Universal spectral correlations in open Floquet systems with localized leaks

This paper demonstrates that introducing localized leaks in time-reversal symmetric Floquet systems, such as the leaky quantum standard map, induces universal short-range spectral correlations governed by the non-Hermitian $\mathrm{AI}^{\dagger}$ symmetry class (modeled by the truncated circular orthogonal ensemble) rather than the unconstrained Ginibre ensemble, with the latter only emerging in the global density of states under sufficiently strong leakage.

The Gist

The Big Picture: A Noisy Room and a Leaky Bucket

Imagine you have a perfectly sealed, echoey room (a closed quantum system). If you clap your hands, the sound bounces around in a very specific, predictable pattern based on the room's shape. In the world of physics, this "sound" is like the energy levels of a particle. When the room is chaotic (like a room full of weirdly shaped mirrors), the sound patterns follow a famous rule called Wigner-Dyson statistics. Think of this as a "perfectly organized chaos" where the notes are spaced out just right so they don't clash.

Now, imagine you punch a hole in the wall of that room (a leak). Sound starts escaping. The room is no longer sealed; it's an open system. The sound doesn't just bounce; it fades away. In physics terms, the energy levels become "complex" (they have a decay rate, like a sound getting quieter).

The Big Question: When you make a hole in this chaotic room, do the remaining sound patterns still follow the old "organized chaos" rules? Or do they turn into something completely random and messy?

The Experiment: The "Kicked Rotor"

The researchers studied a specific model called the Quantum Kicked Rotor.

• The Analogy: Imagine a spinning top that gets kicked periodically.
• The Closed Version: If the top is in a vacuum, it spins forever. Its behavior is chaotic but follows the "Circular Orthogonal Ensemble" (COE) rules.
• The Open Version: They introduced a "leak." Imagine a specific zone on the floor where, if the top touches it, it disappears (escapes). This is the Leaky Quantum Standard Map (L-QSM).

They wanted to see if the "music" (the spectrum of energy levels) of this leaking top followed the rules of a Ginibre Ensemble (a standard model for random, leaking systems) or something else.

The Surprise: It's Not Just "Random"

For a long time, physicists thought that if you opened a chaotic system, it would just become a generic, random mess (the Ginibre Unitary Ensemble). Imagine throwing a handful of marbles onto a table; they scatter randomly. That's what they expected.

But they found something different.

Because the original system had a special symmetry (it looked the same if you played the movie backward, known as Time-Reversal Symmetry), the leaking system didn't become totally random. Instead, it became a Complex Symmetric mess.

The Analogy:

• The Old Expectation (GinUE): Like a crowd of people running in a panic, bumping into each other in every direction.
• The Reality (GinAI†): Like a crowd of people running in a panic, but they are all wearing identical, mirrored costumes. They still run chaotically, but their movements have a hidden "mirror" structure. They are less free than the totally random crowd, but more chaotic than the sealed room.

The paper proves that the "leaky" system follows the rules of this GinAI† class (a specific type of non-Hermitian symmetry), not the totally random class.

Key Findings Explained Simply

1. The "Stickiness" of Chaos

In the chaotic room, some paths are "sticky." If a particle gets near a certain spot, it tends to hang around there for a while before escaping.

• What they found: If you put the leak in a "sticky" spot, particles escape faster. If you put it in a "slippery" spot, they hang around longer.
• The Result: This affects the global picture (how many particles are left), but it does not change the local rules of how the particles bump into each other. The "local" rules are stubborn and stick to the GinAI† pattern regardless of where the leak is.

2. The Size of the Hole Matters (But Not How You Think)

You might think, "If I make the hole tiny, the system should act like the sealed room. If I make it huge, it should act like pure randomness."

• The Twist: The transition isn't about the percentage of the hole, but the number of "channels" (columns in the math matrix) you remove.
• The Analogy: Imagine a choir of 1,000 singers.
  • If you silence just one singer, the choir still sounds like a choir (COE statistics).
  • If you silence 100 singers, the choir suddenly starts sounding like the "Complex Symmetric" mess (GinAI†).
  • Crucial Point: As the choir gets bigger (more singers), you need to silence fewer of them (a smaller percentage) to trigger this change. A tiny leak in a huge system is enough to break the old rules.

