betaFeedback

WorksMathematics

Universality of Large-Scale Geometry For Critical Random-Cluster Models

Abstract

The understanding of the behaviour of critical lattice models in two dimensions has made immense progress in the past couple of decades, with conformal invariance as its conjectured end goal. We consider here the critical random-cluster model (or FK-percolation) on a large class of graphs — namely bi-periodic isoradial graphs — with cluster-weight~q∈[1,4]q \in [1,4]. Following up on the establishment of rotational invariance at criticality in \cite{Rotation}, we prove that the large scale behaviour of these models is asymptotically the same on all such graphs. These include the critical homogeneous models on the square, triangular and hexagonal lattices.

Introduction

Motivation

The successful import of the renormalisation group formalism from quantum field theory was a landmark achievement in statistical mechanics1 that, while remaining elusive from a rigorous point of view, yielded two concrete predictions:

  1. Critical statistical mechanics models should be universal in the sense that their scaling limits should not depend on the choice of underlying lattice.

  1. The scaling limits should enjoy conformal symmetries.

In particular, since the planar conformal group is infinite dimensional, scaling limits of 2D statistical mechanics models should enjoy incredibly rich symmetries.

With Schramm’s introduction of the SLE in [28], this conjectural picture took on a much more tangible form, and the subsequent twenty years have seen a staggering amount of progress, establishing the conjecture of conformal invariance in a number of different models (see [1, 2, 4, 6, 16, 17, 23, 29, 30, 31] for a non-exhaustive list).

However, with the exception of [2, 16, 17], these approaches are characterised by exploiting very particular features of discrete holomorphicity in the models and as such, do not offer robust methods for treating other models.

Another approach is to focus on universality. That is, rather than taking a single model and trying to compute its scaling limit directly, take a class of related models and show that all of their possible scaling limits must agree.

Such an approach has also seen some development in the last fifteen years, starting with bond percolation in [20], extending to more general random-cluster models in [11] and lately, being used to establish not only universality across a class of square lattices, but rotational invariance of any possible scaling limit in [10].

In this paper, we combine the results of [10] with some of the conceptual techniques from [20] in order to extend the universality result of [10] to a larger class of graphs. As a special case, we establish large scale rotational invariance of the critical random-cluster model on any bi-periodic isoradial lattice.

Main Result

In order to state our results properly, we give a brief introduction to some of the objects with which we will be working, all of which will be treated in detail in § 2.

We say that an embedded planar graph G=(V,E)G=(V,E) is isoradial if every face of GG is inscribed in a circle of radius 1 and the centre of the corresponding circle lies within that face. We say that GG is bi-periodic if there exist two linearly independent vectors v,w∈Z2v,w\in\mathbb{Z}^{2} such that GG is invariant under translation by vv and ww.

For a finite graph G=(V,E)G=(V,E), the random-cluster model is the probability measure on {0,1}E\{0,1\}^{E} given by

ϕp,q,G[{ω}]=1Zp,q,Gqκ(ω)∏e∈Epeω(e)(1−pe)1−ω(e),\phi_{p,q,G}[\{\omega\}]=\frac{1}{Z_{p,q,G}}q^{\kappa(\omega)}\prod_{e\in E}p_e^{\omega(e)}(1-p_e)^{1-\omega(e)},

where p=(pe)e∈E∈(0,1)Ep=(p_e)_{e\in E}\in(0,1)^{E} is a vector of probabilities, q∈(0,∞)q\in(0,\infty) is called the cluster-weight, κ\kappa is the number of components in ω\omega and ZZ is a normalising constant called the partition function.

To each edge ee of an isoradial graph, we shall attach a certain weight pep_e depending only on the local geometry of the graph in a way that shall be made precise later on. For this choice of weights, we shall simply refer to the corresponding finite volume measures as ϕG\phi_G, the so-called critical model with isoradial weights. By usual weak compactness elements, this gives rise to corresponding infinite-volume measures ϕG\phi_G on infinite isoradial graphs GG, which turn out to be unique in the regime of interest for us, namely q∈[1,4]q\in[1,4]. We will also write ϕδG\phi_{\delta G} for this same measure on the graph GG rescaled by δ>0\delta>0.

An example of an isoradial graph is the square lattice 2Z2\sqrt{2}\mathbb{Z}^{2}; note the rescaling factor 2\sqrt{2} that makes the faces of Z2\mathbb{Z}^{2} be inscribed in circles of radius 1 rather than 1/21/\sqrt{2}. Then, ϕ2Z2\phi_{\sqrt{2}\mathbb{Z}^{2}} is the critical random-cluster measure [3] with edge-parameters q1+q\frac{\sqrt{q}}{1+\sqrt{q}} for all edges, which has been extensively studied in recent years [9, 12, 13, 14, 18, 27].

It is among this class of random-cluster models that universality is expected when q∈[1,4]q\in[1,4] – that is, when the model on Z2\mathbb{Z}^{2} undergoes a continuous phase transition (see [9, 14] with more recent proofs in [27, 18]. That the phase transition is of the same type on all bi-periodic isoradial graphs was proved in [11].

To make a precise statement, we fix a metric that is naturally associated with re-scaled percolation configurations. For two oriented loops, γ,γ′:S1→C\gamma,\gamma': S^1 \to\mathbb{C}, we define their natural distance

d(γ,γ′)=inf⁡φ∥γ∘φ−γ′∥∞,d(\gamma,\gamma') = \inf_{\varphi} \lVert\gamma\circ\varphi- \gamma' \rVert_{\infty},

where φ\varphi ranges over all orientation-preserving homeomorphisms of S1S^1 into itself.

For two sets of oriented loops F\mathcal{F} and G\mathcal{G}, we define their Camia-Newman distance dCNd_{\mathrm{CN}} such that dCN(F,G)≤ϵd_{\mathrm{CN}}(\mathcal{F},\mathcal{G}) \le\epsilon if and only if for every loop γ∈F\gamma\in\mathcal{F} with ∥γ∥∞≤1ϵ\lVert\gamma\rVert_{\infty} \le\frac{1}{\epsilon}, there exists a loop γ′∈G\gamma' \in\mathcal{G} such that d(γ,γ′)<ϵd(\gamma,\gamma') < \epsilon, and vice-versa.2

For convenience, we will also sometimes consider the global3 Camia-Newman metric dglobal CNd_{\mathrm{global\ CN}} which is defined similarly, except that we do not restrict our attention to loops with ∥γ∥∞≤1ϵ\lVert\gamma\rVert_{\infty} \le\frac{1}{\epsilon}.

Finally, for two probability measures ϕ,ϕ′\phi,\phi' on oriented families of loops, write

dCN(ϕ,ϕ′):=inf⁡{ϵ>0∣∃ coupling P of ϕ and ϕ′ with P[dCN(ω,ω′)>ϵ]<ϵ}.d_{\mathrm{CN}}(\phi,\phi') := \inf\left\{ \epsilon> 0 \mid\exists\ \text{coupling }\mathbb{P}\text{ of }\phi\text{ and }\phi'\text{ with }\mathbb{P}[d_{\mathrm{CN}}(\omega,\omega') > \epsilon] < \epsilon\right\}.

In our applications, the loop configurations will be the set of cluster interfaces surrounding primal and dual clusters of a percolation configuration ω\omega on some isoradial graph G\mathrm{G}. The exact definition is given in § 2, see also Figure 4. For now, the only thing to mention is that these loops will always be oriented so that the dual edges lie on their left.

Figure 4. An isoradial graph (G⋄^{\diamond} is in black), with a primal configuration (in red), its dual configuration (in blue) and the loop representation (in orange).

The identification between percolation and loop configurations is one-to-one, and we shall use the Camia-Newman distance between percolation configurations by extension.

Theorem 1.1. Fix q∈[1,4]q \in[1,4]. For any bi-periodic isoradial graph G\mathrm{G}, there exists c>0c > 0 such that, for all δ≤1\delta\le1,

dCN(ϕδG,ϕδ2Z2)≤δc.(1)d_{\mathrm{CN}}(\phi_{\delta\mathrm{G}},\phi_{\delta\sqrt{2}\mathbb{Z}^2}) \le\delta^c. \tag*{(1)}

The assumption of bi-periodicity of G\mathrm{G} above may be weakened, in the spirit of [20], but not completely removed. We believe the result to hold for general isoradial graphs, but our proof does rely on bi-periodicity.

Two issues arise when working with non-periodic graphs: the potential existence of edges of length arbitrarily close to 0 or 2 (this is excluded by the bounded angles property of [20]) and the potential lack of regularity within G\mathrm{G} (such as the existence of a square grid structure, which was termed the square grid property in [20]). To extend Theorem 1.1 to general isoradial graphs, one would need to tackle these difficulties. The authors believe they are best tackled together and that they are surmountable, but do require new ideas.

Corollary 1.2. For any q∈[1,4]q \in[1,4] and any bi-periodic isoradial graph G\mathrm{G}, any sub-sequential limit of ϕδG\phi_{\delta\mathrm{G}} as δ↘0\delta\searrow0 is invariant under rotations.

Proof. Consider a bi-periodic, isoradial graph G\mathbb{G} and a rotation G~\widetilde{\mathbb{G}} of this same graph. As they are both bi-periodic and isoradial, (1) implies

dCN(ϕδG,ϕδG~)⟶δ→00.d_{\mathrm{CN}}\left(\phi_{\delta\mathbb{G}},\phi_{\delta\widetilde{\mathbb{G}}}\right)\underset{\delta\to0}{\longrightarrow}0.

This implies the result. □

It was recently proved in [1] that the critical random-cluster model on Z2\mathbb{Z}^2 with cluster weight q=4q=4 converges to the conformal loop ensemble CLE4\mathrm{CLE}_4. By Theorem 1.1, this extends readily to ϕG\phi_{\mathbb{G}} for any bi-periodic isoradial graph G\mathbb{G}.

Corollary 1.3. Fix q=4q=4 and G\mathbb{G} a bi-periodic isoradial graph. Then ϕδG\phi_{\delta\mathbb{G}} converges, as δ↘0\delta\searrow0, to the full-plane nested CLE4\mathrm{CLE}_4. This includes the homogeneous models on the triangular and hexagonal lattices with critical parameters 1/21/2 and 4/54/5, respectively.

The full-plane CLE4\mathrm{CLE}_4 is the local limit of the process in increasing domains; see [21]. The convergence above holds for the dCNd_{\mathrm{CN}} distance.

We do not claim the convergence in domains of G\mathbb{G}. Indeed, to deduce the convergence in domains from that in the full plane, one requires a form of continuity of the measure in the shape of the domain in which it is defined. Such a continuity for ϕZ2\phi_{\mathbb{Z}^2} is the byproduct of the convergence in [1]. For isoradial graphs, we do not have an a priori proof of such continuity, which is why we restrict Corollary 1.3 to the full plane.

For q=2q=2, the scaling limit of the random-cluster model (also known in this case as the FK-Ising model) was determined on isoradial graphs and beyond in [7, 5] using direct analysis of each model. For q∈[1,2)∪(2,4)q\in[1,2)\cup(2,4), the scaling limit on Z2\mathbb{Z}^2 and isoradial graphs alike remains conjectural.

Structure of the Paper

Section 2 is dedicated to the necessary background material for the paper. In particular, we review isoradial graphs and the associated critical random-cluster models as well the star-triangle transform and its consequences. The latter is our main tool for comparing a priori different models.

Section 3 then focuses on a particular composite type of transformation, which we will call a cube flip; it is essentially a large-scale star-triangle transformation. It is reminiscent of the interchanging of square lattices in [10], but in finite volume. The estimates of this section are crucial, but follow arguments similar to [10]. To avoid repeating tedious details, proofs here will be sketched, but may be followed without prior knowledge of [10].

Section 4 turns to the proof of Theorem 1.1. First, the graph is transformed so as to be composed of large cubes; these transformations are mesoscopic, and therefore incur an error which is easily controlled. Then, rapid fire applications of the cube flip of Section 3 transform this graph into a square lattice, and the estimates obtained there easily bound the ensuing errors.

Background

Isoradial Graphs

In this section, we introduce the concepts related to isoradial graphs which we will need throughout the paper. First, let us establish a bit of standard notation: Z2\mathbb{Z}^2 will henceforth always refer to the square lattice in its usual isoradial embedding, which is the rotation of 2Z2\sqrt{2}\mathbb{Z}^2 by π4\frac{\pi}{4}. Set ΛN\Lambda_N to the Euclidean box [−N,N]2[-N,N]^2; when working with a fixed graph GG, we will sometimes identify ΛN\Lambda_N to the subgraph of GG it contains.

One may note that any isoradial embedding of GG also gives rise to an isoradial embedding of its dual graph G∗=(V∗,E∗)G^*=(V^*,E^*) by embedding dual vertices as the centres of the circles inscribing the faces of GG.

To any isoradial graph GG, we associate a rhombic tiling of the plane G⋄=(V⋄,E⋄)G^\diamond=(V^\diamond,E^\diamond) as follows: V⋄=V∪V∗V^\diamond=V\cup V^* and we draw an edge between every centre v∗∈V∗v^*\in V^* of a face and each of the vertices v∈Vv\in V incident to that face. It is easy to check that G⋄G^\diamond is an embedded planar graph and that each of its faces is a rhombus of side-length 1. Note that (Z2)⋄=Z2(\mathbb{Z}^2)^\diamond=\mathbb{Z}^2.

