, "Order-disorder in hexagonal lattices", Physica 16, 425-455 (1950). DOI: 10.1016/0031-8914(50)90130-3. https://doi.org/10.1016/0031-8914(50)90130-3
6. G. F. Newell, "Crystal Statistics of a Two-Dimensional Triangular Ising Lattice", Physical Review 79, 876-882 (1950). DOI: 10.1103/PhysRev.79.876. https://doi.org/10.1103/PhysRev.79.876
7. G. F. Newell and E. W. Montroll, "On the Theory of the Ising Model of Ferromagnetism", Reviews of Modern Physics 25, 353-389 (1953). DOI: 10.1103/RevModPhys.25.353. https://doi.org/10.1103/RevModPhys.25.353
8. H. A. Kramers and G. H. Wannier, "Statistics of the Two-Dimensional Ferromagnet. Part I", Physical Review 60, 252-262 (1941). DOI: 10.1103/PhysRev.60.252. https://doi.org/10.1103/PhysRev.60.252
9. Chungpeng Fan and F. Y. Wu, "Ising Model with Second-Neighbor Interaction. I. Some Exact Results and an Approximate Solution", Physical Review 179, 560-569 (1969). DOI: 10.1103/PhysRev.179.560. https://doi.org/10.1103/PhysRev.179.560
10. J. Oitmaa, "The square-lattice Ising model with first and second neighbour interactions", Journal of Physics A: Mathematical and General 14, 1159-1168 (1981). DOI: 10.1088/0305-4470/14/5/035. https://doi.org/10.1088/0305-4470/14/5/035
11. Harold J. W. Zandvliet, "The 2D Ising square lattice with nearest- and next-nearest-neighbor interactions", Europhysics Letters 73, 747 (2006). DOI: 10.1209/epl/i2005-10451-1. https://doi.org/10.1209/epl/i2005-10451-1
12. Junqi Yin and D. P. Landau, "Phase diagram and critical behavior of the square-lattice Ising model with competing nearest-neighbor and next-nearest-neighbor interactions", Physical Review E 80, 051117 (2009). DOI: 10.1103/PhysRevE.80.051117. https://doi.org/10.1103/PhysRevE.80.051117
13. Harold J. W. Zandvliet, "Phase diagram of the square 2D Ising lattice with nearest neighbor and next-nearest neighbor interactions", Phase Transitions 96, 187-195 (2023). DOI: 10.1080/01411594.2022.2162897. https://doi.org/10.1080/01411594.2022.2162897
14. Z.-D. Zhang, "Conjectures on the exact solution of three-dimensional (3D) simple orthorhombic Ising lattices", Philosophical Magazine 87, 5309-5419 (2007). DOI: 10.1080/14786430701646325. https://doi.org/10.1080/14786430701646325
15. F. Y. Wu, B. M. McCoy, M. E. Fisher and L. Chayes, "Comment on a recent conjectured solution of the three-dimensional Ising model", Philosophical Magazine 88, 3093-3095 (2008). DOI: 10.1080/14786430802537738. https://doi.org/10.1080/14786430802537738
16. Jacques H. H. Perk, "Comment on 'Conjectures on exact solution of three-dimensional (3D) simple orthorhombic Ising lattices'", Philosophical Magazine 89, 761-764 (2009). DOI: 10.1080/14786430902776970. https://doi.org/10.1080/14786430902776970
17. Z. D. Zhang, O. Suzuki and N. H. March, "Clifford algebra approach of 3D Ising model", Advances in Applied Clifford Algebras 29, 12 (2019). DOI: 10.1007/s00006-018-0923-2. https://doi.org/10.1007/s00006-018-0923-2
18. O. Suzuki and Z. D. Zhang, "A Method of Riemann-Hilbert Problem for Zhang's Conjecture 1 in a Ferromagnetic 3D Ising Model: Trivialization of Topological Structure", Mathematics 9(7), 776 (2021). DOI: 10.3390/math9070776. https://doi.org/10.3390/math9070776
19. Z. D. Zhang and O. Suzuki, "A Method of the Riemann-Hilbert Problem for Zhang's Conjecture 2 in a Ferromagnetic 3D Ising Model: Topological Phases", Mathematics 9(22), 2936 (2021). DOI: 10.3390/math9222936. https://doi.org/10.3390/math9222936
20. Alan M. Ferrenberg and D. P. Landau, "Critical behavior of the three-dimensional Ising model: A high-resolution Monte Carlo study", Physical Review B 44, 5081-5091 (1991). DOI: 10.1103/PhysRevB.44.5081. https://doi.org/10.1103/PhysRevB.44.5081