Derivation of height field theory for the two-dimensional classical dimer model from a Grassmann-integral representation

This paper provides a constructive derivation of the continuum height field theory for the two-dimensional classical dimer model on square and honeycomb lattices by starting from an exact Grassmann integral representation, taking the continuum limit to obtain massless Dirac fermions, and applying bosonization to map the system to a height model that fully describes its long-distance correlations and topological properties.

The Gist

Imagine a giant, flat floor covered in a grid of tiles. Now, imagine you have a pile of dominoes (which the paper calls "dimers"). Your goal is to cover the entire floor with these dominoes so that:

1. Every single tile is covered by exactly one domino.
2. No dominoes overlap.
3. There are no empty tiles left over.

This is the Classical Dimer Model. It sounds simple, but when you have a huge floor (a "lattice") and you try to count all the possible ways to arrange these dominoes, the math gets incredibly complicated. The paper by Stephen Powell is a guidebook on how to translate this messy, tile-by-tile puzzle into a smooth, flowing language that physicists can easily understand.

Here is the story of how he does it, using some creative analogies.

1. The Problem: Too Many Dominoes

If you look at a small floor, you can count the arrangements. But on a massive floor, the number of ways to arrange the dominoes is astronomical. Physicists want to know: If I look at two dominoes far apart, are they related? Do they "talk" to each other?

The paper says that even though the dominoes are discrete (separate objects), when you zoom out far enough, they behave like a smooth, continuous fluid. This fluid has a special property: it has zero divergence. Think of it like water flowing in a pipe where no water is created or destroyed; whatever flows in must flow out.

2. The First Translation: The "Height Map"

To understand this fluid, the author introduces a clever trick: the Height Field.

Imagine the floor isn't flat. Instead, imagine that every time you place a domino, it slightly changes the "height" of the ground around it.

• If a domino is placed in one direction, the ground rises a tiny bit.
• If it's placed the other way, the ground falls.

Because of the rule that every tile must be covered, the ground can't just slope up forever; it has to form a smooth, wavy landscape. The author shows that the complicated rules of the dominoes can be replaced by a simple rule: The landscape is a smooth, wavy surface.

This is the "Height Theory." Instead of counting dominoes, we are now studying the ripples on a calm lake.

3. The Secret Ingredient: The "Ghost" Dominoes

How did the author prove that the dominoes turn into a smooth lake? He used a mathematical tool called Grassmann Integrals.

Think of Grassmann variables as "ghost" particles. They aren't real dominoes you can touch; they are mathematical ghosts that help you do the counting.

• The author starts by writing the rules of the domino game using these ghosts.
• He then shows that these ghosts behave exactly like Dirac Fermions.

The Fermion Analogy:
Imagine the ghosts are like tiny, invisible particles zooming around the floor. In the world of physics, "Dirac Fermions" are particles that move at a constant speed and have a very specific, simple energy pattern. The author proves that if you look at the "ghosts" of the domino model, they are exactly these simple, fast-moving particles.

4. The Magic Trick: Bosonization

Now we have a problem: We have a smooth lake (the Height Field) and we have invisible particles (the Fermions). How do we connect them?

The author uses a technique called Bosonization.

• The Analogy: Imagine you have a choir of singers (the fermions). Individually, they are distinct people. But if you listen to the choir from far away, you don't hear individual voices; you hear a smooth, continuous sound wave (the boson/height field).
• Bosonization is the mathematical recipe that says: "A crowd of these specific particles is exactly the same thing as a smooth wave."

By applying this recipe, the author transforms the "ghost particle" math back into the "smooth lake" math.

5. The Result: A Perfect Match

The paper's main achievement is showing that this translation works perfectly for two different types of floors:

1. The Square Grid: Like a standard checkerboard.
2. The Honeycomb Grid: Like a beehive pattern.

Even though the dominoes sit on different shapes, the "smooth lake" they create is identical. The author also adds "sensors" (called source terms) to the math. These sensors measure:

• The Flux: How much "wind" is blowing through the system (related to the overall tilt of the lake).
• The Magnetization: A specific pattern of order that might appear in the dominoes.

He proves that these are the only two things you need to measure to understand how the dominoes behave over long distances.

6. Why This Matters (According to the Paper)

Before this paper, physicists often had to guess the "smooth lake" theory and then check if it matched the dominoes. This paper does the opposite: it starts with the exact domino rules, uses the "ghost" math to solve it, and derives the smooth lake theory as a natural result.

It confirms that:

• The dominoes really do act like a smooth, wavy surface.
• The "ripples" in this surface tell us everything we need to know about how far-apart dominoes relate to each other.
• The math works the same way for square floors and honeycomb floors.

