Emergent planarity at criticality: continuity, current mixing, crossing estimates, and Pfaffian order–disorder structure for two-dimensional finite-range Ising models

The Inventiones mathematicae paper of Aizenman, Duminil-Copin, Tassion and Warzel (216 (2019) 661–743) proved that at criticality the boundary spin correlations of two-dimensional Ising models with finite-range ferromagnetic interactions are asymptotically Pfaffian ('emergent planarity'), and posed four linked open problems completing the picture for this non-solvable, non-planar regime: (i) prove that the spontaneous magnetization vanishes at the critical point, i.e. that the phase transition of every finite-range ferromagnetic translation-invariant Ising model on Z^2 is continuous; (ii) prove polynomial ratio-weak mixing for the critical sourceless random current; (iii) prove uniform Russo–Seymour–Welsh-type crossing bounds for the current percolation at every scale; (iv) strengthen the asymptotic Pfaffian relation for bulk order–disorder correlations so that the error is measured relative to the order–disorder correlations themselves. The four questions (Section 7, Questions 7.1–7.4 of the source paper) remain open as of August 2026: continuity is known only for planar interactions (Yang's exact solution for the square lattice; Duminil-Copin–Lis for planar periodic graphs; Duminil-Copin–Sidoravicius–Tassion for planar random-cluster/Potts models) and for nearest-neighbour models in d≥3; current mixing is proven only in the supercritical regime (Hansen–Klausen); the nearest-neighbour current geometry is understood through conformally invariant scaling limits (Chen–Xia; Duminil-Copin–Lis–Qian) without extension to finite range; and Pfaffian universality beyond boundary correlations is established only for weak perturbations of the nearest-neighbour model (Cava–Giuliani–Greenblatt).

Background

The Ising model on the lattice $\mathbb{Z}^2$ assigns spins $\sigma_x\in\{\pm1\}$ to lattice sites with Hamiltonian $H(\sigma)=-\sum_{x,y}J_{x,y}\sigma_x\sigma_y$. When the couplings are ferromagnetic ($J_{x,y}\ge0$), translation-invariant ($J_{x,y}=J(x-y)$), finite-range ($J_{x,y}=0$ for $\|x-y\|>R$), and connect the lattice, the model undergoes a phase transition at a critical inverse temperature $\beta_c$. Its order parameter is the spontaneous magnetization $m^*(\beta)=\langle\sigma_0\rangle_\beta^+$, the magnetization in the infinite-volume Gibbs state obtained with all-plus boundary conditions; the transition is continuous if $m^*(\beta_c)=0$ and first-order otherwise. Crucially, for generic finite-range couplings the interaction graph on $\mathbb{Z}^2$ is not planar (planarity is a special case: for instance, couplings along a single diagonal direction still give the planar triangular lattice, and such exactly planar cases are covered by the planar results cited below), so the exact machinery that solves planar models (Onsager's free energy, Yang's magnetization formula, and the Kac–Ward, Grassmann and Pfaffian determinant methods, which express planar Ising partition functions and correlations through free-fermion determinants) no longer applies to the general finite-range model, while the model is still genuinely two-dimensional.

For the nearest-neighbour model the critical behaviour is understood in depth. Yang (1952) proved $m^*(\beta_c)=0$ on the square lattice from the exact solution. Duminil-Copin–Sidoravicius–Tassion proved continuity of the phase transition for the planar random-cluster and Potts models with cluster weight $1\le q\le4$ on $\mathbb{Z}^2$ (the case $q=2$ is the Ising model; the random-cluster model is the Fortuin–Kasteleyn percolation representation of Potts/Ising systems). Duminil-Copin–Lis proved that the spontaneous magnetization of the Ising model on any planar periodic graph vanishes at criticality, via a correspondence between double random currents and dimer models. Aizenman–Duminil-Copin–Sidoravicius proved continuity for the nearest-neighbour model on $\mathbb{Z}^3$ using the random current representation (introduced below), the case $d\ge4$ having been settled earlier through reflection positivity (a positivity property of reflection-symmetric couplings) and the infrared bound it yields on spin correlations. The two-dimensional finite-range case — planar geometry without planarity of the interaction graph — is covered by none of these methods.

