Bosonization of primary fields for the critical Ising model on multiply connected planar domains

This paper establishes bosonization identities for the scaling limits of critical Ising model correlations on finitely-connected planar domains by expressing them in terms of compactified Gaussian free field correlations, utilizing a limiting version of the Hejhal-Fay identity and operator product expansions to derive explicit formulas involving the domain's period matrix, Green's function, and Abelian differentials.

The Gist

Imagine you are trying to understand a very complex, chaotic dance floor. This dance floor is the Critical Ising Model, a famous mathematical model used to describe how tiny magnets (spins) on a grid decide to point up or down. When the temperature is just right (the "critical" point), these spins don't just act randomly; they form intricate, large-scale patterns that look like a fluid or a wave.

For decades, physicists and mathematicians have known that these patterns exist and have even calculated them for simple, empty rooms (like a flat sheet of paper). But what happens if the room has obstacles? What if the room is shaped like a donut, or has islands in the middle, or has walls with different rules? This is the problem of multiply connected domains (rooms with holes or multiple boundaries).

This paper, by Bayraktaroglu, Izyurov, Virtanen, and Webb, solves a major puzzle: How do we translate the complex dance of these magnetic spins into a simpler, more familiar language?

Here is the breakdown of their discovery using everyday analogies:

1. The Problem: Two Different Languages

The authors are trying to connect two different ways of describing the same physical reality:

• Language A (The Spin Model): This is the "hard" way. It involves counting billions of tiny magnets and figuring out how they influence each other across a complex shape. It's like trying to predict the weather by tracking every single air molecule. It's accurate but incredibly messy.
• Language B (The Bosonic Field): This is the "easy" way. It describes the system using a Gaussian Free Field (GFF). Think of this as a smooth, vibrating rubber sheet or a calm lake with ripples. Mathematically, this is much easier to handle because it's a "free" field (no complicated interactions).

The goal of the paper is Bosonization: proving that the messy, complex dance of the magnets (Language A) is actually just a fancy, disguised version of the smooth ripples on the rubber sheet (Language B).

2. The Big Reveal: The "Square" Connection

The main result of the paper is a formula that says:

The square of the magnetic dance = The ripples on the rubber sheet.

Why the square? In physics, sometimes you need to look at two copies of the system side-by-side to see the pattern clearly. The authors prove that if you take the "correlation" (how likely two points are to act together) of the magnets, square it, and look at it through the right lens, it matches perfectly with the correlation of the smooth ripples.

This is huge because the "ripples" (Language B) can be calculated using standard tools like:

• Green's Functions: Think of these as "distance maps" that tell you how much one point on the rubber sheet affects another.
• Harmonic Measures: These are like "probability maps" telling you how likely a random walker is to hit a specific part of the wall.
• Theta Functions: These are complex mathematical patterns that act like the "DNA" of the shape of the room.

3. The Method: Pinching the Holes

How did they prove this? They used a clever mathematical trick involving Riemann Surfaces (which are like multi-layered, twisted versions of our 2D world).

Imagine the room with holes (like a donut). The authors imagined a process called "pinching the handles."

1. They took the complex shape and attached tiny, thin "handles" to it, turning it into a giant, closed surface (like a sphere with many handles).
2. On this giant surface, there is a famous, old mathematical identity (the Hejhal–Fay identity) that relates the "spin" of the surface to the "ripples."
3. Then, they slowly shrank those tiny handles until they vanished (pinched them off).
4. As the handles shrank, the complex math on the giant surface "collapsed" down into the specific math for the room with holes.

It's like taking a complex origami crane, unfolding it into a flat sheet of paper to see the crease pattern clearly, and then folding it back up. By watching how the math behaves as the handles disappear, they proved the connection holds true for the original complex shape.

4. Why Does This Matter?

Before this paper, if you wanted to know how magnets behave in a weirdly shaped room (like a room with three islands in the middle), you had to solve a nightmare of equations that often had no clear answer.