Thus, to every isoradial graph, we can attach a certain rhombic tiling of the plane and any rhombic tiling gives rise to exactly two different isoradial graphs with the only ambiguity being which vertices to consider primal and which to consider dual. See Figure 1.

An isoradial domain and its various representations

Figure 1. An isoradial domain and its various representations. To the left, we see the graph in red, in the middle, its dual graph superimposed in blue and the associated rhombic tiling in black and to the right, we see the rhombic tiling, the angle associated to an edge and two tracks of the graph.

In this way, each edge e∈Ee\in E is contained in a unique rhombus R(e)∈(V⋄)∗R(e)\in(V^\diamond)^* and we define the angle θe\theta_e to be the interior angle of R(e)R(e) at one of the vertices opposite of ee (which is well-defined by symmetry).

For a given rhombic tiling G⋄G^\diamond, define a track of G⋄G^\diamond to be a sequence of parallel edges (en)n∈Z⊆E⋄(e_n)_{n\in\mathbb{Z}}\subseteq E^\diamond such that for every nn, ene_n and en+1e_{n+1} are edges of the same rhombus R∈(V⋄)∗R\in(V^\diamond)^* while ene_n and en+2e_{n+2} are not. See Figure 1. The transverse angle of the track (en)n∈Z(e_n)_{n\in\mathbb{Z}} is the angle between each of the ene_n and the horizontal axis; we will generally take it to be in [0,π)[0,\pi).

We say that two distinct tracks t=(en)n∈Nt=(e_n)_{n\in\mathbb{N}} and t′=(fn)n∈Nt'=(f_n)_{n\in\mathbb{N}} intersect if there exist n,mn,m such that en,fme_n,f_m are part of the same rhombus RR. We say that two tracks tt and t′t' are parallel if they do not intersect.

We will often identify a track with the mid-points znz_n of the corresponding edges ene_n. We will also, without comment, refer to the tracks of G⋄\mathbb{G}^{\diamond} as the tracks of G\mathbb{G} (and thus, G\mathbb{G} and G∗\mathbb{G}^{*} have the same tracks).

We say that a track t=(zn)n∈Zt=(z_n)_{n\in\mathbb{Z}} has an asymptotic direction if there exists a vector vtv_t and a C>0C>0 such that, uniformly in nn and mm,

∣zn−zm−vt(n−m)∣≤C\lvert z_n-z_m-v_t(n-m)\rvert\le C

Proposition 2.1. Suppose that G\mathbb{G} is bi-periodic. Then, every track of G\mathbb{G} has an asymptotic direction.

Remark 2.2. It is worth noting that two tracks with non-parallel asymptotic directions necessarily intersect. As two tracks can intersect at most once, for bi-periodic graphs, tracks with parallel asymptotic directions cannot intersect. Therefore, the above proposition says that for any bi-periodic isoradial graph G\mathbb{G}, the tracks of G\mathbb{G} can be partitioned into a finite number of track types labelled by their asymptotic directions.

Proof. Let t=(zn)n∈Zt=(z_n)_{n\in\mathbb{Z}} be a track of G\mathbb{G} and let vv and ww be the vectors of translation corresponding to the actions of (1,0)(1,0) and (0,1)(0,1) respectively. By the pigeonhole principle, there exist distinct n,m∈Zn,m\in\mathbb{Z} such that znz_n and zmz_m represent the same class in G/(Zv⊕Zw)\mathbb{G}/(\mathbb{Z}v\oplus\mathbb{Z}w). Accordingly, we have (j,k)∈Z2∖{0}(j,k)\in\mathbb{Z}^{2}\setminus\{0\} such that zn−zm=jv+kwz_n-z_m=jv+kw. It follows that tt is invariant under translation by jv+kwjv+kw. It is straightforward to check that jv+kwn−m\frac{jv+kw}{n-m} is an asymptotic direction of tt.

Finally, note that any biperiodic graph also has the bounded angles property [20] — since the set of transverse angles of its tracks is finite — and the square grid property [20].

The Star-Triangle Transform and the Track-Exchange Operator

The core ingredient in our treatment of universality among random-cluster models is the star-triangle transform (also known as the Yang-Baxter relation), which is a bit of integrability found in a staggering number of places throughout the planar statistical mechanics literature. It was first discovered by Kennelly in 1899 in the context of electrical networks [22] before appearing in Onsager’s famous exact solution of the 2D Ising model [25].

In the present section, we shall focus on the purely graphical aspect of the star-triangle transformation, as that is going to guide the surgery procedures of the paper.

We call an isomorphic copy of the complete graph on three vertices a triangle. The set of dual edges to a triangle is called a star. One might note that, in the rhombic tiling description of an isoradial graph, this corresponds to three neighbouring rhombi forming a hexagon. See Figure 2.

Figure 2. To the left we see a triangle in red, its dual star in blue and the associated hexagon of rhombi in G⋄\mathbb{G}^{\diamond}. To the right, the hexagon has been rotated by π\pi, replacing the triangle by a star.

For any isoradial graph G\mathbb{G} and any star or triangle in G\mathbb{G}, we can define the star-triangle transform S(G)S(\mathbb{G}), which replaces the star by a triangle or vice versa. From the rhombic tiling point of view, this corresponds to rotating the three rhombi forming the hexagon by π\pi. Thus, the star-triangle transform preserves isoradiality of the graph.

Now, if two tracks tt and ss run next to each other before intersecting, we may apply multiple consecutive star triangle transforms to locally change the relative position of the two tracks. Such a sequence of transformations is called a finite track-exchange; see Figure 3. As in [10], infinite track-exchanges may also be defined; see also Remark 2.9 below.

Figure 3. A finite track-exchange in a wider isoradial graph G\mathbb{G}. Successive star-triangle transforms applied in the coloured hexagon move the blue rhombus to the left, thus exchanging the two horizontal rows of rhombi.

The Random-Cluster Model

We now introduce the random-cluster model on isoradial graphs in more detail. We also recall its elementary properties that we are going to need later on. We refer to the monographs [8, 19] for thorough introductions to the random-cluster model.

Definition 2.3. For a finite subgraph G=(V,E)G=(V,E) of an isoradial graph G\mathbb{G}, q≥1q \ge1, vector p=pe∈[0,1]E\boldsymbol{p}=p_e \in[0,1]^E and partition ξ\xi of the vertex boundary of GG, the random-cluster measure on GG with boundary condition ξ\xi is the measure ϕp,q,Gξ\phi^{\xi}_{p,q,G} on {0,1}E\{0,1\}^E given by

ϕp,q,Gξ[{ω}]=1Zp,q,Gξqκξ(ω)∏e∈Epeω(e)(1−pe)1−ω(e),\phi^{\xi}_{p,q,G}[\{\omega\}] = \frac{1}{Z^{\xi}_{p,q,G}}q^{\kappa^{\xi}(\omega)}\prod_{e\in E}p_e^{\omega(e)}(1-p_e)^{1-\omega(e)},

where κξ\kappa^{\xi} is the number of connected components in the graph

Gωξ=(V,Eω)/∼ξ,G^{\xi}_{\omega}=(V,E_{\omega})/\sim_{\xi},

Eω:={e∈E∣ω(e)=1}E_{\omega}:=\{e\in E\mid\omega(e)=1\}, ∼ξ\sim_{\xi} is the equivalence relation with equivalence classes given by ξ\xi and Zp,q,GξZ^{\xi}_{p,q,G} is a normalising constant called the partition function.

EωE_{\omega} is called the set of open edges and E∖EωE\setminus E_{\omega} is called the set of closed edges.

Proposition 2.4 (Domain Markov Property). If G1=(V1,E1)⊆G2=(V2,E2)G_1=(V_1,E_1)\subseteq G_2=(V_2,E_2) are two finite subgraphs of an isoradial graph G\mathbb{G}, we write ω1:=ω∣E1\omega_1:=\omega|_{E_1} and ω2:=ω∣E2∖E1\omega_2:=\omega|_{E_2\setminus E_1} and have

ϕp,q,G2ξ[ω1∈A∣ω2]=ϕp,q,G1ξω2[A],\phi_{p,q,G_2}^{\xi}[\omega_1\in A\mid\omega_2]=\phi_{p,q,G_1}^{\xi_{\omega_2}}[A],

where vv and ww belong to the same class of ξω2\xi_{\omega_2} if and only if they are connected by a path in ((V2∖V1),Eω2)/∼ξ((V_2\setminus V_1),E_{\omega_2})/\sim_{\xi}.

Definition 2.5. For an isoradial graph G=(V,E)\mathbb{G}=(\mathbb{V},\mathbb{E}), we say that a probability measure ϕp,q\phi_{p,q} on {0,1}E\{0,1\}^{\mathbb{E}} is an infinite-volume random-cluster measure on G\mathbb{G} if, for any finite subgraph G=(V,E)G=(V,E) of G\mathbb{G}, we write ω1=ω∣E\omega_1=\omega|_E and ω2=ω∣Ec\omega_2=\omega|_{E^c} and almost surely have

ϕp,q[ω1∈A∣ω2]=ϕp,q,Gξω2[A],\phi_{p,q}[\omega_1\in A\mid\omega_2]=\phi_{p,q,G}^{\xi_{\omega_2}}[A],

where vv and ww belong to the same element of ξω2\xi_{\omega_2} if and only if they are connected by a path in (V∖V,Eω2)(V\setminus V,E_{\omega_2}).

For an isoradial graph G=(V,E)\mathbb{G}=(\mathbb{V},\mathbb{E}) and ω∈{0,1}E\omega\in\{0,1\}^{\mathbb{E}} a FK-percolation configuration. We define the dual configuration ω∗\omega^* on G∗\mathbb{G}^* (defined in § 2.1) by

ω∗(e∗)=1−ω(e)∀e∈E,\omega^*(e^*)=1-\omega(e)\qquad\forall e\in\mathbb{E},

where e∗e^* is the edge dual to ee.

The medial graph of G\mathbb{G} is the dual of G⋄\mathbb{G}^{\diamond}. Its vertices are the intersection points between the primal and dual edges of G\mathbb{G}; vertices of the medial graph are connected if they are centres of adjacent rhombi. The loop representation of ω\omega is the partition of the edges of the medial graph into loops and bi-infinite paths in such a way that no path or loop intersects any open edge of ω\omega or open dual edge of ω∗\omega^*. These may be represented as in Figure 4 using bent edges and small turns at the vertices of the medial graph.

The loops are oriented to have the dual open edges on their left. Throughout this paper, the loop representation of ω\omega contains no infinite paths; it is formed entirely of finite loops.

Isoradial weights and the Star-Triangle Transform

In this paper, we shall only be interested in so-called critical models with isoradial weights. These weights are chosen so as to be compatible with the isoradial embedding of the graph and the star-triangle transform. For a given star-triangle transform S\mathrm{S} and ω∈{0,1}E\omega\in\{0,1\}^{\mathbb{E}}, we can associate a random configuration S(ω)∈{0,1}S(E)\mathrm{S}(\omega)\in\{0,1\}^{\mathrm{S}(\mathbb{E})} as follows:

  1. ω(e)=S(ω)(e)\omega(e)=\mathbf{S}(\omega)(e) for every ee not affected by S\mathbf{S}.

  1. For the edges affected by S\mathbf{S}, we pair configurations as indicated in Figure 5. In cases of ambiguity, we select the result with probabilities proportional to the random-cluster weights4. Solid red lines are open edges and dashed blue ones are closed edges.

Pairing of percolation configurations under the star-triangle transform

Figure 5. The pairing of percolation configurations under the star-triangle transform. We have omitted symmetric cases.

Remark 2.6. We may note that the connection field of ω\omega is preserved, meaning that for v,w∈V∩S(V)v,w\in V\cap\mathbf{S}(V), vv is connected to ww in ω\omega if and only if they are in S(ω)\mathbf{S}(\omega).

Define the isoradial weights

fq(θ)={qsin⁡(r(π−θ))sin⁡(rθ)+qsin⁡(r(ϕ−θ))q<4,2π−2θ2π−θq=4,f^{q}(\theta)= \begin{cases} \frac{\sqrt{q}\sin(r(\pi-\theta))}{\sin(r\theta)+\sqrt{q}\sin(r(\phi-\theta))} & q<4,\\ \frac{2\pi-2\theta}{2\pi-\theta} & q=4, \end{cases}

where r=1πcos⁡−1(q/2)r=\frac{1}{\pi}\cos^{-1}(\sqrt{q}/2). The following is essentially due to Kenyon [24], but the precise statement is to be found in [11].

Proposition 2.7 ([11], Proposition 2.3). For every q∈[1,4]q\in[1,4], any finite part GG of an isoradial graph GG and any boundary condition ξ\xi, the star-triangle transform preserves the random-cluster measure with parameters peG=fq(θe)p_e^{G}=f^{q}(\theta_e). I.e., for these edge weights, we have

S(ϕGξ)=ϕS(G)ξ.\mathbf{S}(\phi_{G}^{\xi})=\phi_{\mathbf{S}(G)}^{\xi}.

