Theory of Anderson localization on the hyperbolic plane

This paper presents a unified framework for studying Anderson localization on the hyperbolic plane by deriving a two-parameter flow that interpolates between low- and high-dimensional behaviors, revealing an extended critical line separating metallic and insulating phases.

The Gist

Imagine you are walking through a strange, magical landscape. In our normal world, if you walk a short distance, the ground looks flat. If you walk a long distance, it still looks flat; the world is just "big."

But in the world of this paper, the Hyperbolic Plane, the rules of space change depending on how far you look.

• Up close: If you stand on a tiny patch of this ground, it feels like a normal, flat sheet of paper (2-dimensional).
• From far away: If you zoom out, the ground doesn't just get bigger; it explodes outward. The amount of space available grows so fast that, effectively, the world feels infinite-dimensional. It's like standing in a room where the walls keep stretching away faster than you can walk, creating a vast, endless maze.

The scientists in this paper wanted to understand what happens to quantum particles (tiny bits of matter that act like waves) when they try to move through this weird, expanding landscape, especially when the landscape is messy or "disordered" (full of bumps and obstacles).

The Problem: Getting Lost vs. Getting Stuck

In physics, there's a famous phenomenon called Anderson Localization. Think of it like this:

• In a normal, flat world: If a particle is moving and hits random bumps, it usually gets confused. It bounces back and forth, interfering with itself, until it gets "stuck" in one spot. It can't travel far. This is called an insulator.
• In a very high-dimensional world: If the space is huge and has infinite directions to escape, the particle has so many ways to run away that it rarely gets stuck. It keeps moving freely. This is called a metal (or conductor).

Usually, a system is either one or the other. But the Hyperbolic Plane is special because it is both at the same time. It starts as a "stuck" world up close and becomes a "free" world far away.

The Solution: A Unified Map

The authors built a new mathematical map to describe this transition. They didn't just look at the "stuck" part or the "free" part separately; they created a single theory that connects them.

They used a tool called a Renormalization Group (RG) flow. Imagine you are looking at a map through a telescope that changes its zoom level:

1. Zoomed in (Short distances): The map looks like a flat, messy street. The particle gets confused by the bumps and tends to localize (get stuck).
2. Zoomed out (Long distances): The map reveals the exponential growth of space. The particle realizes there are too many escape routes to get stuck, so it starts flowing freely.

The paper's main discovery is a two-parameter flow. They found a way to track two things simultaneously:

1. Conductivity: How easily the particle moves.
2. Curvature: How "curved" or "expanding" the space is at that scale.

The Critical Line

By plotting these two factors, they found a Critical Line (a dividing line on their map).

• Above the line: The space is curved enough, or the disorder is low enough, that the particle stays free. It's a Metal.
• Below the line: The disorder is too strong, or the space isn't expanding fast enough to help, so the particle gets trapped. It's an Insulator.

The most surprising part is that this isn't a sharp switch. Because the space changes its "dimension" as you look at it, the transition is a smooth crossover. The paper shows exactly how a system can slide from being an insulator to a metal as you change the scale of observation.

The "Two-Terminal" Surprise

The authors also calculated what would happen if you tried to measure the resistance of this space (like measuring how hard it is to push electricity through a wire).

They found a counter-intuitive result: The size of the outer world doesn't matter.

Imagine a giant, expanding ring. If you try to push current from the center to the edge:

• In a normal ring, making the ring wider adds more resistance.
• In this hyperbolic ring, because the outer area is so vast (exponentially growing), it acts like a giant, infinite parallel highway. Even if the center is narrow, the massive outer area provides so many escape routes that the total resistance stops increasing once you get past a certain point. The resistance is determined almost entirely by the small region near the center, not the huge, expanding outer rim.

Summary

In simple terms, this paper explains how a quantum particle behaves in a space that feels flat up close but infinite far away. They created a unified theory showing that the particle's ability to move (conductivity) depends on a delicate balance between how messy the space is and how fast the space is expanding. They mapped out exactly where the particle gets stuck and where it flows freely, revealing that in this strange geometry, the "size" of the universe doesn't make it harder to conduct electricity; in fact, the vastness of the space helps the current flow.

Technical Summary

1. Problem Statement

The paper addresses the fundamental question of how Anderson localization (the absence of diffusion in disordered systems) behaves in a system where the effective dimensionality is scale-dependent.

• The System: The two-dimensional hyperbolic plane ($H^2$) with constant negative curvature radius $L$.
• The Paradox:
  • At short distances ($r \ll L$), the geometry is locally flat ($d_{eff} = 2$). In 2D, generic non-interacting systems are known to be localized for arbitrarily weak disorder due to constructive interference of semiclassical trajectories (weak localization).
  • At large distances ($r \gg L$), the volume grows exponentially ($V \sim e^{r/L}$), rendering the space effectively infinite-dimensional ($d_{eff} \to \infty$). In high dimensions, localization typically requires strong disorder to overcome the "escape" into the vast volume.
• The Gap: Previous studies treated these two regimes (weak disorder/low $r$ and strong disorder/high $r$) separately. There was no unified theoretical framework describing the dimensional crossover and the resulting phase diagram connecting the metallic and insulating phases.