Summary

The paper is a bridge. It takes a rigid, blocky puzzle (dominoes on a grid) and uses a set of mathematical "ghosts" to show that, deep down, the puzzle is actually a story about smooth, flowing waves. It proves that the complex behavior of the tiles is just the shadow of a simple, elegant wave equation.

Technical Summary

Technical Summary: Derivation of Height Field Theory for the 2D Classical Dimer Model

Problem Statement
The classical dimer model on bipartite lattices (specifically square and honeycomb) exhibits a Coulomb phase characterized by algebraic correlations and topological order. While the long-wavelength physics of this system is widely understood to be described by a continuum height field theory, previous derivations often rely on heuristic coarse-graining or transfer-matrix methods that treat spatial dimensions asymmetrically. The challenge addressed in this work is to provide a constructive, rigorous derivation of this height field theory starting directly from the exact microscopic representation of the dimer model, treating both spatial dimensions on an equal footing and explicitly incorporating source terms to describe dimer observables.

Methodology
The author employs a multi-step constructive approach:

1. Grassmann Integral Representation: The partition function of the dimer model, including sources coupled to the flux density ($B_\ell$) and the valence-bond-solid (VBS) order parameter (magnetization, $\Psi_i$), is first expressed exactly as an integral over Grassmann variables. This is achieved by mapping dimer configurations to directed loop configurations on a Kasteleyn graph, utilizing a fixed reference configuration to define the mapping.
2. Continuum Limit (Dirac Fermions): The discrete Grassmann integral is expanded around the "Dirac points" (zero modes) of the fermionic dispersion relation. By retaining only slowly varying modes near these points, the lattice theory is mapped to a continuum theory of massless Dirac fermions in two-dimensional Euclidean space. The source terms couple to these fermions as a $U(1)$ gauge field (coupling to flux) and a mass term (coupling to the VBS order parameter).
3. Bosonization: The continuum fermionic path integral is then mapped to a bosonic field theory using standard bosonization identities within the path-integral formulation. This step transforms the fermionic operators into a real scalar field, identified as the continuum height field $h(\mathbf{r})$.
4. Boundary Conditions and Flux: The derivation explicitly handles periodic boundary conditions by summing over spin sectors (boundary phases) for the fermions. This summation is shown to generate the correct topological sectors for the bosonic height field, specifically linking the boundary discontinuity (tilt) of the height to the global flux of the dimer configuration.

Key Contributions and Results

• Constructive Derivation: The paper derives the continuum height field theory directly from the Grassmann integral representation, avoiding the intermediate step of transfer matrices. This method maintains explicit real-space structure until the continuum limit is taken, allowing for a unified treatment of both square and honeycomb lattices.
• Exact Mapping of Observables: The work explicitly derives the relationship between microscopic dimer occupation numbers and the continuum fields. It demonstrates that the flux density corresponds to the gradient of the height field ($B_\mu \propto \epsilon_{\mu\nu}\partial_\nu h$), and the VBS order parameter corresponds to vertex operators ($\Psi \propto e^{\pm 2\pi i h}$).
• Source Term Necessity: The author proves that coupling to the flux density and the local VBS order parameter are the only source terms required to describe the asymptotic long-distance correlations of the dimer model. Other potential source terms vanish in the continuum limit due to oscillating phase factors.
• Flux Distribution: The derivation recovers the exact distribution of the global flux $\Phi$, showing it arises from the summation over fermion boundary phases, which translates to a sum over the topological sectors (winding numbers) of the height field.
• Correlation Functions: Using the derived action, the paper reconstructs the standard algebraic decay of dimer-dimer correlation functions for both square and honeycomb lattices, confirming agreement with known exact results.

Significance and Claims
The paper claims that its primary significance lies in the constructive nature of the derivation. Unlike previous approaches that often posit the height theory based on symmetry arguments or fix coefficients by matching to exact solutions, this work derives the theory and its coefficients (specifically the stiffness $\kappa = \pi$) directly from the microscopic model.

The author emphasizes that this approach:

• Treats the square and honeycomb lattices on an equal footing, yielding identical continuum theories (up to lattice-specific scaling factors).
• Clarifies the origin of the boundary conditions for the height field, showing they emerge naturally from the fermion boundary phases.
• Provides a rigorous justification for the specific form of the source terms coupling to the dimer observables, demonstrating that no other terms are necessary for long-wavelength physics.

The work does not propose new experimental applications or future extensions beyond noting that the method could potentially be adapted to systems with arbitrary link potentials, spatial disorder, or quasicrystals, provided the Dirac points in the spectrum remain intact. It also acknowledges that while the height theory is well-established, this derivation offers a more direct link between the microscopic dimer variables and the continuum field operators.