Remark 2.8. The star-triangle transform neither removes nor adds new boundary vertices so that it does, indeed, make sense to regard the same ξ\xi on both sides of the equation.

Throughout the rest of the paper, we will assume that q∈[1,4]q\in[1,4] is fixed. In this range of parameters, it was shown in [11] that for bi-periodic isoradial graphs GG, there is a unique infinite-volume measure, which we denote by ϕG\phi_G. When GG is not bi-periodic, we will require that it has a unique infinite volume measure — we expect this to be the case for all isoradial graphs in this regime.

Remark 2.9. As shown in [10], if an infinite isoradial graph G\mathbb{G} has a unique infinite volume measure ϕG\phi_{\mathbb{G}} and contains two parallel tracks ss and tt that run next to each other (i.e. no rhombus is contained strictly between ss and tt), one may perform an infinite track-exchange of these two tracks.

More precisely, if we write G′\mathbb{G}' for the isoradial graph obtained by exchanging the rhombi of tt and ss, there exists a random map T\mathrm{T} between configurations on G\mathbb{G} and configurations on G′\mathbb{G}' such that

ω∼ϕG⟹T(ω)∼ϕG′\omega\sim\phi_{\mathbb{G}} \quad\Longrightarrow\quad\mathrm{T}(\omega) \sim\phi_{\mathbb{G}'}

where ϕG′\phi_{\mathbb{G}'} is the unique infinite volume measure on G′\mathbb{G}'. Furthermore, T\mathrm{T} preserves the connectivity between vertices common to G\mathbb{G} and G′\mathbb{G}' – that is, all vertices except those contained between ss and tt. Indeed, T\mathrm{T} may be viewed as a bi-infinite sequence of successive star-triangle transformation. Direct constructions of T\mathrm{T} are also possible, see for instance [26].

Let us briefly remind the reader of the RSW theory, which has played a central role in the study of planar critical percolation models.

Definition 2.10. Let QQ be rotation and translation of [0,4n]×[0,n][0,4n] \times[0,n] for n≥1n \ge1. Let C(Q)\mathcal{C}(Q) be the event that there exists a path of open edges crossing QQ between its two short sides. We say that an isoradial graph G\mathbb{G} has the strong RSW property with parameter c>0c > 0 if, for every QQ as above and n∈Nn \in\mathbb{N}, every event BB depending on the edges at distance at least nn from QQ, we have

ϕG[C(Q)∣B]∈[c,1−c]\phi_{\mathbb{G}}[\mathcal{C}(Q) \mid B] \in[c,1-c]

We write G∈RSW(c)\mathbb{G} \in RSW(c) if the above holds for G\mathbb{G}.

The results in [11] imply that for any bi-periodic isoradial graph G\mathbb{G}, ϕG\phi_{\mathbb{G}} admits a strong RSW property. One standard consequence of RSW is a mixing property. See e.g. [12 Proposition 2.9].

Proposition 2.11 (Mixing). For every c>0c > 0, there exists C>0C > 0 such that for every G∈RSW(c)\mathbb{G} \in RSW(c) and events AA depending only on edges in Λr\Lambda_r and BB depending only on edges outside of ΛR\Lambda_R, we have

∣ϕG[A∩B]−ϕG[A]ϕG[B]∣≤C(rR)cϕG[A]ϕG[B].\left|\phi_{\mathbb{G}}[A \cap B] - \phi_{\mathbb{G}}[A]\phi_{\mathbb{G}}[B]\right| \le C\left(\frac{r}{R}\right)^c \phi_{\mathbb{G}}[A]\phi_{\mathbb{G}}[B].

Extensions of isoradial domains

An isoradial domain DD is a finite, simply connected set of rhombi of a rhombic tiling G⋄\mathbb{G}^{\diamond}. We call the isoradial graph G\mathbb{G} an extension of DD. It is called a natural extension if it is such that there exists a unique infinite volume measure ϕG\phi_{\mathbb{G}} on G\mathbb{G}. In particular, biperiodic graphs are natural extensions.

Proposition 2.12. Let DD be an isoradial domain and G\mathbb{G} and G′\mathbb{G}' be two natural extensions of DD. Then, the marginals of ϕG\phi_{\mathbb{G}} and ϕG′\phi_{\mathbb{G}'} on DD are the same. In fact, even the distribution of the boundary conditions on DD induced by the outside configuration is the same under both measures.

Proof. The proof has two parts. We first define a particular extension G0\mathrm{G}_0 of D\mathcal{D}; we will construct it by adding rhombi to D\mathcal{D} sequentially. There are two stages to the construction: first we convexify D\mathcal{D}, and then we complete it into an isoradial graph in the full plane.

Write t1,…,tkt_1,\ldots,t_k for the tracks traversing D\mathcal{D}. Recall that any rhombus of an isoradial graph corresponds to the intersection of two tracks. Thus, the potential rhombi formed by the tracks t1,…,tkt_1,\ldots,t_k is that of the pairs {ti,tj}\{t_i,t_j\} with i≠ji\ne j. Some of these intersections may appear in D\mathcal{D}, but not necessarily all of them. If {ti,tj}\{t_i,t_j\} corresponds to no rhombus of D\mathcal{D}, we say that the intersection is missing. We say that {ti,tj}\{t_i,t_j\} is accessible if it corresponds to a rhombus which shares two edges with existing adjacent rhombi on the boundary of D\mathcal{D}.

Consider the lexicographical order on the pairs {ti,tj}\{t_i,t_j\} with i<ji<j. Let {ti,tj}\{t_i,t_j\} denote the smallest pair so that {ti,tj}\{t_i,t_j\} is both missing and accessible. If such a pair exists, add the rhombus and restart the procedure. By repeatedly doing this, we arrive at a stage where no such pair of tracks exists. Let D′\mathcal{D}' denote the isoradial domain obtained by this procedure.

We claim that the union of the rhombi of D′\mathcal{D}' form a convex set, and that D′\mathcal{D}' contains all intersections between tracks of distinct transverse angles. Indeed, if no rhombus as above may be added, the edges forming the boundary of D′\mathcal{D}' have angles in [0,2π)[0,2\pi) that form an increasing sequence along the boundary in counterclockwise order (if starting from the lowest angle). In other words, the boundary can be parametrised by a curve which only ever turns to the left, and so, D′\mathcal{D}' is convex. Moreover, it implies that any track enters and exits D′\mathcal{D}' at opposing sides. For this to be true for all finitely many angles appearing, each angle must intersect every other angle. See Figure 6 for an illustration.

Figure 6. Left: The default extension G0\mathrm{G}_0 of an isoradial domain D\mathcal{D}. The convexification D′\mathcal{D}' is surrounded by the bold line. Right: A different extension GΛ\mathrm{G}_{\Lambda} of D\mathcal{D}. The convexified part Λ′\Lambda' is also highlighted; outside of it we use the same track angles as for G0\mathrm{G}_0.

Finally, we construct G0\mathrm{G}_0. Consider two angles θ1≠θ2\theta_1\ne\theta_2 so that none of the tracks t1,…,tNt_1,\ldots,t_N has these particular transverse angles. Complete D′\mathcal{D}' with tracks of transverse angles θ1\theta_1 and θ2\theta_2 so that the tracks t1,…,tkt_1,\ldots,t_k intersect all those with angle θ1\theta_1, but none of those with angle θ2\theta_2.

For simplicity of exposition, we will consider θ1=0\theta_1=0 henceforth. We call the tracks of transverse angle θ1\theta_1 vertical and those of transverse angle θ2\theta_2 horizontal.

Notice that G0\mathrm{G}_0 has the structure of a square lattice outside of D′\mathcal{D}'. As a result, the RSW estimates of [11 Thm. 1.1] apply for rectangles sufficiently far from D′\mathcal{D}', which implies that G0\mathrm{G}_0 has a unique infinite volume FK-percolation measure. Thus, G0\mathrm{G}_0 is a natural extension of D\mathcal{D}. We will consider G0\mathrm{G}_0 to be the default extension of D\mathcal{D}.

In the second part of the proof, we show that any other natural extension of D\mathcal{D} produces the same measure on D\mathcal{D} as ϕG0\phi_{\mathcal{G}_0}. Fix such a natural extension G1\mathcal{G}_1. Let ω0\omega_0 be a configuration on D\mathcal{D} and ε>0\varepsilon> 0. Let Λ\Lambda be a large enough connected part of G1\mathcal{G}_1 so that

max⁡ξ,ξ′∣ϕΛξ[ω=ω0 on D]−ϕΛξ′[ω=ω0 on D]∣≤ε.(2)\max_{\xi,\xi'} \left|\phi_{\Lambda}^{\xi}[\omega=\omega_0\text{ on }\mathcal{D}]-\phi_{\Lambda}^{\xi'}[\omega=\omega_0\text{ on }\mathcal{D}]\right|\leq\varepsilon. \tag*{(2)}

Define now GΛ\mathcal{G}_{\Lambda} as the completion of Λ\Lambda obtained as above, by first convexifying Λ\Lambda to obtain Λ′\Lambda', then completing it with tracks of transverse angles θ1\theta_1 and θ2\theta_2. In light of (2), we find

∣ϕG1[ω=ω0 on D]−ϕGΛ[ω=ω0 on D]∣≤ε.(3)\left|\phi_{\mathcal{G}_1}[\omega=\omega_0\text{ on }\mathcal{D}]-\phi_{\mathcal{G}_{\Lambda}}[\omega=\omega_0\text{ on }\mathcal{D}]\right|\leq\varepsilon. \tag*{(3)}

Next, we apply star-triangle transformations to GΛ\mathcal{G}_{\Lambda} so as to create an arbitrarily large portion around D\mathcal{D} identical to G0\mathcal{G}_0. These transformations will not affect D\mathcal{D}. This is done in two steps.

Step 1: We first show that by applying star-triangle transformations to GΛ\mathcal{G}_{\Lambda} outside of D\mathcal{D}, we may reconstruct D′\mathcal{D}'. See Figure 7 for an illustration of this step.

Figure 7. The procedure used in Step 1 of the transformation of GΛ\mathcal{G}_{\Lambda} into G0\mathcal{G}_0. Left: the part D′\mathcal{D}' of G0\mathcal{G}_0 with A\mathcal{A} in blue and the rhombus RjR_j marked in red. Right: the convex part Λ′\Lambda' of GΛ\mathcal{G}_{\Lambda} with the rhombus R~\widetilde{R} in red. This is the vertex which needs to be moved to the position occupied by RjR_j; we do so by applying star triangle transformations in the delimited region. Here, two transformations suffice.

Write R1,R2,…R_1,R_2,\ldots for the rhombi of D′∖D\mathcal{D}'\setminus\mathcal{D} in the order in which they were added when convexifying D\mathcal{D}. Let jj be maximal so that R1,…,Rj−1R_1,\ldots,R_{j-1} appear in Λ′\Lambda' at the same position as in D′\mathcal{D}'. Finally, write A=D∪R1∪⋯∪Rj−1\mathcal{A}=\mathcal{D}\cup R_1\cup\cdots\cup R_{j-1}.

If A=D′\mathcal{A}=\mathcal{D}', we are done, and we may move on to Step 2. Otherwise, observe that RjR_j is the intersection of two tracks, ss and tt, which cross D\mathcal{D}. These two tracks necessarily intersect in Λ′\Lambda' (as Λ′\Lambda' is convex), and we denote by R~\widetilde{R} the rhombus of GΛ\mathcal{G}_{\Lambda} corresponding to their intersection.

Since A\mathcal{A} is connected, there exists a finite region in GΛ\mathcal{G}_{\Lambda} delimited by ss, tt and A\mathcal{A}. Write B\mathcal{B} for the set of rhombi contained in this region, including those contained on ss and tt, but not those contained in A\mathcal{A}.

Note that since Rj≠R~R_j\ne\widetilde{R} by assumption, ∣B∣>1|\mathcal{B}|>1. Therefore, there necessarily exists a hexagon that, when flipped, reduces the number of rhombi in B\mathcal{B} by at least one. This transformation involves the rhombi in B\mathcal{B}, and therefore does not affect A\mathcal{A}, and in particular not D\mathcal{D}.

Applying such transformations repeatedly, we reduce BB to a single rhombus, which in turn yields a new graph G~Λ\widetilde{G}_{\Lambda} with A~=A∪{Rj}\widetilde{A}=A\cup\{R_j\}. By repeated applications of this argument, AA may be increased until it is equal to D′D^{\prime}.

Step 2: We may now assume that GΛG_{\Lambda} contains D′D^{\prime}, and we wish to apply star-triangle transformations to it, outside of D′D^{\prime}, so as to recreate an arbitrarily large portion of G0G_{0}. See Figure 8 for an illustration of this step.

Figure 8. The procedure used in Step 2 of the transformation of GΛG_{\Lambda} into G0G_{0}. The graph GΛG_{\Lambda} may now be considered to contain D′D^{\prime} within the convex region Λ′\Lambda^{\prime}. The tracks that intersect Λ′\Lambda^{\prime} but not D′D^{\prime} (here, there is only one, marked in red) need to be sent to infinity. We do this by star-triangle transformation and track-exchanges. Here the red track may be moved upwards indefinitely.