2. Methodology

The authors employ a multi-pronged approach combining continuum field theory, renormalization group (RG) analysis, and lattice discretization:

A. Effective Field Theory (Continuum Limit)

• They utilize the nonlinear $\sigma$-model extended to curved Riemannian backgrounds.
• Action: $S[Q] = -\frac{\sigma}{16} \int d^2x \sqrt{g} \, \text{str}(Q \Delta Q)$, where $Q$ is a supermatrix field, $\sigma$ is the conductivity, and $\Delta$ is the Laplace-Beltrami operator on $H^2$.
• Symmetry: The model uses supersymmetry (class AI) to handle disorder averaging non-perturbatively.

B. Renormalization Group (RG) Analysis

• Mode Decomposition: Instead of standard Fourier modes, they use the Fourier-Helgason transform, which utilizes the eigenfunctions of the hyperbolic Laplacian.
• Flow Equations: By integrating out "fast" modes (high eigenvalues) and rescaling, they derive coupled RG flow equations for the conductivity $\sigma(\ell)$ and a dimensionless curvature parameter $u(\ell) = \Lambda L e^{-\ell}$ (where $\Lambda$ is the UV cutoff and $\ell$ is the RG scale):
  $$\frac{d\sigma}{d\ell} = -\frac{1}{2\pi} \frac{u^2 \tanh(\pi u)}{u^2 + 1/4}, \quad \frac{du}{d\ell} = -u$$
• Physical Interpretation:
  • As $u \to \infty$ (flat limit), the flow recovers standard 2D weak localization ($\sigma$ decreases linearly).
  • As $u \to 0$ (large scale/high curvature), the spectral gap and low mode density suppress the localization effect, mimicking high-dimensional behavior.

C. Lattice Discretization (Strong Disorder Regime)

• To analyze the regime where the continuum RG flow breaks down (near $u \sim 1$), they discretize $H^2$ onto a lattice.
• Discretization Choice: They compare several lattices (Bethe, Husimi cactus, regular tessellations) and select the Poisson-Delaunay (PD) triangulation. This choice is optimal because it:
  1. Reproduces exponential area growth.
  2. Contains loops (unlike Bethe lattices) but avoids extended loop networks (unlike regular tessellations).
  3. Matches the continuum Laplacian spectrum most accurately.
• Localization Mechanism: They generalize the Bethe lattice recursion relations to include loops and inhomogeneous bond weights. They find that short loops do not qualitatively alter the localization criterion; the transition is still driven by the competition between disorder strength and the coordination number (effective dimensionality).

3. Key Contributions and Results

A. Unified Two-Parameter Flow

The primary result is the derivation of a two-parameter flow in the $(\sigma, u)$ plane.

• The Separatrix: A critical line (separatrix) divides the parameter space into two phases:
  1. Metallic Phase: For initial conditions above the separatrix, the RG flow terminates at a finite, non-zero conductivity ($\sigma > 0$) as $u \to 1$. The system remains metallic.
  2. Insulating Phase: Below the separatrix, the conductivity flows to zero ($\sigma \to 0$) before the curvature scale is reached, leading to localization.
• Critical Line: The transition is not a single point but an extended critical line defined by the condition that the conductivity is suppressed to small values exactly when the curvature scale becomes relevant ($u \sim 1$).

B. Nature of the Transition

• The transition is characterized by mean-field-like critical exponents (similar to Bethe lattices):
  • Correlation length exponent: $\nu = 1$.
  • Order parameter exponent: $\beta = 1$.
• The authors demonstrate that the "crossover" from 2D to infinite-dimensional behavior is smooth in the flow but results in a sharp phase transition boundary due to the interplay between the RG flow and the lattice discretization scale.

C. Transport Properties (Resistance)

The paper derives the scale-dependent resistance $R(d)$ for a two-terminal setup on $H^2$:
$$R(d) = \frac{\rho(d)}{2\pi} \ln \left( \frac{\tanh(d/2L)}{\tanh(\epsilon/2L)} \right)$$

• Small $d$ ($d \ll L$): Recovers the standard 2D logarithmic resistance ($R \sim \ln d$).
• Large $d$ ($d \gg L$): The resistance saturates. Due to the exponential growth of the area, the outer regions of the system act as a giant parallel shunt. The total resistance is dominated by the resistivity of the region near the center electrode ($r \lesssim L$). This implies that arbitrarily large hyperbolic systems can have finite resistance even if the bulk is metallic.

4. Significance

• Theoretical Unification: The work provides the first unified framework bridging the gap between low-dimensional (loop-interference driven) and high-dimensional (strong-disorder driven) Anderson localization.
• Dimensionality as a Tunable Parameter: It establishes $H^2$ as a unique paradigm where effective dimensionality is a continuous function of scale, offering a new perspective on universality classes in phase transitions.
• Holography and Tensor Networks: The authors suggest implications for low-dimensional holography (AdS/CFT). Since the hyperbolic plane geometry is intrinsic to the constant time slices of AdS spaces, and the $\sigma$-model on $H^2$ relates to random matchgate tensor networks, these localization structures may inform the understanding of boundary correlations in holographic duals.
• Experimental Relevance: While $H^2$ is a mathematical abstraction, the results are relevant for systems with hyperbolic geometry, such as certain metamaterials, photonic lattices, and potentially in the study of quantum chaos and many-body localization in high-connectivity networks.

In summary, Altland et al. demonstrate that the hyperbolic plane hosts a robust metallic phase protected by its negative curvature, with a well-defined critical line separating it from the insulating phase, fundamentally altering the standard picture of Anderson localization in 2D.