More scaling limits for 1d random Schrödinger operators with critically decaying and vanishing potentials

This paper extends the scaling limit theory for one-dimensional random Schrödinger operators with critically decaying potentials by characterizing the limits of transfer matrices and eigenvalue point processes for intermediate decaying profiles via coupled stochastic differential equations, thereby identifying new point processes with properties similar to the $\text{Sch}_\tau$ process and describing the corresponding eigenfunction shapes.

The Gist

Imagine you are standing in a long, narrow hallway made of $n$ tiles. This hallway represents a one-dimensional world where tiny particles (like electrons) are trying to walk from one end to the other.

In a perfect, empty hallway, the particles would walk smoothly and predictably. But in this paper, the author, Yi Han, studies a chaotic hallway. Every few tiles, there is a random "bump" or "obstacle" (a potential) that pushes the particle off course. The strength of these bumps changes as you walk down the hall.

The big question in physics is: Do the particles get stuck in one spot (localized), or do they eventually wander freely through the whole hallway (delocalized)?

This paper investigates a specific, tricky middle ground where the bumps get weaker as you walk, but not too fast and not too slow. Here is the breakdown using simple analogies:

1. The Three Types of Hallways

To understand the new discovery, we first need to know the two extreme cases the author is bridging:

• The "Vanishing" Hallway: Imagine the bumps are huge at the start but disappear completely by the time you reach the end. In this case, the particles eventually behave like they are in a perfect hallway. The pattern of their energy levels looks like a rigid, predictable fence (called a "Clock process").
• The "Decaying" Hallway: Imagine the bumps are strongest at the very beginning and get weaker the further you go, but they never quite disappear. Here, the particles get stuck in specific spots. The energy levels look like a random, messy crowd (called "Poisson statistics" or "Sine process").

The Author's Discovery:
Yi Han looks at a hallway where the bumps fade away at a critical speed—a "Goldilocks" speed that is neither too fast nor too slow. This is the "mixed vanishing-decaying" model. It's like a hallway where the obstacles are fading just fast enough to keep things interesting, but slow enough to keep the chaos alive.

2. The Magic of "Transfer Matrices" (The Compass)

To figure out what the particles are doing, the author uses a mathematical tool called a Transfer Matrix.

• The Analogy: Imagine you are walking through the hallway with a compass. Every time you step on a tile with a bump, the compass spins a little bit. The "Transfer Matrix" is the record of all those spins.
• The Result: The author proves that if you watch this compass spin as the hallway gets infinitely long, its movement follows a specific, random dance known as a Stochastic Differential Equation (SDE). It's like the compass is being pushed by a gentle, random wind (Brownian motion) while also trying to march forward in a straight line.

3. The New "Energy Fingerprint" (Point Processes)

When you measure the energy of the particles in this hallway, you get a list of numbers (eigenvalues). In the extreme cases, these numbers form either a perfect grid or total chaos.

In this new "Goldilocks" hallway, the author finds a brand new pattern.

• The Analogy: Think of the energy levels as raindrops hitting a tin roof.
  • In the "Clock" case, they hit in a perfect rhythm: tick-tock-tick-tock.
  • In the "Poisson" case, they hit randomly: clack... clack-clack... clack.
  • In this new case, the raindrops hit in a pattern that is neither perfectly rhythmic nor totally random. It's a unique, complex rhythm that the author calls $\eta$Sch. It's like a jazz drum solo that has a structure but feels improvised.

The author describes this new pattern using a "Brownian Carousel"—a fancy way of saying the energy levels are determined by a spinning wheel that is being jostled by random noise.

4. The Shape of the Particles (Eigenfunctions)

The paper also asks: Where are the particles actually standing?

• In the "stuck" phase, a particle is a tiny dot in one corner.
• In the "free" phase, the particle is a smooth wave spread out evenly.

In this new middle ground, the author finds that the particles form a strange, wavy shape.

• The Analogy: Imagine a spotlight shining on a dancer. In the old models, the light was either a tight beam on one spot or a floodlight covering the whole stage. In this new model, the spotlight is a fuzzy, moving cloud that is brightest in the middle and fades out at the edges, but the exact shape of the cloud is determined by a random, rolling hill (a Brownian motion).