For ease of exposition, we will assume that we may perform infinitely many star-triangle transformations and infinite track-exchanges, and will transform GΛG_{\Lambda} into G0G_{0} exactly. This may be rendered formal and re-interpreted to show how GΛG_{\Lambda} may be transformed into G0G_{0} on an arbitrarily large region by a finite number of transformations.

We refer to the tracks of GΛG_{\Lambda} that cross Λ′\Lambda^{\prime} but not D′D^{\prime} as extra tracks. If no such tracks exist, then D′=Λ′D^{\prime}=\Lambda^{\prime}, and GΛ=G0G_{\Lambda}=G_{0}. Assume next that at least one extra track exists. We will eliminate these tracks by “pushing them to infinity” using star-triangle transformations and track-exchanges.

As extra tracks cross neither D′D^{\prime} nor any of the horizontal tracks, they either run above or below D′D^{\prime}. We will limit ourselves to eliminating those running above D′D^{\prime}. Among these, consider the extra track tt with minimal transverse angle; this is the extra track which is topmost to the left of Λ′\Lambda^{\prime} and bottommost to the right. We write s1,s2,…s_{1},s_{2},\ldots for the horizontal tracks above Λ′\Lambda^{\prime}, in vertically increasing order.

Applying star-triangle transformations on the right of tt, this track may be moved to run adjacent to s1s_{1} by a procedure which is very similar to the one used in step 1. Then, applying track-exchanges between (the image of) tt and the horizontal tracks s1,s2,…s_{1},s_{2},\ldots, tt may be sent upwards to infinity. This concludes Step 2.

Observe now that none of the star-triangle transformations and track-exchanges applied above to GΛ\mathcal{G}_{\Lambda} affects the marginal of the measure in D\mathcal{D} or even the boundary conditions on that domain (i.e. the connections between points on the boundary). Moreover, since these transformations modify GΛ\mathcal{G}_{\Lambda} so as to contain an arbitrary large region around D\mathcal{D} that is identical to G0\mathcal{G}_0, and since the latter has a unique infinite-volume measure, we conclude that

ϕGΛ[ω=ω0 on D]=ϕG0[ω=ω0 on D].\phi_{\mathcal{G}_{\Lambda}}[\omega= \omega_0\text{ on }\mathcal{D}] = \phi_{\mathcal{G}_0}[\omega= \omega_0\text{ on }\mathcal{D}].

Combining the above with (3) and using the fact that ε\varepsilon is arbitrary, we conclude that

ϕG1[ω=ω0 on D]=ϕG0[ω=ω0 on D],\phi_{\mathcal{G}_1}[\omega= \omega_0\text{ on }\mathcal{D}] = \phi_{\mathcal{G}_0}[\omega= \omega_0\text{ on }\mathcal{D}],

as desired. The same applies to the boundary conditions induced on D\mathcal{D} by the configuration outside.

Irreducibility of pavings of convex domains

We close this section with a useful combinatorial observation.

Lemma 2.13. Let D\mathcal{D} and D′\mathcal{D}' be two isoradial domains that occupy the same convex set of R2\mathbb{R}^2. Write 2M2M for the length of the boundary of D\mathcal{D} and D′\mathcal{D}'. There exists a sequence S1,…,SLS_1,\ldots,S_L of at most M3M^3 star triangle transformations that may be applied in order to D\mathcal{D} and such that

(SL∘⋯∘S1)(D)=D′.(4)(S_L \circ\cdots\circ S_1)(\mathcal{D}) = \mathcal{D}'. \tag*{(4)}

Proof. It suffices to prove (4) for a specific choice of D′\mathcal{D}', with L≤M3/2L \le M^3/2. This specific choice of D′\mathcal{D}' is defined as follows:

Write C\mathcal{C} for the convex set occupied by D\mathcal{D} and e1,…,e2Me_1,\ldots,e_{2M} for the edges of the boundary of C\mathcal{C}, in counterclockwise order, so that e1,…,eMe_1,\ldots,e_M have increasing angles θ1,…,θM\theta_1,\ldots,\theta_M in [0,π)[0,\pi). For any paving of C\mathcal{C} with rhombi, write tit_i for the track entering through eie_i. Such a track necessarily exits through eM+k+i0−(i−i0)e_{M+k+i_0-(i-i_0)}, where i0i_0 is the first index with θi0=θi\theta_{i_0}=\theta_i and k+i0k+i_0 is the last index with θk+i0=θi\theta_{k+i_0}=\theta_i.

We define the standard paving D′\mathcal{D}' of C\mathcal{C} inductively. Set i0i_0 equal to smallest index jj such that θj=θM\theta_j=\theta_M and set k=M−i0k=M-i_0. We first require that tMt_M runs along the left boundary of C\mathcal{C} from eMe_M to e2M−ke_{2M-k}. This fixes M−k−1M-k-1 rhombi in C\mathcal{C}. The rest of C\mathcal{C} is surrounded by a convex perimeter with angles θ1,…,θM−1,θ1,…,θM−1\theta_1,\ldots,\theta_{M-1},\theta_1,\ldots,\theta_{M-1}. Proceeding by induction, this constructs the standard paving.

Now that D′\mathcal{D}' is defined, we explain how to pass from D\mathcal{D} to D′\mathcal{D}' using star-triangle transformations. This is again done by induction. Consider the track tMt_M of D\mathcal{D}. If this is in the same position as in D′\mathcal{D}', fix this region and proceed with transforming the rest of the domain, which corresponds to the case M−1M-1.

If tMt_M in D\mathcal{D} differs from its position in D′\mathcal{D}', “push” it to the left. Indeed, our assumption implies the existence of at least one rhombus in D\mathcal{D} which lies strictly to the left of tMt_M. One may then find such a rhombus RR that has two sides adjacent to tMt_M. For instance, one may orient the tracks t1,…,tM−kt_1,\ldots,t_{M-k} from the left of tMt_M towards the right, inducing a partial order on rhombi to the left of tMt_M. Then, pick RR to be any maximal rhombus of this partial ordering. See Figure 2.6. Applying a star triangle-transform to RR and its two neighbouring rhombi on tMt_M, we obtain a graph S(D)\mathbf{S}(\mathcal{D}) with RR to the right tMt_M. Repeating this procedure pushes tMt_M to the left, in the same position as in $\mathcal{D}'`.

Figure 9. A convex set C\mathcal{C} with its standard paving on the left. In the middle, another paving of C\mathcal{C}, where tMt_M is not the leftmost track. This implies the existence of a hexagon which may be flipped with the star-triangle transform and moved to the right of tMt_M, such as the intersection of the blue and green track, which is maximal for the ordering of rhombi given by the orientations of the tracks t1,…,tM−1t_1,\ldots,t_{M-1} from left to right.

The number of star-triangle transformations needed to push tMt_M to the left is bounded by the overall number of rhombi in D\mathcal{D}, which is at most M(M−1)/2M(M-1)/2 - corresponding to the case where all angles are distinct. Since this procedure is then applied to tM−1t_{M-1} in the remaining part, then to tM−2t_{M-2} etc, we conclude that the total number of star-triangle transformations needed to pass from D\mathcal{D} to D′\mathcal{D}' is bounded above by

∑l=1Ml(l−1)2≤M32.\sum_{l=1}^{M} \frac{l(l-1)}{2} \le\frac{M^3}{2}.

This concludes the proof.

The Cube Flip Lemma

In the context of isoradial graphs, the star-triangle transformation may be viewed as the permutation of three adjacent rhombi which together form a hexagon. Here, we consider a large-scale version of this transformation and describe its effect on the configuration.

Given three unit vectors vv, ww and zz not contained in the same closed half-plane, we define the cube C(v,w,z,N)\mathcal{C}(v,w,z,N) to be the isoradial domain obtained by gluing translates of three N×NN \times N rhombi, each paved with N2N^2 unit rhombi with sides v×wv \times w, w×zw \times z and z×vz \times v, respectively, as in Figure 10, left diagram. Furthermore, we define the flipped cube CF(v,w,z,N)=C(−v,−w,−z,N)\mathcal{C}^{F}(v,w,z,N)=\mathcal{C}(-v,-w,-z,N). If G\mathbb{G} is an extension of C(v,w,z,N)\mathcal{C}(v,w,z,N), write GF\mathbb{G}^{F} for the same graph, but with C(v,w,z,N)\mathcal{C}(v,w,z,N) replaced by the flipped cube CF(v,w,z,N)\mathcal{C}^{F}(v,w,z,N).

Figure 10. Left: a cube with the block of old rhombi in yellow, the new ones in red and the transverse ones in blue. Right: the flipped cube, with the three blocks exchanged. Middle: a stage GkNG_{kN} in the transformation from the initial to the flipped cube. The frozen block is hashed and the active one appears bright. To get to G(k+1)NG_{(k+1)N}, all new rhombi in the active block are pushed down by track-exchanges. The blue transverse rhombi are also pushed to the left side.

As vv, ww, zz and NN will mostly be fixed in this section, we will omit them from notation when no ambiguity is possible. The purpose of this section is to prove the following:

Theorem 3.1 (Cube flip). For any three unit vectors zz, vv, ww not contained in the same closed half-plane, there exists a constant c>0c>0 with the following property: For any N≥1N\ge1 and any natural extension G\mathbb{G} of C(v,w,z,N)\mathcal{C}(v,w,z,N), there exists a coupling P\mathbb{P} between ω1∼ϕG\omega_1\sim\phi_{\mathbb{G}} and ω2∼ϕGF\omega_2\sim\phi_{\mathbb{G}^{F}} such that

(i) ω1(e)=ω2(e)\omega_1(e)=\omega_2(e) for any edge ee outside of C\mathcal{C};

:::`

(ii) the connections induced by ω1∩C\omega_1 \cap C on the vertices of ∂C\partial C are the same as those induced by ω2∩CF\omega_2 \cap C^{\mathsf{F}};

(iii)

P[dglobal CN(ω1,ω2)≥N1−c]≤1Nc.\mathbb{P}\left[d_{\mathrm{global}\,\mathrm{CN}}(\omega_1,\omega_2) \ge N^{1-c}\right] \le\frac{1}{N^c}.

The proof of the above requires a slight re-working of the results of [10]. Indeed, the flipping of the cube from CC to CFC^{\mathsf{F}} may be done by a series of star-triangle transformations that resemble the track-exchanges of [10]. Some additional details need to be considered due to the boundary effects that appear when flipping the cube, rather than infinite horizontal tracks. Given the complexity of the argument in [10], we will give a short sketch below. The rest of the section is dedicated to the proof of Theorem 3.1.

RSW theory in the cube

In the proof of Theorem 3.1, the RSW theory will be used inside CC. Notice that C(v,w,z,N)C(v,w,z,N), or rather its infinite version C(v,w,z,∞)C(v,w,z,\infty), is not biperiodic, and therefore, the results of [11, 20] do not apply as such. Nevertheless, the RSW theory inside C(v,w,z,∞)C(v,w,z,\infty) is deduced easily from [11].

Lemma 3.2 (RSW in infinite cubes). For every v,w,zv,w,z as above, C(v,w,z,∞)C(v,w,z,\infty) satisfies the strong RSW property RSW⁡(c)\operatorname{RSW}(c) for some c>0c > 0.

The constant c>0c > 0 above may be shown to be independent of v,w,zv,w,z, but this is not important for our purposes.

A piece of notation will be useful: we write L(v,w,z)\mathbb{L}(v,w,z) for the isoradial embedding of the square lattice formed of one half-plane paved by rhombi of sides v,zv,z and the other half-plane paved by rhombi of sides w,zw,z; the common boundary of the half-planes is the line with direction zz passing through 00.

Proof. Let QQ be a rectangle as in Definition 2.10 and consider the three pieces Q1Q^1, Q2Q^2, Q3Q^3 of QQ lying inside L(v,w,z)\mathbb{L}(v,w,z), L(w,z,v)\mathbb{L}(w,z,v) and L(z,v,w)\mathbb{L}(z,v,w) respectively. These are finite, potentially empty, subgraphs of square lattices, which have the RSW property by [11].

Now, deterministically, some QjQ^j must contain a translated, rotated and scaled copy of [0,4]×[0,12][0,4] \times[0,\frac{1}{2}], and thus, one may cross QQ simply by crossing QjQ^j. This proves the desired by Proposition 2.12 and the RSW property of the associated square lattice. ∎

For convenience, we will avoid proving Theorem 3.1 for general extensions of the cube and focus on the special case of the infinite cube.

Lemma 3.3 (Reduction to the infinite cube). If Theorem 3.1 holds with G=C(v,w,z,∞)\mathbb{G}=\mathcal{C}(v,w,z,\infty), then it holds for any natural extension G\mathbb{G}.

Proof. Fix v,w,zv,w,z and NN as in the statement of Theorem 3.1. Assume Theorem 3.1 holds for G=C(v,w,z,∞)\mathbb{G}=\mathcal{C}(v,w,z,\infty), where we flip only C(v,w,z,N)\mathcal{C}(v,w,z,N). Finally, fix a natural extension G\mathbb{G} of C(v,w,z,∞)\mathcal{C}(v,w,z,\infty).