3. The "Circular Law" vs. The "Hole"

• Global View: If you look at the whole picture of where the particles are, they usually cluster near the edge of a circle (like a ring of light). This is because the "long-lived" particles are still holding on.
• The Extreme Case: Only if you make the leak massive (removing almost the whole system) does the distribution flatten out into a perfect, uniform circle (the Ginibre Circular Law).
• Takeaway: The "local" rules (how particles interact with neighbors) change very quickly with a small leak. The "global" rules (the overall shape of the distribution) take a huge leak to change.

The Bottom Line

This paper is like discovering that if you poke a small hole in a very complex, chaotic machine, the machine doesn't just fall apart into random noise. Instead, it rearranges itself into a new, specific type of organized chaos (the AI† class).

• The Leak: Acts like a filter that reveals a hidden symmetry.
• The Result: Even when a system is losing energy and leaking out, it remembers its past (its time-reversal symmetry) and follows a very specific set of "leaky" rules, distinct from total randomness.

In everyday terms: It's like realizing that even if you open a window in a stormy room, the wind doesn't just blow randomly; it creates a specific, predictable pattern of turbulence that depends on the shape of the window and the room, not just the wind itself.

Technical Summary

1. Problem Statement

The paper addresses the spectral statistics of open quantum chaotic systems, specifically Floquet systems (periodically driven) that possess time-reversal symmetry.

• Context: In isolated (closed) chaotic quantum systems, spectral correlations follow the Circular Orthogonal Ensemble (COE) statistics. When such systems are opened to an environment, the dynamics become non-unitary, eigenvalues become complex, and the system is often modeled by non-Hermitian random matrix theory.
• The Gap: Previous conjectures (e.g., Grobe-Haake-Sommer) suggested that open chaotic systems generally follow the statistics of the unconstrained Ginibre Unitary Ensemble (GinUE), characterized by uniform eigenvalue distribution and cubic level repulsion. However, this assumes a specific type of coupling.
• Specific Question: What happens when the "opening" is not a general dissipative coupling (modeled by master equations) but a spatially localized leak (probability escaping through a specific region of phase space)? Does the system still follow GinUE statistics, or does the underlying symmetry of the closed system (COE) impose a different universality class?

2. Methodology

The authors employ a combination of numerical simulations of a specific quantum model and comparisons with Random Matrix Theory (RMT) benchmarks.

• Physical Model: The Leaky Quantum Standard Map (L-QSM).
  • Based on the quantum kicked rotor with kick strength $K=10$ (strongly chaotic regime).
  • Leak Implementation: A "leak" is introduced by truncating the unitary Floquet propagator $\hat{U}$. Specifically, columns of the matrix corresponding to position states within a specific interval (the leak region) are set to zero. This creates a non-unitary operator $\tilde{U} = \hat{U}\hat{\Pi}$, where $\hat{\Pi}$ is a projection operator.
  • Parameters: Matrix dimension $N$ (up to $10^4$), leak width $\Delta q$, and leak position $\bar{q}_L$.
• Random Matrix Benchmarks:
  1. Truncated Circular Orthogonal Ensemble (TCOE): Matrices drawn from COE with $n_L$ columns set to zero. This serves as the direct RMT analog of the L-QSM.
  2. Ginibre Unitary Ensemble (GinUE): Unconstrained complex random matrices (Symmetry Class A).
  3. Ginibre Ensemble with Transpose Symmetry (GinAI†): Complex symmetric random matrices (Symmetry Class AI†), constructed as $G = (X + X^T)/\sqrt{2}$.
• Analysis Techniques:
  • Global Properties: Density of states (DOS) and eigenvalue distribution in the complex plane.
  • Local Properties: Short-range spectral correlations using:
    • Nearest-neighbor eigenvalue spacing distribution $P(s)$.
    • Ratio of consecutive spacings (radial $r$ and angular $\theta$ components).
  • Filtering: Short-lived resonances (near the origin) are filtered out to focus on the "bulk" of the spectrum where universal behavior is expected.