5. Why Does This Matter?

This paper is important because nature often lives in the "middle." Real materials aren't perfectly ordered or perfectly chaotic; they usually have imperfections that fade out over distance.

By finding the mathematical rules for this "in-between" state, the author:

1. Connects the dots: Shows how the world transitions from "stuck" to "free."
2. Creates new math: Defines a new type of random pattern (the $\eta$Sch process) that didn't exist in textbooks before.
3. Predicts behavior: Gives scientists a way to calculate how likely it is to find two energy levels very close together (repulsion) or very far apart (gaps).

Summary

Yi Han has discovered a new "flavor" of randomness in the quantum world. By studying a hallway where obstacles fade at just the right speed, they found a new, unique rhythm to the universe's energy levels—a rhythm that is more complex than a metronome but more structured than a coin toss. It's a beautiful piece of math that helps us understand how order emerges from chaos.

Technical Summary

1. Problem Statement and Context

The paper investigates the spectral properties of one-dimensional random Schrödinger operators ($H_n$) defined on a finite lattice $\{0, 1, \dots, n\}$. The operator is given by:
$$(H_n\psi)_\ell = \psi_{\ell-1} + \psi_{\ell+1} + v_{\ell,n}\psi_\ell$$
with Dirichlet boundary conditions. The potential $v_{\ell,n}$ involves independent, identically distributed (i.i.d.) random variables $\omega_\ell$ with mean 0 and variance 1, scaled by a deterministic profile.

Background:
Previous work by Kritchevski, Valkó, and Virag [1] established scaling limits for two extreme cases at the critical decay rate $\alpha = 1/2$:

1. Vanishing Model: $v_{\ell,n} \propto \omega_\ell / \sqrt{n}$. The eigenvalue point process converges to the Sch$_\tau$ process (defined via solutions to a specific SDE).
2. Decaying Model: $v_{\ell,n} \propto \omega_\ell / \sqrt{\ell}$. The eigenvalue point process converges to the Sine$_\beta$ process (from Random Matrix Theory).

The Gap:
While the behavior for $\alpha > 1/2$ (Clock process) and $\alpha < 1/2$ (Poisson process) is understood, the intermediate regime where the potential decay profile lies between the pure vanishing and pure decaying cases had not been rigorously characterized. This paper fills that gap by analyzing a mixed vanishing-decaying model.

2. Methodology

The author employs a combination of stochastic analysis, transfer matrix methods, and the theory of point processes.

• Model Definition: The paper introduces a generalized potential profile:
  $$v_{k,n} = \sigma \frac{\omega_k}{n^\eta (n + 1 - k)^{1/2 - \eta}}$$
  where $\eta \in [0, 1/2]$.

  • $\eta = 1/2$ recovers the vanishing model.
  • $\eta = 0$ recovers the decaying model (after index reversal).
  • $\eta \in (0, 1/2)$ represents the intermediate regime.

• Transfer Matrices and SDE Limits:
  The core methodology involves analyzing the transfer matrices $M_n^\lambda$ associated with the operator. By diagonalizing the transfer matrix of the free operator and rescaling time ($t = \ell/n$), the author derives the scaling limit of the normalized transfer matrices $Q_n^\lambda$.

  • The limit is shown to satisfy a Stochastic Differential Equation (SDE) driven by Brownian motions.
  • Crucially, the diffusion coefficient in the SDE contains a singular term $(1-t)^{-(1/2-\eta)}$, which diverges as $t \to 1$ but remains integrable for $\eta > 0$.

• Prüfer Coordinates:
  To study the eigenvalues, the transfer matrix evolution is mapped to Prüfer coordinates $(r_\lambda, \theta_\lambda)$, where $\theta_\lambda$ represents the phase. The evolution of the phase is governed by an SDE resembling a "Brownian Carousel."

• Tightness and Convergence:
  The author establishes tightness estimates for the transfer matrices and the number of eigenvalues in small intervals (Wegner-type estimates). This allows for the convergence of the point process of eigenvalues to a limiting measure defined by the zeros of an analytic function derived from the SDE solution.

• Eigenfunction Shape Analysis:
  The paper analyzes the joint scaling limit of eigenvalue-eigenvector pairs. By uniformly selecting an eigenvalue, the local density of the eigenvector is shown to converge to a functional of a Brownian motion and a uniform random variable.