Let ω1∼ϕC(v,w,z,∞)\omega_1\sim\phi_{\mathcal{C}(v,w,z,\infty)} and ω2∼ϕ(C(v,w,z,∞))F~\omega_2\sim\phi_{(\mathcal{C}(v,w,z,\infty))^{\widetilde{\mathsf{F}}}} (where, in the latter, we only flip C(v,w,z,N)\mathcal{C}(v,w,z,N)) be coupled so as to satisfy (i), (ii), and (iii) of Theorem 3.1. By Proposition 2.12, ω1\omega_1 may also be coupled to ω∼ϕG\omega\sim\phi_{\mathbb{G}} in such a way that

  • ω(e)=ω1(e)\omega(e)=\omega_1(e) for ee in C(v,w,z,N)\mathcal{C}(v,w,z,N),

  • for all v,w∈C(v,w,z,N)v,w\in\mathcal{C}(v,w,z,N), we have v⟷ωwv\mathrel{\mathop{\longleftrightarrow}^{\omega}}w if and only if v⟷ω1wv\mathrel{\mathop{\longleftrightarrow}^{\omega_1}}w and

  • ω\omega is conditionally independent of ω2\omega_2 given ω1\omega_1.

Define now a configuration ω~\widetilde{\omega} on GF~\mathbb{G}^{\widetilde{\mathsf{F}}} by

ω~(e)={ω(e)e∉C(v,w,z,N),ω2(e)e∈C(v,w,z,N).\widetilde{\omega}(e)= \begin{cases} \omega(e) & e\notin\mathcal{C}(v,w,z,N),\\ \omega_2(e) & e\in\mathcal{C}(v,w,z,N). \end{cases}

Then, ω~∼ϕGF~\widetilde{\omega}\sim\phi_{\mathbb{G}^{\widetilde{\mathsf{F}}}} by Proposition 2.12 and the fact that the coupling between ω1\omega_1 and ω2\omega_2 satisfies (ii). Furthermore, the coupling of ω\omega and ω~\widetilde{\omega} satisfies (i), (ii) and (iii) because that of ω1\omega_1 and ω2\omega_2 does.

Nail lattices and homotopy distance

Fix unit vectors v,w,zv,w,z not belonging to the same half-plane and N≥1N\geq1 large. With Lemma 3.3 in hand, we now turn to proof of Theorem 3.1 for G=C(v,w,z,∞)\mathbb{G}=\mathcal{C}(v,w,z,\infty). Write C=C(v,w,z,N)\mathcal{C}=\mathcal{C}(v,w,z,N).

We will encode the distance dCNd_{\mathcal{C}_{N}} via the notion of nails and homotopy classes — see [10 Sec. 3.5]. We give a brief description below for completeness.

Several scales will come into play, all depending on NN, the size of the cube. In principle, they are independent, but we will choose them as follows for some small ε>0\varepsilon>0:

  • R=N1−ε3R=N^{1-\varepsilon^3} is the distance between the nails;

  • rmax⁡=N1−ε2r_{\max}=N^{1-\varepsilon^2} is the maximal distance between a nail and the corresponding lattice point;

  • rmin⁡=N1−εr_{\min}=N^{1-\varepsilon} is the scale of mesoscopic clusters;

  • rIIC=N1−Cεr_{\mathrm{IIC}}=N^{1-C\varepsilon} is the scale at which we will compare extrema of nails with the IIC; CC is a large constant chosen as in [10 eq. (18)]$.

Once these quantities are fixed, we define the following notions: We will henceforth assume that ε\varepsilon is sufficiently small and NN is large, in particular so that rIIC≫1r_{\mathrm{IIC}}\gg1.

The lattice L\mathcal{L} is the set of points of RZ2R\mathbb{Z}^{2} contained in CC, at a distance at least RR from CcC^{c}. For a configuration ω\omega on GG and a finite cluster CC of ω\omega, define its extremal values in the directions −v-v and −z-z by

T(C)=max⁡y∈C⟨−v,y⟩;B(C)=min⁡y∈C⟨−v,y⟩;L(C)=min⁡y∈C⟨−z,y⟩;R(C)=max⁡y∈C⟨−z,y⟩.\mathrm{T}(C)=\max_{y\in C}\langle-v,y\rangle;\quad\mathrm{B}(C)=\min_{y\in C}\langle-v,y\rangle;\quad\mathrm{L}(C)=\min_{y\in C}\langle-z,y\rangle;\quad\mathrm{R}(C)=\max_{y\in C}\langle-z,y\rangle.

The points realising these minimas and maximas are called the extrema of CC. If multiple points satisfy one of the conditions, we choose one according to an arbitrary rule5.

We say that CC is a nail at some point x∈Lx\in\mathcal{L} if

∣T(C)−⟨−v,x⟩∣≤rmax⁡;∣B(C)−⟨−v,x⟩∣≤rmax⁡;\left|\mathrm{T}(C)-\langle-v,x\rangle\right|\le r_{\max};\qquad\left|\mathrm{B}(C)-\langle-v,x\rangle\right|\le r_{\max};
∣L(C)−⟨−z,x⟩∣≤rmax⁡and∣R(C)−⟨−z,x⟩∣≤rmax⁡.\left|\mathrm{L}(C)-\langle-z,x\rangle\right|\le r_{\max}\qquad\text{and}\qquad\left|\mathrm{R}(C)-\langle-z,x\rangle\right|\le r_{\max}.

A lattice of nails N=(N(x))x∈L\mathbf{N}=(N(x))_{x\in\mathcal{L}} is a choice of one nail N(x)N(x) for each lattice point x∈Lx\in\mathcal{L}. Note that there need not be a unique lattice of nails a given configuration - and there might not even be one at all.

Given a configuration and a lattice of nails N=(N(x))x∈L\mathbf{N}=(N(x))_{x\in\mathcal{L}}, we will encode the large loops and paths of ω∩C\omega\cap C as follows: For each x∈Lx\in\mathcal{L}, let uxu_x be a point in N(x)N(x). For an edge xyxy of L\mathcal{L}, let [ux,uy][u_x,u_y] be the oriented segment of the straight line between uxu_x and uyu_y.

For a simple oriented curve γ\gamma with image in R2∖{ux:x∈L}\mathbb{R}^{2}\setminus\{u_x:x\in\mathcal{L}\} let [ux1,uy1],…,[uxK,uyK][u_{x_1},u_{y_1}],\ldots,[u_{x_K},u_{y_K}] be the sequence of oriented segments crossed by γ\gamma; the segments are oriented from left to right as γ\gamma crosses them. The word associated to γ\gamma is the subsequence of (x1,y1),…,(xK,yK)(x_1,y_1),\ldots,(x_K,y_K) obtained by inductively removing all consecutive pairs of the form (x,y)(x,y), (y,x)(y,x). When γ\gamma is an oriented loop, we consider the word modulo cyclic permutations. A word and, by extension, the path or loop it encodes is called trivial if the associated word contains at most one letter.

Recall that the ω\omega is encoded by an infinite family of oriented loops. Write [ω]N[\omega]_{\mathbf{N}} for the multi-set of non-trivial words of all loops associated to ω\omega that intersect CC. We call [ω]N[\omega]_{\mathbf{N}} the homotopy information of ω\omega with respect to the lattice of nails N\mathbf{N}.

The following lemma shows how the homotopy information gives information controls the distance dCNd_{\mathrm{CN}}.

Lemma 3.4. Let P\mathbb{P} be a coupling between ω1∼ϕG\omega_1\sim\phi_G and ω2∼ϕGF\omega_2\sim\phi_{G_F} with ω1∣Cc=ω2∣Cc\omega_1|_{C^c}=\omega_2|_{C^c}. Then, for any ε>0\varepsilon>0 there exists δ>0\delta>0, independent of NN, such that

P[dglobal CN(ω1,ω2)≤N1−δ]≤P[ω1,ω2 contain lattices of nails N1,N2 with [ω1]N1=[ω2]N2]+1Nδ.\mathbb{P}\left[d_{\mathrm{global\ CN}}(\omega_1,\omega_2)\le N^{1-\delta}\right] \le\mathbb{P}\left[\omega_1,\omega_2\text{ contain lattices of nails }\mathbf{N}_1,\mathbf{N}_2\text{ with }[\omega_1]_{\mathbf{N}_1}=[\omega_2]_{\mathbf{N}_2}\right]+\frac{1}{N^{\delta}}.

Proof. The result is a straightforward extension of [10 Prop. 3.9].

Coupling for the cube flip

The cube flip may be viewed as a composition of N2N^{2} finite track-exchanges, where each finite track-exchange is the sequential application of NN star-triangle transformations involving the two same tracks of transverse angles vv and ww and NN tracks of transverse angle zz. The track-exchanges will be grouped into 2N−12N-1 batches, each batch containing simultaneous, non-overlapping applications of several finite track-exchanges.

To fix ideas, we will call the rhombi with sides vv and zz the old rhombi (yellow in Figure 10), those of sides ww and zz the new rhombi (red in Figure 10) and the others the transverse rhombi (blue in Figure 10). We will perform the cube flip by exchanging old and new rhombi via track-exchanges. These are described on G\mathbb{G}, but only affect the region in CC. We will group them as follows:

Fix G0=G\mathbb{G}_0=\mathbb{G}. Suppose Gt\mathbb{G}_t was constructed. To obtain Gt+1\mathbb{G}_{t+1}, exchange any horizontal track of old rhombi with a horizontal track of new rhombi if the two tracks are adjacent, and the track of new rhombi is on the top. Let StS_t denote the finite composition of star-triangle transformations forming these multiple track-exchanges.

The rhombi in CC are gradually permuted, but remain in the same hexagonal region of the plane, which we will always call C⊆GtC\subseteq\mathbb{G}_t for clarity. In any Gt\mathbb{G}_t, CC contains an active block formed of alternating horizontal tracks of old and new rhombi and up to four frozen blocks, formed entirely of old, new or transverse rhombi. The latter type of rhombi form two blocks, one to the left and one to the right of the active block. The old and new blocks lie below and above the active block, and exchange positions after NN transformations. In G2N\mathbb{G}_{2N}, all old rhombi live above all new rhombi, and the process terminates. We have G2N=Gf\mathbb{G}_{2N}=\mathbb{G}^{\mathrm{f}}.

These transformations will also act on configurations to produce a Markov chain of configurations ωt\omega_t sampled according to ϕGt\phi_{\mathbb{G}_t}. To define the chain, we add a piece of notation.

A cluster CC in some configuration ωt\omega_t on some lattice Gt\mathbb{G}_t is called mesoscopic if

C⊆Candrmin⁡≤T(C)−B(C)≤rmax⁡.C\subseteq\mathcal{C}\qquad\text{and}\qquad r_{\min}\leq\mathrm{T}(C)-\mathrm{B}(C)\leq r_{\max}.

Write Meso⁡(ωt)\operatorname{Meso}(\omega_t) for the set of mesoscopic clusters of ωt\omega_t and

H(ω)=(T(C),B(C),L(C),R(C))C∈Meso⁡(ω)\mathcal{H}(\omega)=\left(\mathrm{T}(C),\mathrm{B}(C),\mathrm{L}(C),\mathrm{R}(C)\right)_{C\in\operatorname{Meso}(\omega)}

for the extremal coordinates of all mesoscopic clusters.

We now give the explicit construction of the chain (ωt)t=0,…,2N(\omega_t)_{t=0,\ldots,2N}. Sample ω0∼ϕG0\omega_0\sim\phi_{\mathbb{G}_0}. For t≥0t\geq0 even, given ω0,…,ωt\omega_0,\ldots,\omega_t, we first sample an intermediate configuration ωt+1/2∼ϕGt\omega_{t+1/2}\sim\phi_{\mathbb{G}_t} as follows:

Consider the boxes ΛrIIC(zj)\Lambda_{r_{\mathrm{IIC}}}(z_j) around the extrema zjz_j of mesoscopic clusters. We say that an overlap error occurs (in ωt\omega_t) if any two such boxes overlap, if any such box intersects more than one cluster of diameter rmin⁡/4r_{\min}/4, if any such box intersects more than one block, or if any such box is at a distance less than rmin⁡r_{\min} from CcC^{c}. In this case, we set ωt+1/2=ωt\omega_{t+1/2}=\omega_t. Otherwise, set ωt+1/2=ωt\omega_{t+1/2}=\omega_t for all edges outside of the boxes ΛrIIC(zj)\Lambda_{r_{\mathrm{IIC}}}(z_j) and resample the configuration in each box according to the measure

ϕGt[ ⋅∣ω=ωt outside ⋃jΛrIIC(zj) and H(ω)=H(ωt) ].\phi_{\mathbb{G}_t}\left[\,\cdot\mid\omega=\omega_t\text{ outside }\bigcup_j\Lambda_{r_{\mathrm{IIC}}}(z_j)\text{ and }\mathcal{H}(\omega)=\mathcal{H}(\omega_t)\,\right].

Now, independently of the definition of ωt+1/2\omega_{t+1/2}, set

ωt+1=St(ωt+1/2)andωt+2=St+1(ωt).\omega_{t+1}=S_t(\omega_{t+1/2})\qquad\text{and}\qquad\omega_{t+2}=S_{t+1}(\omega_t).

