Random matrix theory of integrability-to-chaos transition

This paper proposes a simple random matrix ensemble based on the power-law statistics of nonintegrable perturbation matrix elements to quantitatively describe the universal level spacing distributions observed during the transition from quantum integrability to chaos.

The Gist

Imagine you are a music critic trying to understand the "vibe" of a symphony orchestra.

In the world of quantum physics, the "notes" an atom or a particle plays are its energy levels. Scientists have long known two extreme ways these notes can be arranged:

1. The "Perfectly Organized" Orchestra (Integrable): Imagine a choir where everyone sings a note, and the gaps between the notes are completely random, like people picking numbers out of a hat. There's no pattern, no tension. In physics, this is called Poisson statistics. It happens in systems that are perfectly predictable and "solvable."
2. The "Chaotic Jam Session" (Chaos): Now imagine a jazz band where the musicians are so connected that they can't help but react to each other. If one person plays a low note, the next person avoids playing a low note right after. They "repel" each other to create a complex, interwoven sound. This is called Wigner-Dyson statistics. It happens in systems that are chaotic and unpredictable.

The Missing Middle Ground

For decades, physicists have been great at describing the "Perfectly Organized" and the "Chaotic Jam Session." But what happens when you take a perfectly organized choir and slowly start introducing a chaotic jazz drummer? The music changes from organized to chaotic.

This is the Integrability-to-Chaos Transition.

Until now, scientists didn't have a good recipe to describe the music during this transition. They had some rough guesses (like the "Brody distribution"), but they were just mathematical curves fitted to the data, not a deep understanding of why the music sounded that way. It was like trying to describe a sunset by just saying, "It's sort of orange and pink," without understanding the physics of light scattering.

The Paper's Big Idea: The "Secret Sauce"

The authors of this paper (Ben Craps, Marine De Clerck, Oleg Evnin, and Maxim Pavlov) discovered the secret ingredient that controls this transition.

They realized that to predict how the energy levels behave during the transition, you don't need to know the exact details of the whole system. You only need to look at the "matrix elements" of the chaotic part.

Here is a simple analogy:
Imagine the "Perfectly Organized" orchestra is a grid of empty seats (the energy levels). The "Chaotic Drummer" is a person walking through the aisle, tapping people on the shoulder.

• The Integrable System is just the empty seats.
• The Chaos is the drummer tapping people.
• The Transition is what happens when the drummer starts tapping.

The authors found that the pattern of the music (the spacing between notes) depends entirely on how the drummer taps.

• Does he tap lightly or hard?
• Does he tap random people, or people sitting next to each other?
• What is the distribution of the strength of his taps?

They built a Random Matrix Model (a computer simulation) that mimics this. Instead of simulating the whole complex orchestra, they just simulated:

1. A list of empty seats (the original energy levels).
2. A list of "tap strengths" (the chaotic perturbation).

When they ran their simulation using the actual distribution of tap strengths found in real physical systems, it perfectly recreated the messy, intermediate music of the transition.

The Surprising Discovery: The "Power Law"

While studying these "tap strengths," the authors found something truly weird and beautiful.

In almost every physical system they tested (from spinning atoms to vibrating particles), the distribution of these tap strengths followed a Power Law.

The Analogy:
Imagine you are counting how many people in a city have a certain amount of money.

• In a normal distribution (like height), most people are average, with very few giants or dwarfs.
• In a Power Law, you have a few billionaires, a lot of middle-class people, and a massive number of people with very little money. It's a "long tail."

The authors found that the "taps" of the chaotic perturbation behave exactly like this. There are a few huge interactions, many medium ones, and a vast ocean of tiny, almost invisible interactions. This "Power Law" shape seems to be a universal rule of nature, appearing in totally different systems like spinning chains of atoms and vibrating bosons.

Why Does This Matter?

1. A New Universal Language: They provided a simple "recipe" (a random matrix ensemble) that works for almost any system transitioning from order to chaos. You don't need to solve the impossible math of the whole system; you just need to know the statistics of the "taps."
2. Reading the Future: If you measure the energy levels of a new, unknown quantum system (like a new type of superconductor or a quantum computer chip), you can look at the spacing between the levels. Using their model, you can work backward to figure out the hidden properties of the system's "chaotic taps" without ever seeing the atoms directly.
3. Solving the "Middle" Problem: They finally gave a quantitative, mathematical description of that messy, in-between state where order is breaking down into chaos.

In a Nutshell

The paper says: "We used to think the transition from order to chaos in quantum systems was too messy to predict. But we found that it's actually controlled by a simple, universal pattern in how the system's parts interact. By studying this pattern (which looks like a 'power law'), we can build a simple model that predicts exactly how the system will behave, no matter how complex it actually is."

It's like realizing that while a storm looks chaotic, the wind speed follows a simple, predictable rule that lets you forecast the weather.

Technical Summary

1. Problem Statement

The transition from quantum integrability to quantum chaos is a fundamental problem in theoretical physics.

• Integrable Systems: Characterized by uncorrelated energy levels following a Poisson distribution ($P_P(s) = e^{-s}$).
• Chaotic Systems: Characterized by level repulsion and correlations described by Random Matrix Theory (RMT), specifically the Gaussian Orthogonal Ensemble (GOE) leading to the Wigner-Dyson distribution ($P_W(s) \approx \frac{\pi s}{2} e^{-\pi s^2/4}$).
• The Gap: While the limits are well-understood, the intermediate regime (where an integrable Hamiltonian is perturbed by a chaotic deformation) lacks a universal, quantitative description. Existing phenomenological fits (e.g., Brody distribution) or basic RMT models (e.g., Rosenzweig-Porter) fail to accurately reproduce the specific level spacing statistics observed in diverse physical systems during this crossover.

