🔬 condensed matter

Two dimensional inhomogeneous classical systems at criticality

This paper demonstrates that slowly varying inhomogeneous deformations of two-dimensional critical classical lattice models, such as the Ising and six-vertex models, are effectively described by conformal field theory in curved space, enabling the determination of underlying metrics and the calculation of phase boundaries like the arctic curve.

Original authors: Jean-Marie Stéphan

Published 2026-08-05

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Jean-Marie Stéphan

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

Imagine the universe as a giant, intricate dance floor where tiny particles are constantly waltzing, bumping into each other, and forming patterns. In the world of physics, scientists study these dances using "lattice models," which are like grid-based maps where every step a particle takes is recorded. Usually, these dance floors are perfectly uniform: the music is the same everywhere, the floor is flat, and the rules don't change from one corner to the next. This uniformity makes the math easier, but it's a bit boring. Real life, however, is messy. Gravity pulls harder in some places, temperatures vary, and materials aren't perfectly smooth.

When these particle dances reach a special "critical" point—a moment where the system is on the edge between order and chaos, like water just about to boil or a magnet losing its magnetism—they behave in a magical way. They become "scale-invariant," meaning a zoomed-in view looks just like a zoomed-out view. Physicists describe this behavior using something called Conformal Field Theory (CFT), which is like a universal language for these critical dances. Usually, this language assumes the dance floor is flat and the music is the same everywhere. But what happens if we tilt the floor or change the tempo slowly as you walk across the room? Does the dance still follow the same rules, or does the whole rhythm break? This is the big question: how do we describe the beautiful, chaotic patterns of critical systems when the stage itself is warped and uneven?

In this paper, the author Jean-Marie Stéphan takes a deep dive into exactly that scenario. He asks what happens when we take two famous, well-understood particle dances—the Ising model (which explains how magnets work) and the six-vertex model (which describes how paths or water molecules arrange themselves)—and slowly warp the rules of the game across the grid. Instead of changing the rules abruptly, he changes them gently, like a slow fade in volume or a gradual slope in the floor.

The main finding is surprisingly elegant: even when the rules change from place to place, the system doesn't fall apart. Instead, the entire dance floor effectively becomes "curved." The author shows that these inhomogeneous systems can be described by the same mathematical language (CFT) used for flat floors, but with a twist: the geometry of the space itself bends. It's as if the particles are dancing on a trampoline that is being gently stretched and squeezed in different spots. The paper identifies exactly what this "curved space" looks like for these specific models.

To prove this, the author uses two main examples. First, he looks at the Ising model. By making the magnetic connections between spins (the dancers) slightly stronger or weaker depending on where they are, he shows that the "speed" at which information travels through the system changes locally. This creates a curved space-time metric, a mathematical description of the warped floor. He confirms this by checking the energy gaps and how spins talk to each other, finding that the math matches perfectly with the idea of a curved stage.

Second, he tackles the more complex six-vertex model with "domain wall" boundaries. Imagine a square room where all the dancers enter from the left wall and must exit through the top wall. In a uniform room, they form a beautiful, predictable shape in the middle, surrounded by frozen, rigid corners. This boundary is called the "arctic curve" (because the frozen parts look like ice). When the author introduces the slow, inhomogeneous changes to the rules, the shape of this ice changes. He uses a clever mix of hydrodynamics (treating the particles like a flowing fluid) and a method called the "tangent method" to predict exactly how the arctic curve bends and warps.

The paper doesn't just guess; it checks these ideas with high-precision computer simulations. For the free (non-interacting) version of the six-vertex model, the simulations match the curved-space predictions almost perfectly. For the interacting version (where particles bump into each other more), the author derives a new formula for the arctic curve and shows that the simulations agree with it, even though the math is much harder. The results suggest that the "curved space" description is a robust way to understand these systems, even when the rules are slowly changing. The paper essentially hands us a new pair of glasses: one that lets us see the hidden curvature in the dance of particles, turning a messy, uneven room into a beautifully warped stage where the laws of physics still hold, just in a different shape.

Technical Summary: Two Dimensional Inhomogeneous Classical Systems at Criticality

Problem Statement
This paper investigates the behavior of critical two-dimensional classical lattice models under slow, spatially inhomogeneous deformations. While standard statistical mechanics models are typically homogeneous, leading to Conformal Field Theories (CFTs) with translation invariance, the author explores systems where Boltzmann weights vary slowly across the lattice without breaking criticality. The central question is how such deformations modify the underlying field theory description, specifically whether they induce a curved space-time metric and how integrability is preserved or utilized in this context. The study focuses on two primary examples: the Ising model and the six-vertex model with domain wall boundary conditions (DWBC).

Methodology
The author employs a combination of exact lattice techniques, field theory mapping, and hydrodynamic approaches:

Ising Model Analysis:
- The author utilizes the transfer matrix formalism and the Jordan-Wigner transformation to map the inhomogeneous Ising model to a free fermion system with space-dependent parameters.
- They introduce a deformation where coupling constants $K_j$ and $K'_j$ vary slowly as functions of position, maintaining the local critical condition $\sinh^2 K_j^* \sinh^2 K'_j = 1$ .
- By analyzing the gap of the transfer matrix and spin-spin correlation functions, they derive the effective space-time metric. They employ Weyl rescaling and conformal transformation laws for primary fields to predict correlation functions in the curved background.
Six-Vertex Model Analysis (Free Fermion Case, $\Delta=0$ ):
- For the free fermion limit, the author adapts an imaginary-time hydrodynamic theory. They treat the average vertex occupancies as a density profile $\rho(x,y)$ governed by a complex dispersion relation.
- They solve the hydrodynamic equation $i\partial_y z + \partial_x \epsilon(z, \lambda(x), \mu(y)) = 0$ using the method of complex characteristics to determine the density profile and the "arctic curve" (the boundary between fluctuating and frozen phases).
- They interpret the fluctuations around the mean height profile as a free scalar field theory in curved space, identifying the metric tensor via the Hessian of the free energy Lagrangian.
Six-Vertex Model Analysis (Interacting Case, $\Delta \neq 0$ ):
- For the interacting case, exact limit shapes are not generally known. The author utilizes the "tangent method" to determine the arctic curve.
- They derive asymptotic formulas for the partition function and boundary probabilities (contact points) using determinant representations (Izergin-Korepin) and saddle-point approximations.
- They construct a family of curved rays (geodesics) in the $(x,y)$ plane. The arctic curve is identified as the caustic (envelope) of these rays, determined by the conditions $R(\omega) = 0$ and $dR(\omega)/d\omega = 0$ .

Key Contributions and Results

Curved Space CFT Identification: The primary finding is that slowly varying inhomogeneous deformations of critical lattice models result in a field theory description in curved space.
- For the Ising model, the deformation leads to a space-dependent Fermi speed $v_F(x) = \sinh[2K_0 f(x/N)]$ . The underlying Euclidean metric is identified as $ds^2 = dx^2 + v_F(x)^2 dy^2$ . The author verifies this by showing that the transfer matrix gap scales with an effective system size $\tilde{N} = \int dx/v_F(x)$ and that spin-spin correlations match the predictions of CFT in this curved metric, including necessary prefactor adjustments for the lattice-to-field mapping.
- For the six-vertex model, the metric is similarly curved but arises from inhomogeneous spectral parameters (rapidities) $\lambda(x)$ and $\mu(y)$ . In the free case, the hydrodynamic approach yields an exact density profile and confirms the CFT description with central charge $c=1$ .
Arctic Curves in Inhomogeneous Systems:
- In the free six-vertex model, the author derives an exact formula for the density profile and the arctic curve for arbitrary slow deformations of the spectral parameters.
- In the interacting six-vertex model, they successfully generalize the tangent method to inhomogeneous weights. They derive a central equation (Eq. 80) defining a family of curved geodesics. The arctic curve is shown to be the caustic of these rays. This result generalizes previous homogeneous findings to the inhomogeneous regime.
Numerical Verification:
- The theoretical predictions for the Ising model gap and correlations are verified numerically with high precision, showing excellent agreement with the curved space CFT formulas.
- For the six-vertex model, Monte Carlo simulations of the arctic curve for both free and interacting cases confirm the theoretical predictions. The author notes that finite-size effects are significant but can be effectively removed using Tracy-Widom scaling near the edge.

Significance and Claims
The paper claims to provide a systematic framework for understanding critical lattice models with slow spatial inhomogeneities. Its significance lies in:

Unifying Lattice and Field Theory: It demonstrates that "clean" inhomogeneities in lattice models naturally map to CFTs in curved space, providing a concrete lattice realization of such field theories.
Preservation of Integrability: It shows that specific inhomogeneous deformations (varying spectral parameters) preserve the integrability of the six-vertex model, allowing for exact solutions even in the presence of spatial variation.
Generalization of Limit Shapes: It extends the understanding of limit shapes and arctic curves from homogeneous systems to inhomogeneous ones, providing exact analytical tools (hydrodynamics for free cases, tangent method for interacting cases) to determine phase boundaries.

The author modestly notes that while they have not derived the curved space metric directly from the lattice for the interacting six-vertex model (relying instead on the tangent method and asymptotic analysis), the consistency between the hydrodynamic approach (free case) and the tangent method (interacting case) strongly supports the validity of the curved space interpretation. They suggest these results may extend to other models like the three-state Potts model and higher dimensions, though lattice computations for those remain challenging.

More like this