The resampling procedure performed between ωt\omega_t and ωt+1/2\omega_{t+1/2} is described in more detail in [10 Sec. 3.2]. It is also explained why we resample the environment around the extrema only at even steps.

With these definitions, the main step of the proof is the following statement, corresponding to [10 Prop 3.7]: Denote by Meso⁡+(ω0)\operatorname{Meso}^{+}(\omega_0) the set of clusters C of ω0\omega_0 such that

dist⁡(C,Cc)≥rmax⁡and2rmin⁡≤T(C)−B(C)≤rmax⁡2.\operatorname{dist}(C,C^{c}) \ge r_{\max} \quad\text{and}\quad2r_{\min} \le\mathrm{T}(C)-\mathrm{B}(C) \le\frac{r_{\max}}{2}.

For a cluster CtC_t of ωt\omega_t, we denote by Ct+1C_{t+1} its coupled cluster in ωt+1\omega_{t+1} with the convention that Ct+1=∅C_{t+1}=\varnothing if the cluster ceases to exist. We call a cluster C of ω0\omega_0 stable if for all t∈[0,2N]t \in[0,2N], we have Ct≠∅C_t \ne\varnothing and furthermore,

∣A(C0)−A(Ct)∣≤110rmax⁡for all A∈{T,B,L,R}.\lvert A(C_0)-A(C_t)\rvert\le\frac{1}{10}r_{\max}\qquad\text{for all } A \in\{\mathrm{T},\mathrm{B},\mathrm{L},\mathrm{R}\}.

The above corresponds to [10 Def. 3.5]. The constant 1/101/10 above is arbitrary, and should just be taken sufficiently small. Note that, as opposed to [10 Def. 3.5], we do not mention the drifts Drift⁡vert\operatorname{Drift}_{\mathrm{vert}} for A=T,BA=\mathrm{T},\mathrm{B} and Drift⁡lat\operatorname{Drift}_{\mathrm{lat}} in the definition of stability. Indeed, these were proved in [10 Lem. 4.4] to be zero.

Proposition 3.5. For ε>0\varepsilon>0 small enough, there exists δ=δ(ε)>0\delta=\delta(\varepsilon)>0 so that

P[all clusters C∈Meso⁡+(ω0) are stable]≥1−1Nδ.(5)\mathbb{P}\left[\textit{all clusters } C\in\operatorname{Meso}^{+}(\omega_0)\textit{ are stable}\right]\ge1-\frac{1}{N^{\delta}}. \tag*{(5)}

Proof. The proof of (5) is the same as that of [10 Prop 3.7], with the added input that the drift in both vertical and lateral directions is zero [10 Lem. 4.4].

The only minor difference with [10 Prop 3.7] is the presence of transfer blocks. Indeed, in [10 Prop 3.7], horizontal tracks are infinite, and no transfer blocks exist; only the old and new blocks do.

We have incorporated in the notion of overlap errors the situation when an extremum box intersects the interface between the active and transverse blocks. As such, controlling the expectation of increments for extrema close to such interfaces is not necessary. Also, the probability of an overlap error occurring at a given step may be bounded exactly as in [10 Prop 3.7], with the addition of a factor accounting for the case where the overlap error occurs due to the transverse block.

Another small detail to take into account is in the definition of τstart(C)\tau_{\mathrm{start}}(C) and τend(C)\tau_{\mathrm{end}}(C) for clusters C∈Meso⁡+(ω0)C\in\operatorname{Meso}^{+}(\omega_0). These are roughly the times when the cluster enters and exits, respectively, the active block. In our setting, a point in the plane may enter or exit the active block by transitioning to the transverse blocks. As such, one needs to take both coordinates of the cluster into account when defining τstart(C)\tau_{\mathrm{start}}(C) and τend(C)\tau_{\mathrm{end}}(C), rather than just the vertical coordinate, as was the case in [10 Prop 3.7].

These are only technical difficulties which we will not detail. We will however note that, in this adaptation, the fact that the drifts are null makes all estimates easier.

Finally, following [10], we state a simple consequence of the strong RSW property:

Lemma 3.6. For ε>0\varepsilon>0 small enough, there exists δ=δ(ε)>0\delta=\delta(\varepsilon)>0 so that

P[ω0 contains a lattice of nails in Meso⁡+(ω0)]≥1−1Nδ.\mathbb{P}\left[\omega_0\text{ contains a lattice of nails in }\operatorname{Meso}^{+}(\omega_0)\right]\ge1-\frac{1}{N^{\delta}}.

Proof. The proof is identical to [10 Lem. 3.11].

We are now ready to prove Theorem 3.1 for G=C(v,w,z,∞)\mathbb{G}=\mathcal{C}(v,w,z,\infty).

Proof of Theorem 3.1 for G=C(v,w,z,∞)\mathbf{G}=\mathbf{C}(v,w,z,\infty). Combining Proposition 3.5 and Lemma 3.6 we find that, under the coupling P\mathbb{P} above, for ε>0\varepsilon>0 small enough, there exists δ>0\delta>0 such that

P[ω0 contains a lattice of nails that survives to a lattice of nails of ω2N]≥1−2Nδ.\mathbb{P}\left[\omega_0\text{ contains a lattice of nails that survives to a lattice of nails of }\omega_{2N}\right]\geq1-\frac{2}{N^\delta}.

The construction of the coupling is such that,

  • ω0∼ϕG\omega_0\sim\phi_G, ω2N∼ϕG′\omega_{2N}\sim\phi_{G^\prime}, since, by induction, the laws of the different ωt\omega_t are ϕGt\phi_{G_t};

  • ω0=ω2N\omega_0=\omega_{2N} outside of CC;

  • all connections between vertices outside CC are identical in ω0\omega_0 and ω2N\omega_{2N}: indeed, the transformations St\mathbf{S}_t preserve such connections, and the resampling between ωt\omega_t and ωt+1/2\omega_{t+1/2} is only performed in situations where no such connection may be broken or created (see the definition of overlap error);

  • if N0\mathbf{N}_0 is a lattice of nails in ω0\omega_0 that survives to clusters N2N\mathbf{N}_{2N} in ω2N\omega_{2N}, which also form a lattice of nails, then

[ω0]N0=[ω2N]N2N.[\omega_0]_{\mathbf{N}_0}=[\omega_{2N}]_{\mathbf{N}_{2N}}.

Combining the last display above with Lemma 3.4 provides the desired conclusion after possibly altering δ\delta.

Universality On Bi-Periodic Isoradial Graphs

In this section, we show Theorem 1.1. To be precise, we will prove it in the following form:

Theorem 4.1. Let G\mathbf{G} be a bi-periodic isoradial graph. There exists c>0c>0 such that, for any N≥1N\geq1, there exists a coupling P\mathbb{P} between ω1∼ϕG\omega_1\sim\phi_{\mathbf{G}} and ω2∼ϕZ2\omega_2\sim\phi_{\mathbb{Z}^2} such that

P[dglobal CN(ω1∣ΛN,ω2∣ΛN)>N1−c]<N−c.\mathbb{P}\left[d_{\mathrm{global\,CN}}(\omega_1|_{\Lambda_N},\omega_2|_{\Lambda_N})>N^{1-c}\right]<N^{-c}.

In the above, by ω∣ΛN\omega|_{\Lambda_N}, we mean the family of loops of ω\omega contained in ΛN\Lambda_N. The proof of the above occupies the rest of the section. Before diving in, let us mention how it implies Theorem 1.1.

Proof of Theorem 1.1 given Theorem 4.1. Fix δ>0\delta>0. Let P\mathbb{P} be the coupling of Theorem 4.1 for N=δ−22−cN=\delta^{-\frac{2}{2-c}}, where c>0c>0 is the constant given by Theorem 4.1 — we will assume for simplicity that NN is an integer, otherwise one should round up; we also assume without loss of generality that c<1/4c<1/4.

Observe that6

dglobal CN(δω1∣δΛN,δω2∣δΛN)=δ⋅dglobal CN(ω1∣ΛN,ω2∣ΛN).d_{\mathrm{global\,CN}}(\delta\omega_1|_{\delta\Lambda_N},\delta\omega_2|_{\delta\Lambda_N})=\delta\cdot d_{\mathrm{global\,CN}}(\omega_1|_{\Lambda_N},\omega_2|_{\Lambda_N}).

Then Theorem 4.1 and some standard algebraic manipulations imply

P[dglobal CN(δω1∣Λδ−c/2,δω2∣Λδ−c/2)>δc/4]<δc/2,\mathbb{P}\left[d_{\mathrm{global\,CN}}(\delta\omega_1|_{\Lambda_{\delta^{-c/2}}},\delta\omega_2|_{\Lambda_{\delta^{-c/2}}})>\delta^{c/4}\right]<\delta^{c/2},

which in turn yields (1) for the exponent c/4c/4.

Batched graphs and general strategy

For an isoradial graph G\mathsf{G}, we call a family of tracks t1,…,tKt_1,\ldots,t_K a KK-batch if

  1. the tracks tjt_j do not mutually intersect. Furthermore, they are indexed such that for each jj, the tracks t1,…,tj−1t_1,\ldots,t_{j-1} lie on one side of tjt_j and tj+1,…,tKt_{j+1},\ldots,t_K on the other;

  2. any rhombus between t1t_1 and tKt_K belongs to exactly one of the tracks tjt_j.

We call G\mathsf{G} KK-batched if its tracks may be partitioned into KK-batches. We further call it properly KK-batched if the tracks of each KK-batch have the same transverse angle.

If G\mathsf{G} is an isoradial graph, its associated proper KK-batch, which we write G(K)\mathsf{G}^{(K)}, is the isoradial graph obtained by replacing each track of G\mathsf{G} with KK tracks of the same transverse angle that run next to each other. See Figure 11.

Figure 11. An isoradial domain D\mathcal{D} and the corresponding domain in the proper two-batch G(2)\mathsf{G}^{(2)} (scaled by a factor of 12\frac{1}{2}). Note that stars and triangles of the original rhombic tiling correspond to cubes in G(2)\mathsf{G}^{(2)} to which one may apply the Cube Flip Lemma (Theorem 3.1).

Our proof of Theorem 4.1 is split in two steps. First, we will prove that it applies to any properly KK-batched bi-periodic isoradial graph, where K=N1−δK=N^{1-\delta} for some well chosen δ>0\delta>0. This is done by applying Theorem 3.1. We will then prove that if it applies to G(K)\mathsf{G}^{(K)}, it also does to G\mathsf{G}. This step requires almost no probabilistic estimates.

Proof of Theorem 4.1 for KK-batched graphs

Theorem 4.2. Fix a bi-periodic isoradial graph G\mathsf{G}. There exist C>0C>0 and c∈(0,1)c\in(0,1) such that, for any K≥1K\geq1 and N≥KN\geq K, there exists a coupling P\mathbb{P} between ω1∼ϕG(K)\omega_1\sim\phi_{\mathsf{G}^{(K)}} and ω2∼ϕZ2\omega_2\sim\phi_{\mathbb{Z}^2} such that

P[dglobal CN(ω1∣ΛN,ω2∣ΛN)≥C(NK)3K1−c]≤C(NK)3K−c.\mathbb{P}\left[d_{\mathrm{global\,CN}}\left(\omega_1|_{\Lambda_N},\omega_2|_{\Lambda_N}\right)\geq C\left(\frac{N}{K}\right)^3K^{1-c}\right]\leq C\left(\frac{N}{K}\right)^3K^{-c}.

Remark 4.3. In the above, CC and cc are allowed to depend on G\mathsf{G}. A quick inspection of the proof shows that they actually only depend on the bounded angles property [20] satisfied by G\mathsf{G}. The bi-periodicity of G\mathsf{G} is only needed here to ensure that G\mathsf{G} has this bounded angles property.

We believe that the constants may be taken independently of the bounded angles property, and thus that Theorem 4.2 applies uniformly to all isoradial graphs G\mathsf{G}. This, however, requires controlling cube flips with degenerating angles, which we do not do here.

Proof. Fix NN, KK and a domain D0\mathcal{D}_0 of G\mathcal{G} that contains ΛN/K\Lambda_{N/K} and let D\mathcal{D} denote its convexification. Then, the perimeter 2M2M of D\mathcal{D} satisfies M≤C0N/KM \le C_0N/K for some constant C0C_0 that depends on G\mathcal{G}. Write e1,…,e2Me_1,\ldots,e_{2M} for the edges of the boundary of D\mathcal{D}, in counterclockwise order, so that e1,…,eMe_1,\ldots,e_M have increasing angles in [0,π)[0,\pi). Write θ1,…,θM\theta_1,\ldots,\theta_M for the angles of e1,…,eMe_1,\ldots,e_M.

Next, add rows of 3M3M rhombi left and right of D\mathcal{D}, each with edges with angles 00 and θi\theta_i with i=1,…,Mi=1,\ldots,M except when θi=0\theta_i=0. This corresponds to prolonging the tracks intersecting D\mathcal{D} horizontally, as in Figure 6 and Figure 12. Next, add 3M3M tracks running above the constructed graph, with transverse angle π/2\pi/2, and the same below. Finally, let G\mathcal{G} be the convexification of the graph thus obtained.