The random current representation expresses Ising correlations in terms of integer-valued currents $\mathbf{n}\in\mathbb{Z}_{\ge0}^{E}$ on the edge set $E$, constrained by parity conditions at vertices (the vertices of odd total incident current are the sources of $\mathbf{n}$). The symbol $\mathbf{P}^{\emptyset}_{\mathbb{Z}^2,\beta_c}$ denotes the (single) sourceless random current measure in the full plane at criticality, and the current percolation $\widehat{\mathbf{n}}=\{e:\mathbf{n}_e>0\}$ is the set of edges carrying positive current. The switching lemma for currents relates intersection properties of current paths to correlation identities; in particular the sum of two independent sourceless currents has the law of the $q=2$ random-cluster model, which links the current geometry to percolation. In Bernoulli percolation, the Russo–Seymour–Welsh (RSW) theory provides uniformly positive lower bounds, at criticality, on the probability that a long thin rectangle is crossed in the hard direction (from one short side to the other, spanning the longer dimension); no analogue of this theory is known for current percolation of finite-range models.

In this landscape, Aizenman–Duminil-Copin–Tassion–Warzel proved 'emergent planarity': for finite-range ferromagnetic, translation-invariant, reflection-invariant Ising models on the half-plane $\mathbb{H}=\mathbb{Z}\times\mathbb{Z}_+$, the $2n$-point correlation of boundary spins at $\beta_c$ equals the Pfaffian of the matrix of two-point functions up to a factor $1+o(1)$ as the spin separations tend to infinity. (The Pfaffian $\mathrm{Pf}^n(A)=\sum_{\pi}\mathrm{sgn}(\pi)\prod_{j}A_{\pi(2j-1),\pi(2j)}$, summed over pairings $\pi$ of $\{1,\dots,2n\}$, is the fermionic counterpart of the Wick rule familiar from Gaussian variables; for planar nearest-neighbour models, boundary spin correlations are exactly Pfaffian — a classical theorem of Groeneveld–Boel–Kasteleyn — reflecting the free-fermion structure underlying Onsager's solution.) The proof introduces new stochastic-geometric controls of the current, among them crossing estimates for $\widehat{\mathbf{n}}$ valid only along an infinite sequence of scales. The paper's discussion section then lists four central problems left open — continuity of the order parameter, polynomial mixing of the critical current, crossing estimates at all scales, and the full Pfaffian relation for bulk order–disorder correlations — which together would complete the emergent-planarity picture of the critical finite-range model. The order–disorder variables involved in the fourth question are the Kadanoff–Ceva observables: a disorder operator $\mu_\ell$ flips the sign of every coupling crossed by a line $\ell$, and the paired order–disorder variables $\tau_i$ combine a spin at a site $x_i$ with a disorder line running from a neighbour of $x_i$ to a common 'grand central' region; for planar nearest-neighbour models their correlations are exactly Pfaffian (the source paper's Theorem 6.1).

Problem Statement

For the class of Ising models on $\mathbb{Z}^2$ whose couplings are ferromagnetic ($J_{x,y}\ge0$), translation-invariant ($J_{x,y}=J(x-y)$), finite-range ($J_{x,y}=0$ for $\|x-y\|>R$ for some finite $R$) with connected interaction graph, settle the four linked open questions posed in Section 7 of Aizenman–Duminil-Copin–Tassion–Warzel (Inventiones mathematicae 216 (2019), 661–743; arXiv:1801.04960). Question (i) concerns this full class, as in the source's Question 7.1, which carries no symmetry hypothesis beyond it; Questions (ii)–(iv) concern models that additionally satisfy the reflection invariance $J_{0,(a,b)}=J_{0,(-a,b)}=J_{0,(a,-b)}=J_{0,(b,a)}$ for all $a,b\in\mathbb{Z}$, assumed in the source for the half-plane current-geometry and order–disorder statements on which those questions build:

(i) Continuity of the order parameter: prove that the spontaneous magnetization — the magnetization $m^*(\beta)=\langle\sigma_0\rangle_\beta^+$ in the infinite-volume Gibbs state obtained with all-plus boundary conditions — vanishes at the critical inverse temperature $\beta_c$, i.e. that the phase transition of the two-dimensional finite-range ferromagnetic Ising model is continuous.