2. Methodology

The authors propose a new, physically motivated random matrix ensemble to model the crossover. Their approach relies on the insight that the level spacing statistics in the transition regime are controlled not by the full structure of the Hamiltonian, but specifically by the statistics of the matrix elements of the perturbation in the eigenbasis of the unperturbed integrable system.

The Proposed Ensemble

The physical Hamiltonian is modeled as $H = H_0 + \gamma H_1$, where $H_0$ is integrable and $H_1$ is chaotic. The authors construct a random matrix ensemble $H_{RMT}$ defined as:
$$H_{RMT} = D + \gamma' M$$
Where:

• $D$: A diagonal matrix containing the eigenenergies of the integrable part ($H_0$).
• $M$: A zero-diagonal symmetric matrix representing the perturbation.
• $\gamma'$: A tunable coupling parameter.

Key Innovation: The entries of $M$ are not drawn from a standard Gaussian distribution (as in GOE). Instead, the distribution of the off-diagonal elements of $M$ (denoted $P_M$) is extracted directly from the numerical data of the specific physical perturbation $H_1$ in the eigenbasis of $H_0$. The diagonal entries of $M$ are set to zero because their contribution is negligible in the crossover regime.

Validation Protocol

To compare the physical system and the random matrix ensemble:

1. Spectral Unfolding: The energy spectra are unfolded to remove global density variations, ensuring the mean level spacing is unity.
2. Parameter Matching via $r$-ratio: Instead of matching the coupling constants $\gamma$ and $\gamma'$ directly, the authors match the average $r$-ratio ($\bar{r}$). The $r$-ratio is defined as $r_n = \min(\delta_n, \delta_{n-1}) / \max(\delta_n, \delta_{n-1})$, where $\delta_n$ is the level spacing.
   • $\bar{r} \approx 0.386$ corresponds to Poisson (integrable).
   • $\bar{r} \approx 0.536$ corresponds to GOE (chaotic).
   • By tuning $\gamma'$ in the ensemble to match the $\bar{r}$ of the physical system, they compare the full level spacing distributions at equivalent stages of the transition.

3. Key Contributions

A. A Universal Random Matrix Ensemble

The paper demonstrates that a simple random matrix ensemble, where the perturbation matrix elements follow the empirical distribution of the physical system's perturbation, accurately reproduces the level spacing statistics across the integrability-to-chaos transition. This holds true for systems with vastly different degrees of freedom.

B. Discovery of Universal Power Laws

A major theoretical finding is the statistical nature of the perturbation matrix elements ($P_M$). Across diverse systems (spin chains, resonant bosonic systems, billiards, oscillators), the distribution of the magnitude of off-diagonal matrix elements follows a power law:
$$P_M(x) \sim \frac{e^{-Cx^2}}{x^p}$$
where the exponent $p$ typically lies in the interval $[0, 1]$. The authors argue that this power-law behavior, rather than the specific cutoff or small/large deviations, is the dominant factor controlling the crossover statistics.

C. Superiority Over Existing Models

The proposed ensemble significantly outperforms:

• Brody Distribution: A phenomenological fit that lacks a physical basis and fails for certain spin chains.
• Rosenzweig-Porter (RP) Model: A standard RMT crossover model that fails to quantitatively match the level spacing distributions of specific physical systems, particularly in the intermediate regime.

4. Results

The authors validated their construction on two distinct physical models:

1. Spin-1/2 Ising Chain:

   • System: Integrable transverse field Ising chain ($H_0$) perturbed by a chaotic XY-Heisenberg Hamiltonian ($H_1$).
   • Result: The random matrix ensemble, using the empirically derived $P_M$ from the spin chain, matched the level spacing distribution of the physical system almost perfectly across four different values of the average $r$-ratio ($0.4, 0.42, 0.45, 0.48$). The RP model and Brody fit showed significant deviations.

2. Quantum Resonant Systems (QRS):

   • System: Bosonic many-body models with resonant interactions. An integrable limit ($H_0$) was perturbed by a generic chaotic QRS realization ($H_1$).
   • Result: Similar to the spin chain, the proposed ensemble captured the transition statistics with high precision, while the RP model failed to reproduce the correct curve shape.

3. Additional Systems (Appendix):

   • The power-law behavior of matrix elements was confirmed in rectangular billiards and perturbed anisotropic oscillators, suggesting the universality of the underlying mechanism.

5. Significance and Implications

• Quantitative Bridge: The work provides the first quantitative, predictive framework for the integrability-to-chaos transition that is not merely a phenomenological fit but is rooted in the statistical properties of the Hamiltonian's matrix elements.
• Spectroscopic Diagnostics: The findings suggest that one can infer the statistical properties of a system's Hamiltonian (specifically the perturbation structure) solely from spectroscopic data (level spacing statistics). This is relevant for experimental systems like ultracold atomic gases and quantum processors where full Hamiltonian reconstruction is difficult.
• Connection to ETH: The observed power laws in matrix elements connect to the Eigenstate Thermalization Hypothesis (ETH). The paper suggests that deviations from ETH in integrable dynamics may be governed by these specific power-law distributions of matrix elements.
• Future Directions: The authors identify the analytical derivation of level spacing statistics for ensembles with power-law distributed off-diagonal elements as a significant open problem for future theoretical work, potentially solvable via statistical field theory methods.

In summary, the paper establishes that the "fingerprint" of the integrability-to-chaos transition is encoded in the distribution of the perturbation's matrix elements, which universally exhibit power-law behavior, allowing for the construction of a highly accurate, system-specific random matrix model.