Figure 12. Left: a convex domain D\mathcal{D} of an isoradial graph (green), with the completion G\mathcal{G} around it. Right: the convex region occupied by G\mathcal{G}, paved with the rhombic tiling G~\widetilde{\mathcal{G}}. The tracks between SW and NE are blue, those between NW and SE are orange. Note that the green region is entirely covered by a copy of Z2\mathbb{Z}^{2}. A sequence of at most C1M3C_{1}M^{3} star-triangle transformations allows to pass from G\mathcal{G} to G~\widetilde{\mathcal{G}}.

Note that the edges of the perimeter of G\mathcal{G} have angles:

  • (I) θ1,…,θM\theta_1,\ldots,\theta_M, each appearing twice;

  • (II) 00, appearing 12M12M additional times;

  • (III) π/2\pi/2, appearing an additional 12M12M times.

The angles in category (I) different from 00 and π/2\pi/2 are grouped into four sectors, which we denote by NENE, NWNW, SWSW and SESE, respectively. The tracks entering SWSW exit in NENE and those entering NWNW exist in SWSW. Also note that the distance between ΛN/K\Lambda_{N/K} and Gc\mathcal{G}^{c} is at least 3M3M.

Let G~\widetilde{\mathcal{G}} be the paving of the interior of G\mathcal{G} by rhombi constructed as follows:

  • all tracks entering SWSW run along the bottom and right side of the boundary up to their exit in NENE;

  • all tracks entering NWNW run along the top boundary, then left of the tracks in the point above, up to the time they intersect them and exit in SESE;

  • all other parts of G~\widetilde{\mathcal{G}} are paved by square rhombi - that is, rhombi with angles 00 and π/2\pi/2.

See Figure 12 for an example.

Observe that, in G~\widetilde{\mathcal{G}}, ΛN/K\Lambda_{N/K} is entirely contained in the region paved with square rhombi. Indeed, the tracks with traverse angles other than 00 and π/2\pi/2 run along the boundary of G~\widetilde{\mathcal{G}} and hence, remain within distance 2M2M of it.

By Lemma 2.13, there exists a sequence S1,…,SLS_1,\ldots,S_L of star-triangle transformations with

L≤C1M3L \le C_1 M^3

that transforms G\mathcal{G} into G~\widetilde{\mathcal{G}} where C1=133C_1=13^3. Here, we have used the counts in (I), (II) and (III) to determine the length of the perimeter.

Let L\mathbb{L} be a natural extension of G\mathcal{G}. Then, by applying cube flips S1(K),…,SL(K)S_1^{(K)},\ldots,S_L^{(K)} to L(K)\mathbb{L}^{(K)}, corresponding to the star-triangle transformations S1,…,SLS_1,\ldots,S_L, we transform L(K)\mathbb{L}^{(K)} into a graph L~(K)\widetilde{\mathbb{L}}^{(K)} which contains G~(K)\widetilde{\mathcal{G}}^{(K)}. Let P\mathbb{P} be the coupling between ω1∼ϕL(K)\omega_1\sim\phi_{\mathbb{L}^{(K)}} and ω2∼ϕL~(K)\omega_2\sim\phi_{\widetilde{\mathbb{L}}^{(K)}} obtained by these cube flips. By Theorem 3.1, we find that

P[dglobal CN(ω1,ω2)≥C1M3K1−c]≤C1M31Kc.\mathbb{P}\left[d_{\mathrm{global\ CN}}(\omega_1,\omega_2)\ge C_1M^3K^{1-c}\right]\le C_1M^3\frac{1}{K^c}.

Here, we have used the triangle inequality for the global Camia-Newman distance and a union bound.

Finally, since L(K)\mathbb{L}^{(K)} in ΛN\Lambda_{N} is identical to G(K)\mathbb{G}^{(K)}, L~(K)\widetilde{\mathbb{L}}^{(K)} in ΛN\Lambda_{N} is identical to Z2\mathbb{Z}^{2}, and since L(K)\mathbb{L}^{(K)} and L~(K)\widetilde{\mathbb{L}}^{(K)} are natural extensions, Proposition 2.12 implies that we may extend P\mathbb{P} to include ω1′∼ϕG(K)\omega'_1 \sim\phi_{\mathbb{G}^{(K)}} and ω2′∼ϕZ2\omega'_2 \sim\phi_{\mathbb{Z}^{2}} so that

dglobal CN(ω1∣ΛN,ω1′∣ΛN)=0anddglobal CN(ω2∣ΛN,ω2′∣ΛN)=0.d_{\mathrm{global\ CN}}(\omega_1|_{\Lambda_N},\omega'_1|_{\Lambda_N})=0 \quad\text{and}\quad d_{\mathrm{global\ CN}}(\omega_2|_{\Lambda_N},\omega'_2|_{\Lambda_N})=0.

Combining the last two displays, and using the bound on the value of MM, we find

P[dglobal CN(ω1′∣ΛN,ω2′∣ΛN)≥C2(NK)3K1−c]≤C2(NK)3K−c,\mathbb{P}\left[d_{\mathrm{global\ CN}}(\omega'_1|_{\Lambda_N},\omega'_2|_{\Lambda_N}) \ge C_2\left(\frac{N}{K}\right)^3K^{1-c}\right] \le C_2\left(\frac{N}{K}\right)^3K^{-c},

for some G\mathbb{G}-dependent constant C2C_2, as required.

Proof of Theorem 4.1 for general graphs

This section contains the proof of Theorem 4.1. We start with a proposition.

Proposition 4.4. Fix G\mathbb{G} a bi-periodic isoradial graph. For any c>0c>0, there exists δ>0\delta>0 such that, for all N≥1N\ge1, if we set K=⌈N1−c⌉K=\lceil N^{1-c}\rceil, there exists a coupling P\mathbb{P} between ω1∼ϕG\omega_1\sim\phi_{\mathbb{G}} and ω2∼ϕG(K)\omega_2\sim\phi_{\mathbb{G}^{(K)}} such that

P[dglobal CN(ω1∣ΛN,ω2∣ΛN)>N1−δ]<N−δ.\mathbb{P}\left[d_{\mathrm{global\ CN}}(\omega_1|_{\Lambda_N},\omega_2|_{\Lambda_N})>N^{1-\delta}\right]<N^{-\delta}.

Proof. We ignore the integer parts for simplicity. We will transform G\mathbb{G} into G(K)\mathbb{G}^{(K)} using star-triangle transformations and track-exchanges.

A collection of tracks (tn)n∈Z(t_n)_{n\in\mathbb{Z}} is said to be a track type if

  • (i) the tnt_n are mutually parallel;

  • (ii) no other track in G\mathbb{G} is parallel to tnt_n for any nn;

  • (iii) every other track ss intersects the tnt_n in order (i.e. there exists an orientation of ss such that tnt_n is intersected before tn+1t_{n+1} for all nn).

Since GG is bi-periodic, its tracks may be partitioned into a finite number of track types. Write (tn(j))n∈Z(t_n^{(j)})_{n\in\mathbb{Z}} with j=1,…,Jj=1,\ldots,J for these families.

We start off with the case when all tracks of the same type have the same transverse angle. Let us start by batching the tracks of the type (tn(1))n∈Z(t_n^{(1)})_{n\in\mathbb{Z}}. There exists a finite sequence of transformations S1,…,ST\mathrm{S}_1,\ldots,\mathrm{S}_T with the following properties:

  • each St\mathrm{S}_t is the composition of infinitely many, non-overlapping star-triangle transformations;

  • the tracks tnK(1)t_{nK}^{(1)} with n∈Zn\in\mathbb{Z} are not involved in any star-triangle transformation of any St\mathrm{S}_t;

  • ST∘⋯∘S1(G)\mathrm{S}_T\circ\cdots\circ\mathrm{S}_1(G) is bi-periodic, with the tracks (tn(1))n∈Z(t_n^{(1)})_{n\in\mathbb{Z}} being batched into KK-batches.

The St\mathrm{S}_t’s may be constructed by taking limits of the construction in [20 Lemma 7.1]. That the limit exists is a direct consequence of the fact that the star-triangle transformations are non-overlapping and local.

For a configuration ω0∼ϕG\omega_0\sim\phi_G set

G1=ST∘⋯∘S1(G)andω1=ST∘⋯∘S1(ω0).G_1=\mathrm{S}_T\circ\cdots\circ\mathrm{S}_1(G) \qquad\text{and}\qquad\omega_1=\mathrm{S}_T\circ\cdots\circ\mathrm{S}_1(\omega_0).

Note that the connections in ω0\omega_0 and ω1\omega_1 between points on the tracks (tnK(1))n∈Z(t_{nK}^{(1)})_{n\in\mathbb{Z}} are identical. Indeed, the transformations involved in S1,…,ST\mathrm{S}_1,\ldots,\mathrm{S}_T do not affect the points on these track, and thus preserve connections between them. Furthermore, ω1∼ϕG1\omega_1\sim\phi_{G_1}.

Using the same procedure, define sequentially GjG_j from Gj−1G_{j-1} so that each of the families of tracks (tn(1)),…,(tn(j))(t_n^{(1)}),\ldots,(t_n^{(j)}) are batched into KK-batches in GjG_j. To get GjG_j from Gj−1G_{j-1}, we apply star-triangle transformations that do not affect any of the tracks (tnK(j))n∈Z(t_{nK}^{(j)})_{n\in\mathbb{Z}}. Thus, the connections in ωj−1\omega_{j-1} and ωj\omega_j between points on the tracks (tnK(j))n∈Z(t_{nK}^{(j)})_{n\in\mathbb{Z}} are identical.

The following lemma is a direct consequence of the RSW property and the preservation of connections between points of (tnK(j))n∈Z(t_{nK}^{(j)})_{n\in\mathbb{Z}}. Its proof is postponed to the end of the section.

Lemma 4.5. For any fixed c>0c>0, there exists δ>0\delta>0 such that, for all NN large enough and KK and 1≤j≤J1\leq j\leq J as above,

P[dglobalCN(ωj−1∣ΛN,ωj∣ΛN)>N1−δ]<N−δ.\mathbb{P}\left[d_{\mathrm{global}}^{\mathrm{CN}}(\omega_{j-1}|_{\Lambda_N},\omega_j|_{\Lambda_N})>N^{1-\delta}\right]<N^{-\delta}.

We return to the proof of Proposition 4.4. By repeatedly applying the lemma and using a union bound, we conclude that

P[dglobalCN(ω0∣ΛN,ωJ∣ΛN)>J⋅N1−δ]<J⋅N−δ.\mathbb{P}\left[d_{\mathrm{global}}^{\mathrm{CN}}(\omega_0|_{\Lambda_N},\omega_J|_{\Lambda_N})>J\cdot N^{1-\delta}\right]<J\cdot N^{-\delta}.

Recall that ω0∼ϕG\omega_0\sim\phi_G and observe that ωJ∼ϕGJ=ϕG(K)\omega_J\sim\phi_{G_J}=\phi_{G(K)}. This provides the desired conclusion by adjusting the constants in the case where all track of any given type have the same transverse angle.

We now describe how to treat the case where tracks of one type may have different transverse angles. Due to bi-periodicity, the transverse angles of any track type form a periodic sequence; write ℓi\ell_i for the period of the transverse angles of the tracks of type ii. Define G′G' as the graph obtained by batching ℓiK\ell_iK tracks of track type ii for each ii.

The same reasoning as above provides a coupling between ω∼ϕG\omega\sim\phi_{\mathbb{G}} and ω′∼ϕG′\omega' \sim\phi_{\mathbb{G}'} with

P[dglobal CN(ω∣ΛN,ω′∣ΛN)>C⋅N1−δ]<C⋅N−δ,(6)\mathbb{P}\left[d_{\mathrm{global\ CN}}\left(\omega|_{\Lambda_N},\omega'|_{\Lambda_N}\right)>C\cdot N^{1-\delta}\right]<C\cdot N^{-\delta}, \tag*{(6)}

where CC is a constant that may depend on G\mathbb{G}, but not on NN or c>0c>0. Note, however, that now, G′\mathbb{G}' may be different from G(K)\mathbb{G}^{(K)}, as it may not be properly KK-batched.

To get from G′\mathbb{G}' to G(K)\mathbb{G}^{(K)} we may need to permute tracks of the same type to obtain a properly KK-batched graph G′′\mathbb{G}'', then move certain batches of G′′\mathbb{G}'' to obtain G(K)\mathbb{G}^{(K)}. More precisely we claim the following: There exists a sequence T1(1),…,TL1(1),…,T1(J),…,TLJ(J)\mathrm{T}^{(1)}_1,\ldots,\mathrm{T}^{(1)}_{L_1},\ldots,\mathrm{T}^{(J)}_1,\ldots,\mathrm{T}^{(J)}_{L_J}, each of which is the composition of infinitely many, non-overlapping track-exchanges and a sequence S1(1),…,ST1(1),…,S1(J),…,STJ(J)\mathrm{S}^{(1)}_1,\ldots,\mathrm{S}^{(1)}_{T_1},\ldots,\mathrm{S}^{(J)}_1,\ldots,\mathrm{S}^{(J)}_{T_J}, each of which is the composition of infinitely many non-overlapping star-triangle transformations such that,

(STJ(J)∘⋯∘S1(1)∘TLJ(J)∘⋯∘T1(1))(G′)=G(K),\left(\mathrm{S}^{(J)}_{T_J}\circ\cdots\circ\mathrm{S}^{(1)}_1\circ\mathrm{T}^{(J)}_{L_J}\circ\cdots\circ\mathrm{T}^{(1)}_1\right)(\mathbb{G}')=\mathbb{G}^{(K)},

and the transformations Tj(i)\mathrm{T}^{(i)}_j and Sj(i)\mathrm{S}^{(i)}_j do not affect the tracks (tKℓ in(i))n∈Z\left(t^{(i)}_{K\ell\text{ in}}\right)_{n\in\mathbb{Z}}.