3. Key Contributions

1. Identification of the Correct Universality Class: The paper demonstrates that open Floquet systems with localized leaks and time-reversal symmetry do not follow the unconstrained GinUE statistics. Instead, their short-range spectral correlations belong to the non-Hermitian symmetry class AI† (complex symmetric matrices).
2. Role of Truncation Size: The authors establish that the transition to AI† statistics is governed by the absolute number of removed columns ($n_L$), not just the truncation ratio.
   • If $n_L \gtrsim 100$, the system exhibits GinAI† statistics regardless of the total matrix size $N$ (provided $N$ is large enough).
   • The COE limit (closed system behavior) is only recovered if the truncation is smaller than one full column ($n_L < 1$).
3. Decoupling of Local and Global Statistics: The paper reveals a distinct separation between local and global spectral properties:
   • Local correlations (level repulsion) converge to GinAI† even for relatively small leaks.
   • Global distribution (density of states) only approaches the uniform Ginibre circular law when the leak is very large (strong truncation). For intermediate leaks, eigenvalues concentrate near the unit circle.
4. Quantum-Classical Correspondence in Leaky Systems: The study confirms that "stickiness" (trajectories trapped in phase space regions) affects the global distribution of eigenvalues (specifically the density of long-lived resonances) but does not alter the local universal spectral correlations.

4. Key Results

• Spectral Correlations (Local):
  • The nearest-neighbor spacing distribution $P(s)$ and the ratio distributions ($P(r)$, $P(\theta)$) of the L-QSM show excellent agreement with the TCOE and the GinAI† ensemble.
  • They deviate significantly from the GinUE predictions. Specifically, the AI† class shows a lower peak and broader variance in spacing distributions compared to GinUE, and an enhanced probability of collinear eigenvalue triplets (angular correlations).
• Density of States (Global):
  • For small to intermediate leaks, the eigenvalues of both L-QSM and TCOE accumulate near the unit circle ($|\lambda| \approx 1$), reflecting long-lived resonances. This contrasts with the uniform disk distribution of GinUE.
  • As the leak size $\Delta q$ increases (approaching 0.9), the eigenvalue distribution homogenizes and converges to the Ginibre circular law (uniform distribution), but this requires a very strong leak.
• Scaling with Matrix Size:
  • As the system size $N$ increases, the leak size required to observe GinAI† statistics decreases. The critical parameter is the number of truncated columns $n_L = N \Delta q$. Once $n_L \gtrsim 100$, the universality class is established.
• Symmetry Explanation:
  • Although the open propagator $\tilde{U}$ is not symmetric, its non-zero eigenvalues are identical to those of the reduced operator $\hat{\Pi}\hat{U}\hat{\Pi}$ acting on the surviving subspace. Since the closed system is COE (symmetric), this reduced operator is complex symmetric, placing the system in the AI† class.

5. Significance

• Refinement of Quantum Chaos Theory: The work corrects the assumption that all open chaotic systems follow GinUE statistics. It highlights that the method of openness (localized leak vs. global dissipation) and the underlying symmetries of the closed system are crucial in determining the universality class.
• Experimental Relevance: The findings are directly applicable to experimental platforms with spatially localized losses, such as leaky optical or microwave billiards, where the geometry of the leak dictates the spectral statistics.
• Theoretical Insight: It provides a rigorous link between the truncation of unitary matrices (TCOE) and non-Hermitian random matrix classes, clarifying the interplay between residual unitary structure and emergent non-Hermitian chaos.
• Distinction of Scales: The paper underscores that "chaos" in open systems is multifaceted; local correlations can be universal (AI†) while global distributions remain non-universal (non-circular) until the leak becomes dominant.

In summary, the paper establishes that localized leaks in time-reversal symmetric Floquet systems drive the system into the GinAI† universality class, a result that depends on the absolute number of lost channels rather than the relative leak size, and distinguishes clearly between local spectral correlations and global eigenvalue distributions.