Now, thanks to this paper, you can:

• Plug in the shape: Take the geometry of your room (its period matrix, its boundaries).
• Use the formula: Apply the "Bosonization" recipe provided in the paper.
• Get the answer: Instantly get an explicit formula for the magnetic correlations using the "ripples" (Green's functions and theta functions).

Summary Analogy

Imagine you are trying to understand the sound of a complex jazz band playing in a cathedral with many pillars (the Ising model). It's hard to hear the individual notes.

This paper proves that the sound of the jazz band is mathematically identical to the sound of a single, pure flute playing in a simpler, echo-free room (the Bosonic field), provided you listen to the square of the volume.

The authors didn't just say "it sounds similar." They built a bridge (using the "pinching" trick) that allows you to take the sheet music of the flute (which is easy to read) and instantly know the sheet music of the jazz band, no matter how many pillars are in the room.

In short: They found a universal translator that turns the chaotic language of complex magnets into the elegant, solvable language of smooth waves, for any shape of room you can imagine.

Technical Summary

1. Problem Statement and Context

The paper addresses the rigorous derivation of bosonization identities for the scaling limits of correlation functions in the critical two-dimensional Ising model defined on finitely-connected (multiply connected) planar domains.

• Background: The critical Ising model is a cornerstone of statistical mechanics. While its exact solvability is well-known on the full plane ($\mathbb{C}$) and the torus, extending these results to general multiply connected domains (e.g., annuli, domains with holes) with mixed boundary conditions (wired/free) has been challenging.
• Previous Work: In prior work (specifically [10, 11]), the authors established the existence and conformal covariance of scaling limits for Ising correlations (spin $\sigma$, disorder $\mu$, energy $\varepsilon$, fermions $\psi, \psi^*$) on such domains. However, these limits were expressed via solutions to Riemann boundary value problems, which, while explicit, are not as "closed-form" as the bosonization formulas available for simply connected domains.
• The Gap: There was no rigorous proof connecting these scaling limits to the compactified Gaussian Free Field (GFF) (bosonic theory) in the general multiply connected case. The goal is to prove that the square of an Ising correlation equals a specific correlation of a free bosonic field, expressed explicitly via the domain's period matrix, Green's function, and harmonic measures.

2. Methodology

The proof is a sophisticated synthesis of discrete complex analysis, conformal field theory (CFT) concepts (specifically Operator Product Expansions), and classical Riemann surface theory.

A. The Bosonic Framework

The authors define a "compactified free field" $\Phi = \phi + \xi$ on the domain $\Omega$:

• $\phi$: A standard Gaussian Free Field with Dirichlet boundary conditions.
• $\xi$: An "instanton" component, a random harmonic function taking values in a discrete lattice ($\sqrt{2\pi}\mathbb{Z}$ or shifted versions) on the boundary components. This accounts for the topological sectors (winding numbers) inherent in multiply connected domains.
• Correlations: The bosonic correlations are defined via normal ordering (Wick ordering) of exponentials, derivatives, and trigonometric functions of $\Phi$.

B. The Hejhal–Fay Identity and Schottky Doubles

The core of the proof relies on a limiting argument involving the Schottky double of the domain $\Omega$.

1. Construction of $\hat{\Omega}_\epsilon$: The authors construct a compact Riemann surface $\hat{\Omega}_\epsilon$ by taking the Schottky double of $\Omega$ (two copies of $\Omega$ glued along the boundary) and attaching $n+k$ small handles (pinching limits) near the insertion points of the fields ($v_i$) and the boundary condition change points ($b_j$).
2. Hejhal–Fay Identity: They utilize a classical identity by Hejhal and Fay, which relates the square of a Szegő kernel (a spinor kernel on a Riemann surface) to Abelian differentials and theta functions:
   $$\Lambda(P, Q)^2 = \beta(P, Q) + \sum \partial \partial \log \theta \cdot u_i(P) u_j(Q)$$
3. The Pinching Limit: They analyze the behavior of this identity as the handles degenerate ($\epsilon \to 0$).
   • Left-Hand Side (LHS): The Szegő kernel on the degenerating surface converges to the ratio of Ising correlations involving fermions ($\psi$).
   • Right-Hand Side (RHS): The Abelian differentials and period matrix converge to the Green's function, harmonic measures, and the period matrix of the original domain $\Omega$. The theta function terms converge to the partition function of the instanton component $\xi$.