In particular, Lemma 4.5 applies to each transformation of the type TLi(i)∘⋯∘T1(i)\mathrm{T}^{(i)}_{L_i}\circ\cdots\circ\mathrm{T}^{(i)}_1 or STi(i)∘⋯∘S1(i)\mathrm{S}^{(i)}_{T_i}\circ\cdots\circ\mathrm{S}^{(i)}_1 and extends the coupling P\mathbb{P} to include a configuration ω′′∼ϕG(K)\omega''\sim\phi_{\mathbb{G}^{(K)}} with

P[dglobal CN(ω′∣ΛN,ω′′∣ΛN)>C′⋅N1−δ]<C′⋅N−δ,(7)\mathbb{P}\left[d_{\mathrm{global\ CN}}\left(\omega'|_{\Lambda_N},\omega''|_{\Lambda_N}\right)>C'\cdot N^{1-\delta}\right]<C'\cdot N^{-\delta}, \tag*{(7)}

where C′C' is a constant that may depend on G\mathbb{G}, but not on NN or c>0c>0. Combining (6) and (7) provides the desired conclusion after adjustment of the constants. □

Proof of Theorem 4.1, general case. We may restrict our attention to NN sufficiently large. Let c,C>0c,C>0 be given by Theorem 4.2.

Consider NN large and let K=N1−c/6K=N^{1-c/6}. Applying Theorem 4.2 to G(K)\mathbb{G}^{(K)} and using Proposition 4.4, we obtain a coupling P\mathbb{P} between ω0∼ϕG\omega_0\sim\phi_{\mathbb{G}}, ω1∼ϕG(K)\omega_1\sim\phi_{\mathbb{G}^{(K)}} and ω2∼ϕZ2\omega_2\sim\phi_{\mathbb{Z}^2} such that

P[dglobal CN(ω0∣ΛN,ω2∣ΛN)>CN1−c/2+N1−δ]≤P[dglobal CN(ω0∣ΛN,ω1∣ΛN)>N1−δ]+P[dglobal CN(ω1∣ΛN,ω2∣ΛN)≥CN1−c/2]≤N−δ+CN−c/3,\begin{aligned} \mathbb{P}\left[d_{\mathrm{global\ CN}}\left(\omega_0|_{\Lambda_N},\omega_2|_{\Lambda_N}\right)>CN^{1-c/2}+N^{1-\delta}\right] &\leq\mathbb{P}\left[d_{\mathrm{global\ CN}}\left(\omega_0|_{\Lambda_N},\omega_1|_{\Lambda_N}\right)>N^{1-\delta}\right]\\ &\quad+\mathbb{P}\left[d_{\mathrm{global\ CN}}\left(\omega_1|_{\Lambda_N},\omega_2|_{\Lambda_N}\right)\geq CN^{1-c/2}\right]\\ &\leq N^{-\delta}+CN^{-c/3}, \end{aligned}

where δ>0\delta>0 is given by Lemma 4.5. Adjusting the constants provides the desired conclusion. □

Proof of Lemma 4.5. Fix c>0c>0, NN, KK and jj as in the statement. We will choose δ>0\delta>0 below. We will only work with the tracks of type jj, which we denote (tn)n∈Z(t_n)_{n\in\mathbb{Z}}.

Let Strip⁡n\operatorname{Strip}_n denote the region enclosed by tnKt_{nK} and t(n+1)Kt_{(n+1)K} in ΛN\Lambda_N. Observe that ΛN\Lambda_N is then partitioned by Strip⁡−M,…,Strip⁡M\operatorname{Strip}_{-M},\ldots,\operatorname{Strip}_M for some M≤CN/K=CNcM\leq CN/K=CN^c, where CC is a constant potentially depending on G\mathbb{G}.

Using the fact that connections between the points of tnKt_{nK} are identical in ωj−1\omega_{j-1} and ωj\omega_j, we get a bijection between loops intersecting these tracks in the two configurations. In order for dglobal CN(ωj−1∣ΛN,ωj∣ΛN)>N1−δd_{\mathrm{global\ CN}}(\omega_{j-1}|_{\Lambda_N},\omega_j|_{\Lambda_N})>N^{1-\delta} to occur, one of two things must happen:

a) there exists some −M≤n≤M-M\leq n\leq M and a loop γ\gamma of ωj−1\omega_{j-1} intersecting tnKt_{nK} such that its coupled loop γ′\gamma' of ωj\omega_j satisfies d(γ,γ′)>N1−δd(\gamma,\gamma')>N^{1-\delta} or b) there exists some −M≤n≤M-M \le n \le M and a loop γ\gamma of ωj−1\omega_{j-1} or ωj\omega_j, completely contained in Stripn\mathrm{Strip}_n that has a diameter at least N1−δN^{1-\delta}.

Both points imply the existence of a primal path in either ωj−1\omega_{j-1} or ωj\omega_j, contained in some Stripn\mathrm{Strip}_n with −M≤n≤M-M \le n \le M and of diameter at least 12N1−δ\frac{1}{2}N^{1-\delta}. Write Long for this event. By the strong RSW property and Proposition 2.11 applied to ϕGj−1\phi_{G_{j-1}} and ϕGj\phi_{G_j} and using a union bound, we find that

P[Long]≤C0⋅M⋅exp⁡(−c0N1−δ/K),\mathbb{P}[\mathrm{Long}] \le C_0 \cdot M \cdot\exp(-c_0 N^{1-\delta}/K),

where C0C_0 and c0>0c_0 > 0 are constants potentially depending on GG. Since K=N1−cK = N^{1-c} and M≤CNcM \le CN^c, we conclude that, by taking δ>0\delta> 0 small enough, we have P[Long]≤N−δ\mathbb{P}[\mathrm{Long}] \le N^{-\delta}. This provides the desired conclusion. □\square

Statement on the use of AI. This work was entirely developed and written without the use of AI.

Acknowledgements. The second author acknowledges financial support from the Swiss National Science Foundation (SNSF).

References

  1. [1]Emile Averous, Hugo Duminil-Copin, Tiancheng He, Piet Lammers, and Ioan Manolescu. Conformal invariance of the planar four-state Potts model and its random-cluster representation. 2026. preprint arXiv:2610.08507.arxiv.org/abs/2610.08507
  2. [2]Emile Averous, Hugo Duminil-Copin, Tiancheng He, Piet Lammers, and Ioan Manolescu. Conformal invariance of the six-vertex model’s height function. 2026. preprint arXiv:2610.07454.arxiv.org/abs/2610.07454
  3. [3]Vincent Beffara and Hugo Duminil-Copin. The self-dual point of the two-dimensional random-cluster model is critical for q ≥ 1. Probab. Theory Relat. Fields, 153:511–542, 2012.
  4. [4]Federico Camia and Charles M. Newman. Two-dimensional critical percolation: the full scaling limit. Comm. Math. Phys., 268(1):1–38, 2006.
  5. [5]Dmitry Chelkak. Ising model and s-embeddings of planar graphs. Annales scientifiques de l’École Normale Supérieure, 57(5):1271–1346, 2024.arxiv.org/abs/2006.14559
  6. [6]Dmitry Chelkak, Hugo Duminil-Copin, Clément Hongler, Antti Kemppainen, and Stanislav Smirnov. Convergence of Ising interfaces to Schramm’s SLE curves. C. R. Math. Acad. Sci. Paris, 352(2):157–161, 2014.
  7. [7]Dmitry Chelkak and Stanislav Smirnov. Universality in the 2D Ising model and conformal invariance of fermionic observables. Invent. Math., 189(3):515–580, 2012.arxiv.org/abs/0910.2045
  8. [8]Hugo Duminil-Copin. Lectures on the Ising and Potts models on the hypercubic lattice. PIMS-CRM Summer School in Probability, 2019.
  9. [9]Hugo Duminil-Copin, Maxime Gagnebin, Matan Harel, Ioan Manolescu, and Vincent Tassion. Discontinuity of the phase transition for the planar random-cluster and Potts models with q > 4. Ann. Sci. Éc. Norm. Supér. (4), 54(6):1363–1413, 2021.
  10. [10]Hugo Duminil-Copin, Karol K. Kozlowski, Dmitry Krachun, Ioan Manolescu, and Mendes Oulamara. Rotational invariance in critical planar lattice models. ArXiV pre-print 2012.11672, 2021.
  11. [11]Hugo Duminil-Copin, Jhih-Huang Li, and Ioan Manolescu. Universality for the random-cluster model on isoradial graphs. Electron. J. Probab., 23:Paper No. 96, 70, 2018.
  12. [12]Hugo Duminil-Copin and Ioan Manolescu. Planar random-cluster model: scaling relations. Forum Math. Pi, 10:Paper No. e23, 83, 2022.
  13. [13]Hugo Duminil-Copin, Ioan Manolescu, and Vincent Tassion. Planar random-cluster model: fractal properties of the critical phase. Probab. Theory Relat. Fields, pages 401–449, 2021.
  14. [14]Hugo Duminil-Copin, Vladas Sidoravicius, and Vincent Tassion. Continuity of the phase transition for planar random-cluster and Potts models with 1 ≤ q ≤ 4. Communications in Mathematical Physics, 349(1):47–107, 2017.
  15. [15]Michael E. Fisher. Renormalization group theory: its basis and formulation in statistical physics. Rev. Modern Phys., 70(2):653–681, 1998.
  16. [16]Alessandro Giuliani, VierI Mastropietro, and Fabio Lucio Toninelli. Height fluctuations in interacting dimers. Annales de l’Institut Henri Poincaré, Probabilités et Statistiques, 53(1):98 – 168, 2017.DOI
  17. [17]Alessandro Giuliani, Vieri Mastropietro, and Fabio Lucio Toninelli. Non-integrable dimers: Universal fluctuations of tilted height profiles. Commun. Math. Phys., 377:1883–1959, 2020.DOI
  18. [18]Alexander Glazman and Piet Lammers. Delocalisation and continuity in 2d: Loop O(2), six-vertex, and random-cluster models. Communications in Mathematical Physics, 406, 2025.
  19. [19]Geoffrey R. Grimmett. The random-cluster model. volume 333 of Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences]., 2006.
  20. [20]Geoffrey R. Grimmett and Ioan Manolescu. Bond percolation on isoradial graphs: criticality and universality. Probab. Theory Related Fields, 159(1-2):273–327, 2014.DOI
  21. [21]Antti Kemppainen and Wendelin Werner. The nested simple conformal loop ensembles in the riemann sphere. Probability Theory and Related Fields, (165):835–866, 2016.arxiv.org/abs/1402.2433
  22. [22]Arthur E. Kennelly. The equivalence of triangles and three-pointed stars in conducting networks. Electrical World and Engineer, 34(12):413–414, 1899.
  23. [23]Richard Kenyon. Conformal invariance of domino tiling. Ann. Probab., 28(2):759–795, 2000.arxiv.org/abs/math-ph/9910002
  24. [24]Richard Kenyon. The Laplacian and Dirac operators on critical planar graphs. Invent. Math., 150(2):409–439, 2002.DOI
  25. [25]Lars Onsager. Crystal statistics. I. A two-dimensional model with an order-disorder transition. Phys. Rev. (2), 65:117–149, 1944.DOI
  26. [26]R. Peled and D. Romik. Bijective combinatorial proof of the commutation of transfer matrices in the dense O(1) loop model. Séminaire Lotharingien de Combinatoire, 73(B73b), 2015.arxiv.org/abs/1407.5357
  27. [27]Gourab Ray and Yinon Spinka. A short proof of the discontinuity of phase transition in the planar random-cluster model with q > 4. Communications in Mathematical Physics, 378, 2020.DOI
  28. [28]Oded Schramm. Scaling limits of loop-erased random walks and uniform spanning trees. Israel J. Math., 118:221–288, 2000.
  29. [29]Oded Schramm and Scott Sheffield. Harmonic explorer and its convergence to SLE₄. Ann. Probab., 33(6), 2005.
  30. [30]Oded Schramm and Scott Sheffield. Contour lines of the two-dimensional discrete Gaussian free field. Acta Math., 202(1):21–137, 2009.
  31. [31]Stanislav Smirnov. Critical percolation in the plane: conformal invariance, Cardy’s formula, scaling limits. C. R. Acad. Sci. Paris Sér. I Math., 333(3):239–244, 2001.DOI

Paper details

vlnc:W_recanmps
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 5

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents