🔢 mathematics

Analyticity of the correlation length in subcritical FK-percolation and Potts models

This paper proves that the correlation length of FK-percolation and Potts models on $\mathbb{Z}^d$ is analytic throughout the subcritical regime by utilizing Kato's analytic perturbation theory and the FK-Ruelle operator, thereby demonstrating that rotational invariance of the correlation length occurs only at countably many temperature values.

Original authors: Lucas D'Alimonte, Loïc Gassmann

Published 2026-10-19

📖 7 min read🧠 Deep dive

CC BY 4.0

Original authors: Lucas D'Alimonte, Loïc Gassmann

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the microscopic world of statistical physics, scientists study how vast collections of tiny particles or spins interact to create the large-scale properties we see in everyday materials, such as whether a magnet holds its strength or a liquid freezes into ice. A central concept in this field is the "correlation length," which measures how far the influence of one particle reaches before fading away into randomness. Imagine dropping a stone into a calm pond; the ripples spread out, but eventually, the water becomes still again. The distance the ripples travel before vanishing is analogous to the correlation length. In the "subcritical" phase of these systems—a state where the material is stable and not on the verge of changing its fundamental nature—these influences fade away quickly, following a predictable exponential decay. For decades, physicists have wondered about the mathematical smoothness of this decay rate as the temperature changes. They suspected that if you were to tweak the temperature slightly, the rate at which these influences fade would change in a perfectly smooth, predictable way, without any sudden jumps or jagged edges. This question of "analyticity" is crucial because it tells us whether the system behaves with a consistent, underlying order or if it hides sudden, unexpected shifts in behavior even when it appears stable.

A team of researchers has now provided a definitive answer to this long-standing question for a broad class of models known as FK-percolation and Potts models, which describe how connections form and break in a grid-like structure. These models are not just abstract mathematical toys; they are powerful tools used to understand real-world phenomena, including the behavior of magnets and the flow of fluids through porous rock. The researchers proved that in the stable, subcritical phase, the correlation length is indeed perfectly smooth. It changes in a way that is mathematically "analytic," meaning it can be described by a single, continuous formula that works for any small change in temperature or direction. This result confirms that the system does not suffer from hidden, sudden shifts in its decay rate as it moves through this stable phase. It is a rigorous proof that the behavior of these complex networks is as orderly and predictable as physicists had hoped, ruling out the possibility of strange, irregular changes that might have occurred at specific, isolated temperatures.

To reach this conclusion, the team had to navigate a landscape of complex mathematics that connects probability theory with the study of operators, which are essentially machines that transform one set of numbers into another. They focused on a specific quantity called the "mass," which is the inverse of the correlation length; a larger mass means the influence fades faster. The challenge was that the mass is defined by a limit involving a logarithm, a mathematical operation that is notoriously difficult to handle when trying to prove smoothness. Directly analyzing the logarithm of the connection probabilities often leads to dead ends because the numbers inside the log can vanish or behave erratically when the temperature is treated as a complex variable. Instead of fighting this difficulty head-on, the researchers took a different path. They turned to a method involving "irreducible clusters," which are the smallest, indivisible building blocks of the connections in the model. By breaking down the complex network of connections into these fundamental pieces, they could describe the system using a special operator, often called a Ruelle operator, which acts like a sophisticated counting machine for these clusters.

The breakthrough came when the researchers realized they could treat the temperature parameter not just as a real number, but as a complex number, allowing them to use powerful tools from the theory of how operators change when their inputs are slightly perturbed. They showed that the operator describing the system behaves in a very stable and smooth way even when the temperature is nudged into this complex realm. Because the most important feature of this operator—the one that dictates the decay rate of the system—stays isolated and distinct from the rest of the mathematical noise, the researchers could prove that the decay rate itself must change smoothly. This is similar to how a single, distinct note played on a violin remains clear and unchanged even if the room temperature shifts slightly, whereas a chaotic noise might distort. By proving that the operator's key feature moves smoothly, they proved that the correlation length moves smoothly. This approach allowed them to bypass the difficulties of the logarithm and establish the smoothness of the correlation length across the entire stable phase.

The implications of this finding extend beyond just confirming a mathematical property. The researchers used their result to address a question about the shape of the "Wulff shape," a geometric object that describes the most efficient way for a cluster of connected particles to form in a specific direction. In a perfectly round, rotationally symmetric world, this shape would be a sphere. However, because these models live on a grid, they are expected to feel the underlying structure of that grid, making the shape slightly square or diamond-like rather than perfectly round. The team proved that the correlation length cannot be perfectly rotationally symmetric except at a countable number of isolated temperature points. In other words, the system almost always "feels" the grid it sits on, and its properties change in a way that respects the grid's geometry. They showed that if the shape were to become a perfect sphere, it would have to happen only at very specific, isolated temperatures, and likely never at all. This adds a layer of rigidity to our understanding of these systems, showing that their behavior is tightly constrained by the mathematics of their structure.

The work also clarifies the behavior of these models in two dimensions, where a special relationship called duality exists between the system and its mirror image. The researchers showed that this duality preserves the smoothness of the correlation length, meaning that the properties of the system above a critical temperature are just as smooth as those below it. This resolves a conjecture that had been open for some time, particularly for the Ising model, which is a famous model of magnetism. By proving that the correlation length is analytic, the researchers have closed the door on the possibility of "saturation" phenomena, where the decay rate might suddenly stop changing or behave differently in a specific temperature range. Their results indicate that for these nearest-neighbor models, the decay rate changes continuously and predictably from the coldest temperatures up to the point where the phase transition occurs. This provides a solid foundation for future studies, suggesting that the tools used here can be applied to more complex systems where the interactions between particles are not limited to immediate neighbors.

Ultimately, this paper represents a significant step in understanding the mathematical landscape of statistical mechanics. It demonstrates that even in complex, interacting systems, there is a deep underlying order that ensures smoothness and predictability in the stable phases. The researchers did not just simulate these behaviors or suggest they might be true; they provided a rigorous proof that leaves no room for doubt. By combining advanced techniques from operator theory with a careful analysis of cluster expansions, they have mapped out the behavior of the correlation length with a level of precision that was previously out of reach. This work reassures physicists that the models they use to describe the physical world are mathematically robust, and it opens the door to exploring similar questions in even more complicated settings. The proof stands as a testament to the power of modern mathematical methods to illuminate the hidden structures of the physical universe.

Technical Summary: Analyticity of the Correlation Length in Subcritical FK-Percolation and Potts Models

Problem Statement
The paper addresses the regularity properties of thermodynamic quantities in statistical mechanics lattice models, specifically focusing on the correlation length (or its inverse, the mass) of FK-percolation and Potts models in their subcritical regimes. While the exponential decay of the two-point function (sharpness of the phase transition) was established for these models in [DCRT19], the analyticity of the mass as a function of the percolation parameter $p$ (or inverse temperature $\beta$ ) and the spatial direction remained an open problem, particularly in dimensions $d \geq 3$ . Previous results were limited to perturbative regimes (near $p=0$ or $p=1$ ) or specific integrable cases (e.g., 2D Ising). The authors aim to prove that the correlation length is jointly analytic in both the parameter and the direction throughout the entire subcritical phase $(0, p_c)$ .

Methodology
The authors adopt a strategy distinct from a "direct approach" (extending the two-point function to the complex plane and taking a logarithm), which is hindered by the potential vanishing of the complexified probabilities. Instead, they utilize a framework inspired by Ornstein–Zernike theory and Campanino, Ioffe, and Velenik (CIV) [CIV03, CIV08].

Polar Duality and the Mass Gap: The authors work with the quantity $y(p, t)$ , defined as the inverse of the polar dual of the mass. Geometrically, $y(p, t)$ represents the rate of decay of connection probabilities to a hyperplane in direction $t$ .
FK-Ruelle Operator: They identify $y(p, t)$ as the solution to an equation involving the spectral radius of a specific operator, the FK-Ruelle operator ( $L_{\psi}$ ), acting on a Banach space of functions over chains of irreducible clusters.
Kato's Perturbation Theory: The core of the proof relies on Kato's theory of analytic perturbations of linear operators [Kat95]. To apply this, the authors must demonstrate that the FK-Ruelle operator, when its parameters ( $p$ and the direction $s$ ) are perturbed into the complex plane, forms a holomorphic family of operators.
Cluster Expansion: To establish the holomorphicity of the operator, the authors perform a rigorous cluster expansion (following [Ott20a, Ott20b]) on the FK-Ruelle potential. A key innovation is the introduction of Unique Decoupling (UD) events. They prove that the probabilities of these specific events admit analytic extensions to a complex neighborhood of the real parameter $p$ , with uniform bounds on their logarithmic derivatives. This allows them to control the complexified potential and verify the necessary Hölder and mass-gap conditions (Theorem 3.4).
Implicit Function Theorem: Once the spectral radius of the complexified operator is shown to be analytic (via Kato's theorem), the analyticity of $y(p, t)$ follows from the holomorphic implicit function theorem applied to the identity relating the spectral radius to the mass.

Key Contributions

Theorem 1.2: Proves that for any dimension $d \geq 2$ and cluster weight $q \geq 1$ , the mass function $(p, x) \mapsto m(p, x)$ is jointly analytic on $(0, p_c) \times \mathbb{R}^d \setminus \{0\}$ .
Theorem 1.9: Establishes the joint analyticity of the dual quantity $y(p, t)$ on the same domain.
Complex Regularity of the Ruelle Operator (Theorem 3.4): The paper provides the first proof that the FK-Ruelle operator admits a holomorphic perturbation in the complex parameter $p$ while maintaining the crucial mass-gap and Hölder continuity properties required for the Ruelle-Perron-Frobenius theorem. This is achieved via a novel cluster expansion on general subgraphs of $\mathbb{Z}^d$ .
Rigidity Principle: The authors demonstrate that the analyticity of the correlation length implies a "rigidity principle": any property of the model that can be reformulated as an identity for the correlation length must either hold everywhere in the subcritical regime or only on a countable set of parameters.

Results

Corollary 1.4: As a consequence of the Edwards–Sokal coupling, the correlation length of the Potts model (and thus the Ising model for $q=2$ ) is jointly analytic in $\beta$ and direction for all $d \geq 2$ and $q \geq 2$ . This resolves a conjecture regarding the analyticity of the Ising mass in $d \geq 3$ mentioned in [OV23].
Corollary 1.5 & 1.6: In dimension $d=2$ , the truncated mass and the surface tension are shown to be jointly analytic and strictly positive on their respective domains, utilizing planar duality.
Theorem 1.7 (Non-Rotational Invariance): The authors prove that the set of parameters $p \in (0, p_c)$ for which the Wulff shape is rotationally invariant (a sphere) is finite within any compact sub-interval of $(0, p_c)$ . This provides a rigorous, albeit slightly weaker, confirmation that off-critical lattice models generally lack rotational symmetry, except possibly at isolated points.

Significance and Claims
The paper claims to solve the long-standing problem of the analyticity of the correlation length for FK-percolation and Potts models in the full subcritical regime, extending beyond perturbative expansions.

Absence of Saturation: The result proves the absence of the "saturation phenomenon" (where analyticity fails at an intermediate parameter $\beta_{sat}$ ) for nearest-neighbor FK-percolation. This supports the conjecture that such models are analytic in the entire non-saturated regime.
Methodological Advancement: The work bridges the gap between the "modern renewal picture" of Ornstein–Zernike theory and the rigorous cluster expansion techniques required to handle complex parameters. The authors note that their method relies on the specific form of the FK-Ruelle potential and that the "modern" renewal weights are too intricate to verify the necessary regularity for Kato's theory.
Rigidity: By establishing analyticity, the paper provides a powerful tool to rule out "accidental" symmetries or identities in the subcritical phase, limiting them to isolated parameter values.

The authors remain modest regarding the non-rotational invariance result, acknowledging that while they prove the set of exceptional parameters is finite, they cannot currently prove it is empty (though it is known to be empty for the 2D Ising model). They also note that their method relies on ingredients (exponential mixing, mass gap) that are expected to hold in broader contexts, suggesting their approach could be a pathway to resolving the general conjecture on analyticity in the non-saturated regime for other models.

Technical Summary: Analyticity of the Correlation Length in Subcritical FK-Percolation and Potts Models

More like this