C. Operator Product Expansions (OPE)

Once the identity is established for specific fields (spin $\sigma$ and disorder $\mu$), the authors use fusion rules (OPEs) to extend the result to all primary fields ($\varepsilon, \psi, \psi^*$).

• They verify that the bosonic OPEs (asymptotic expansions of products of fields like $\partial \Phi(z) :e^{\gamma \Phi(w)}:$) match the known Ising OPEs derived in previous work [11].
• This allows them to bootstrap the result from the base case (spins/disorders) to the full set of primary fields.

3. Key Contributions and Results

Main Theorem (Theorem 1.1)

The paper proves the following bosonization identity for the scaling limits of critical Ising correlations on a finitely-connected domain $\Omega$:
$$\langle O_{z_1} \dots O_{z_N} \rangle_\Omega^2 = \langle \hat{O}_{z_1} \dots \hat{O}_{z_N} \rangle_\Omega$$
where:

• $O_{z_i}$ are Ising primary fields ($\sigma, \mu, \varepsilon, \psi, \psi^*$).
• $\hat{O}_{z_i}$ are the corresponding bosonic fields constructed from the compactified GFF $\Phi$.
• The identity holds provided the parity of the number of specific fields (spin/fermion/disorder) is even (otherwise the correlation vanishes).

Explicit Mapping:

• $\sigma_z \leftrightarrow \sqrt{2} : \cos(\frac{\sqrt{2}}{2}\Phi(z)) :$
• $\mu_z \leftrightarrow \sqrt{2} : \sin(\frac{\sqrt{2}}{2}\Phi(z)) :$
• $\varepsilon_z \leftrightarrow -\frac{1}{2} : |\nabla \Phi(z)|^2 :$
• $\psi_z \leftrightarrow 2\sqrt{2}i \partial \Phi(z)$
• $\psi^*_z \leftrightarrow -2\sqrt{2}i \partial \Phi(z)$

Explicit Formulae

The RHS is given explicitly in terms of:

• The Green's function $G_\Omega$.
• The period matrix of the domain (or its Schottky double).
• Harmonic measures of boundary components and arcs.
• Theta functions evaluated at the harmonic measures (Abel-Jacobi map).

Technical Lemmas

• Lemma 4.15: Establishes the convergence of the RHS of the Hejhal–Fay identity to the bosonic correlation involving the instanton component.
• Lemma 4.16: Proves that the Szegő kernel on the degenerating surface converges to the ratio of Ising fermion correlations.
• Theorem 1.3: A specific case of the main theorem for spin and disorder fields, which serves as the base for the inductive proof using OPEs.

4. Significance and Impact

1. Rigorous Bosonization: This work provides the first rigorous proof of bosonization for the critical Ising model on multiply connected domains. Previously, such results were largely heuristic or restricted to simply connected geometries.
2. Explicit Computability: By expressing Ising correlations in terms of the period matrix and theta functions, the authors provide a concrete, computable framework. This moves beyond the "implicit" solutions of boundary value problems, allowing for direct numerical evaluation and asymptotic analysis.
3. Unification of Approaches: The paper bridges the gap between:
   • Probabilistic/Discrete approaches: (Convergence of lattice models to GFF).
   • Algebraic/Geometric approaches: (Riemann surfaces, theta functions, and CFT).
4. Foundation for Future Work: The methodology (pinching handles on Schottky doubles) is robust and could be adapted to other minimal models or off-critical regimes. It also clarifies the role of the "instanton" (topological) sector in the bosonization of lattice models with non-trivial topology.
5. Resolution of Parity Issues: The paper carefully handles the parity constraints (even/odd numbers of spin insertions) and boundary conditions, showing how they map to the specific sectors of the compactified boson.

In summary, this paper is a major advancement in mathematical physics, providing a complete and explicit dictionary between the critical Ising model and the compactified Gaussian Free Field on general planar domains, grounded in rigorous analysis of Riemann surfaces.