3. Key Contributions and Results

A. New Point Process Limits ($\eta_{Sch}$)

The primary result is the characterization of the scaling limit of the eigenvalue point process $\Lambda_n$ for $\eta \in (0, 1/2)$.

• Limit Process: The rescaled eigenvalues converge in distribution to a new point process denoted $\eta_{Sch}$ (or $\eta_{Sch}^*$ when the phase is randomized).
• Definition: $\eta_{Sch}$ is defined as the set of $\lambda$ such that the solution $\theta_\lambda(1)$ to the limiting SDE satisfies $\theta_\lambda(1) \in 2\pi\mathbb{Z}$.
• SDE Characterization: The phase $\theta_\lambda(t)$ solves:
  $$d\theta_\lambda(t) = \lambda dt + \frac{\sigma\rho}{(1-t)^{1/2-\eta}} [dB + \text{Im}(e^{-i\theta_\lambda(t)} dW)]$$
  where $B$ and $W$ are independent Brownian motions.

B. Properties of the New Point Process

The paper details the statistical properties of $\eta_{Sch}$, showing it shares features with both Sch$_\tau$ and Sine$_\beta$ but exhibits unique behavior:

1. Eigenvalue Repulsion: The probability of finding two eigenvalues within a small distance $\epsilon$ decays super-exponentially (faster than polynomial repulsion seen in GOE/GUE).
   $$P(\eta_{Sch}[\mu, \mu+\epsilon] \ge 2) \le 4 \exp(-C (\log(1/\epsilon))^2)$$
2. Large Gap Probability: The probability of a gap of length $\lambda$ decays as $\exp(-c \lambda^2)$. This quadratic decay in the exponent is distinct from the linear decay in Poisson processes and differs from the specific forms in Sine$_\beta$.
3. Central Limit Theorem (CLT): While the paper proves a crude upper bound for the fluctuations of the number of points, it notes that the refined Gaussian fluctuations known for Sch$_\tau$ ($\eta=1/2$) and Sine$_\beta$ ($\eta=0$) are not immediately generalizable to the intermediate $\eta$ case.

C. Shape of Eigenfunctions

The paper generalizes the "shape theorem" for eigenvectors.

• For a uniformly chosen eigenvalue $\mu$, the rescaled eigenvector density $n|\psi_\mu(\lfloor nt \rfloor)|^2$ converges in distribution to a random measure.
• The limit is given by a normalized exponential functional of a Brownian motion $Z$ and a uniform random variable $U$:
  $$\frac{\exp\left( \int_0^t \dots \right)}{\int_0^1 \exp\left( \int_0^s \dots \right) ds}$$
• This result interpolates between the uniform distribution (delocalized, $\eta \to 0$ limit) and the localized exponential decay profiles seen in the vanishing model.

D. Supercritical Decay

The paper also briefly addresses the case where the potential decays faster than the critical rate ($\tau > 1/2$ in the generalized model). In this regime, the point process converges to the Clock process (deterministic spacing), and eigenfunctions become uniformly spread out.

4. Significance

• Unification of Regimes: This work provides a continuous parameterization ($\eta$) that bridges the gap between the vanishing and decaying critical models, demonstrating a smooth transition in the scaling limits of both eigenvalues and eigenvectors.
• New Universality Class: The introduction of the $\eta_{Sch}$ process establishes a new universality class for 1D random Schrödinger operators. It demonstrates that the critical regime is not limited to just the Sch$_\tau$ and Sine$_\beta$ processes but contains a family of processes with distinct gap statistics and repulsion behaviors.
• Methodological Advancement: The rigorous handling of the singular SDE coefficients (where the diffusion term blows up at $t=1$) and the derivation of tightness estimates for these specific profiles provide a robust framework for analyzing other intermediate decay models.
• Physical Insight: The results offer deeper insight into the transition between localized and delocalized phases in disordered systems, specifically how the "memory" of the disorder profile affects the spectral statistics and wavefunction localization.

In summary, Yi Han's paper extends the theory of critical random Schrödinger operators by identifying and characterizing a new family of scaling limits that arise when the potential decay profile is a mixture of vanishing and decaying behaviors, revealing rich and previously unknown statistical structures in the spectrum.