(ii) Polynomial mixing of random currents: prove that there exist $c,C>0$ such that for any two events $A$ (depending on edges inside a box of size $n$) and $B$ (depending on edges outside a concentric box of size $N\ge n$),

$\left|\mathbf{P}^{\emptyset}_{\mathbb{Z}^2,\beta_c}[A\cap B]-\mathbf{P}^{\emptyset}_{\mathbb{Z}^2,\beta_c}[A]\,\mathbf{P}^{\emptyset}_{\mathbb{Z}^2,\beta_c}[B]\right|\le C\left(\frac{n}{N}\right)^{c}\mathbf{P}^{\emptyset}_{\mathbb{Z}^2,\beta_c}[A]\mathbf{P}^{\emptyset}_{\mathbb{Z}^2,\beta_c}[B],$

where $\mathbf{P}^{\emptyset}_{\mathbb{Z}^2,\beta_c}$ is the law of the sourceless critical random current in the full plane (as defined in the background above).

(iii) Crossing estimates for current percolation: prove that there exists $c>0$ such that for every $n$,

$\mathbf{P}^{\emptyset}_{\mathbb{Z}^2,\beta_c}[\text{the box of size }n\times2n\text{ is crossed in the hard direction by a path of positive current}]\ge c,$

that is, by a path of edges each carrying strictly positive current ($\mathbf{n}_e>0$), connecting the two short sides of the rectangle and thus spanning its longer dimension.

(iv) Full asymptotic Pfaffian relation for order–disorder correlations: strengthen the source paper's Corollary 6.5 (equation (6.12)) so that the error term $o(1)\langle\sigma_{x_1}\cdots\sigma_{x_{2n}}\rangle_{\mathbb{Z}^2,\beta_c}$ is replaced by $o(1)\left|\langle\tau_1\cdots\tau_{2n}\rangle_{\mathbb{Z}^2,\beta_c}\right|$, where $\tau_1,\dots,\tau_{2n}$ are the paired order–disorder (Kadanoff–Ceva) variables defined in the background above, thereby establishing the fermionic Pfaffian structure for bulk order–disorder fields in finite-range models.

These four statements constitute one program with a common objective, completing the critical theory of the two-dimensional finite-range Ising model: (ii) and (iii) are the stochastic-geometric controls on the critical sourceless current — the source paper obtains (iii) only along an infinite sequence of scales — which are expected to be the mechanism yielding the physical conclusions (i) and (iv); the source authors state that mixing and RSW-type statements for currents 'would have far-reaching consequences for the study of universality'. A complete answer proves each of (i)–(iv) in the stated generality, with no restriction on the magnitude of the interaction range $R$ beyond finiteness, or on the relative strength of the couplings.

References

1. Michael Aizenman, Hugo Duminil-Copin, Vincent Tassion, Simone Warzel, Emergent planarity in two-dimensional Ising models with finite-range interactions, Inventiones mathematicae 216 (2019), no. 3, 661–743, DOI: 10.1007/s00222-018-00851-4, arXiv:1801.04960, https://arxiv.org/abs/1801.04960
2. C. N. Yang, The Spontaneous Magnetization of a Two-Dimensional Ising Model, Physical Review 85 (1952), 808–816, DOI: 10.1103/PhysRev.85.808, https://doi.org/10.1103/PhysRev.85.808
3. Hugo Duminil-Copin, Marcin Lis, On the double random current nesting field, Probability Theory and Related Fields 175 (2019), no. 3–4, 937–955, DOI: 10.1007/s00440-019-00899-0, arXiv:1712.02305, https://arxiv.org/abs/1712.02305
4. Hugo Duminil-Copin, Vladas Sidoravicius, Vincent Tassion, Continuity of the phase transition for planar random-cluster and Potts models with $1\le q\le4$, Communications in Mathematical Physics 349 (2017), 47–107, DOI: 10.1007/s00220-016-2759-8, arXiv:1505.04159, https://arxiv.org/abs/1505.04159
5. Michael Aizenman, Hugo Duminil-Copin, Vladas Sidoravicius, Random currents and continuity of Ising model's spontaneous magnetization, Communications in Mathematical Physics 334 (2015), 719–742, DOI: 10.1007/s00220-014-2093-y, arXiv:1311.1937, https://arxiv.org/abs/1311.1937
6. Aran Raoufi, Translation-Invariant Gibbs States of Ising model: General Setting, The Annals of Probability 48 (2020), no. 2, DOI: 10.1214/19-aop1374, arXiv:1710.07608, https://arxiv.org/abs/1710.07608
7. Hong-Bin Chen, Jiaming Xia, Conformal invariance of random currents: a stability result, arXiv:2306.10625 (2023), https://arxiv.org/abs/2306.10625
8. Hugo Duminil-Copin, Marcin Lis, Wei Qian, Conformal invariance of double random currents I: identification of the limit, Proceedings of the London Mathematical Society 130 (2025), no. 1, e70022, DOI: 10.1112/plms.70022, arXiv:2107.12985, https://arxiv.org/abs/2107.12985
9. Frederik Ravn Klausen, Aran Raoufi, Mass Scaling of the Near-Critical 2D Ising Model Using Random Currents, Journal of Statistical Physics 188 (2022), no. 3, article 21, DOI: 10.1007/s10955-022-02939-x, arXiv:2105.13673, https://arxiv.org/abs/2105.13673
10. Giulia Cava, Alessandro Giuliani, Rafael Leon Greenblatt, The scaling limit of boundary spin correlations in non-integrable Ising models, Journal of Mathematical Physics 66 (2025), no. 2, 023301, DOI: 10.1063/5.0235381, arXiv:2408.12923, https://arxiv.org/abs/2408.12923
11. Giovanni Antinucci, Alessandro Giuliani, Rafael L. Greenblatt, Non-integrable Ising models in cylindrical geometry: Grassmann representation and infinite volume limit, Annales Henri Poincaré 23 (2022), 1061–1139, DOI: 10.1007/s00023-021-01107-3, arXiv:2104.09328, https://arxiv.org/abs/2104.09328
12. Ulrik Thinggaard Hansen, Frederik Ravn Klausen, The Supercritical Loop O(1) and Random Current models: Uniqueness and Mixing, arXiv:2603.28520 (2026), https://arxiv.org/abs/2603.28520
13. Ulrik Thinggaard Hansen, Frederik Ravn Klausen, The End of the Double Random Current, arXiv:2607.07616 (2026), https://arxiv.org/abs/2607.07616
14. Jianping Jiang, Frederik Ravn Klausen, A simple proof of exponential decay for the near-critical planar Ising model, arXiv:2512.07554 (2025), https://arxiv.org/abs/2512.07554
15. Michael Aizenman, A geometric perspective on the scaling limits of critical Ising and $\varphi^4_d$ models, Current Developments in Mathematics 2020, International Press, 1–39, DOI: 10.4310/CDM.2020.v2020.n1.a1, arXiv:2112.04248, https://arxiv.org/abs/2112.04248
16. Hugo Duminil-Copin, 100 Years of the (Critical) Ising Model on the Hypercubic Lattice, arXiv:2208.00864 (2022), https://arxiv.org/abs/2208.00864
17. Frederik Ravn Klausen, On monotonicity and couplings of random currents and the loop-$\mathrm{O}(1)$-model, ALEA, Lat. Am. J. Probab. Math. Stat. 19 (2022), 151–161, DOI: 10.30757/alea.v19-07, arXiv:2006.15897, https://arxiv.org/abs/2006.15897
18. Michael Aizenman, Simone Warzel, Kac–Ward Formula and Its Extension to Order–Disorder Correlators Through a Graph Zeta Function, Journal of Statistical Physics 173 (2018), 1755–1778, DOI: 10.1007/s10955-018-2184-9, arXiv:1709.06052, https://arxiv.org/abs/1709.06052