Timelike Entanglement Signatures of Ergodicity and Spectral Chaos

This paper demonstrates that timelike entanglement measures derived from the spacetime density kernel in the Rosenzweig-Porter model, including Tsallis entropy, imagitivity, and a newly defined kernel negativity, serve as sharp diagnostics for distinguishing between ergodic, fractal, and localized phases by exhibiting distinct growth patterns, spectral form factor-like structures, and correlations with fractal dimensions.

The Gist

Imagine you are trying to understand how a complex machine works by watching how information flows through it. Usually, scientists look at how information spreads across different parts of a machine at the same moment (spatial entanglement). But this paper asks a different question: What happens if we look at how a system connects to itself across time?

The authors are studying a specific mathematical model called the Rosenzweig-Porter (RP) model. Think of this model as a giant, chaotic switchboard with millions of wires. Depending on how you tune a dial (called $\gamma$), the switchboard behaves in three very different ways:

1. The Ergodic Phase (Chaotic): The wires are all mixed up. If you flip a switch, the signal spreads everywhere instantly and randomly.
2. The Localized Phase (Frozen): The wires are disconnected. A signal stays stuck in one spot and never travels.
3. The Fractal Phase (The Middle Ground): The signal travels, but only to a limited set of places. It's like a maze where you can wander, but you can't reach every corner.

The paper introduces a new tool called the "Spacetime Density Kernel." To understand this, imagine taking a movie of the system and laying all the frames out on a giant table. This "kernel" is a special mathematical object that captures how the system at the beginning of the movie (Time 0) is connected to the system at the end of the movie (Time $t$).

Here is what the authors discovered using this tool, explained through simple analogies:

1. The "Imaginary" Mess (Non-Hermiticity)

In physics, some things are "real" and predictable, while others are "imaginary" and chaotic. The authors found that in the Chaotic (Ergodic) phase, this "imaginary" mess grows very fast and stays high. It's like stirring a cup of coffee: the cream swirls wildly and never settles back into a neat pattern.

• In the Frozen (Localized) phase, there is almost no "imaginary" mess. The coffee stays still.
• In the Fractal phase, it's somewhere in between.

They call this measurement "Imagitivity." It tells them how much the system is scrambling information over time.

2. The "Dip-Ramp-Plateau" Dance

One of their most interesting findings involves a graph that looks like a specific dance move: a Dip, a Ramp, and a Plateau.

• The Dip: The signal drops quickly at the start (like a ball bouncing off the floor).
• The Ramp: The signal slowly climbs back up (like the ball rolling up a hill).
• The Plateau: The signal levels off (the ball reaches the top and stops).

They found that this "dance" happens perfectly in the Chaotic phase. It's a signature of true chaos. However, in the Frozen phase, the "Ramp" disappears entirely; the ball just drops and stops. In the Fractal phase, the ramp is weak and sluggish. This proves that their time-based tool can detect the same "chaos" that traditional methods find, but by looking at time instead of space.

3. The "Kernel Negativity" (The Ghost Signal)

This is the paper's most unique invention. They define a quantity called "Kernel Negativity."
Imagine you have a scale that measures "probability" (how likely something is to happen). In a normal world, probabilities are always positive numbers (0% to 100%).
However, in the Chaotic phase, this "Kernel Negativity" detects negative probabilities. Think of this as a "ghost signal"—a mathematical signature that says, "This system is so chaotic and interconnected that it behaves in ways that defy normal logic."

• Chaotic Phase: High "ghost signals" (high negativity).
• Frozen Phase: No "ghost signals" (zero negativity).
• Fractal Phase: A moderate amount of "ghost signals."

Crucially, the amount of this "negativity" tracks perfectly with how "spread out" the system's energy levels are. If the system is fully chaotic, the negativity is high. If it's frozen, the negativity vanishes.

The Big Picture

The authors essentially built a new "thermometer" for quantum chaos. Instead of just measuring how hot (chaotic) a system is at a single moment, they measure how the system's past and future are tangled together.

• If the system is chaotic: The time-tangle is strong, the "ghost signals" are loud, and the "dance" (dip-ramp-plateau) is clear.
• If the system is frozen: The time-tangle is weak, the "ghost signals" are silent, and the dance is broken.
• If the system is fractal: It's a mix of both.

By using this "timelike entanglement," they can distinguish between these three states of matter with high precision, offering a new way to see how information scrambles and spreads through the quantum world.

Technical Summary

Technical Summary: Timelike Entanglement Signatures of Ergodicity and Spectral Chaos

Problem Statement
The study of quantum chaos and information scrambling typically relies on distinct tools for different temporal regimes: out-of-time-ordered correlators (OTOCs) for early-time scrambling and spectral statistics (e.g., spectral form factors) for late-time ergodicity. However, a unified framework that treats temporal and spatial correlations on comparable footing remains limited. While the Rosenzweig-Porter (RP) model is a standard testbed for studying the crossover between ergodic, fractal (non-ergodic extended), and localized phases, existing diagnostics like inverse participation ratios (IPR) and level statistics characterize these phases as static properties. They do not directly elucidate how temporal correlations and timelike entanglement evolve under real-time dynamics. This work addresses the need for a unified language to diagnose both eigenvector ergodicity and spectral chaos within a single spacetime-kernel framework.

Methodology
The authors investigate a system of $M$ qubits ($N=2^M$) governed by the Rosenzweig-Porter Hamiltonian, $H = H_0 + \frac{\epsilon}{N^{\gamma/2}}V$, where $H_0$ is a diagonal matrix of random on-site energies and $V$ is drawn from the Gaussian Orthogonal Ensemble (GOE). The control parameter $\gamma$ dictates the phase:

• $0 < \gamma < 1$: Ergodic phase (Wigner-Dyson statistics).
• $1 < \gamma < 2$: Fractal phase (non-ergodic extended).
• $\gamma > 2$: Localized phase (Poisson statistics).

To probe timelike correlations between subsystem $A$ (at time $0$) and subsystem $B$ (at time $t$), the authors construct a spacetime density kernel $T_{AB}(t)$. This operator acts on $H_A \otimes H_B$ and is defined via the trace-pairing identity:
$$\text{Tr}[T_{AB}(t) (O_A \otimes O_B)] = \text{Tr}[\rho O_A U^\dagger_t O_B U_t]$$
where $U_t = e^{-iHt}$ and $\rho$ is a random pure initial state. Unlike a standard reduced density matrix, $T_{AB}(t)$ is generally non-Hermitian.

The study evaluates the time evolution of several diagnostics derived from this kernel:

1. Schatten Norms: Specifically the Frobenius norm ($\|T_{AB}(t)\|_2^2$), quantifying the strength of two-time correlations.
2. Tsallis Entropy: The second Tsallis entropy $Z_2(t) = \text{Tr}[T_{AB}(t)^2]$, which can be complex.
3. Imagitivity: The entanglement $p$-imagitivity $I_2(t) = \|T_{AB}(t) - T^\dagger_{AB}(t)\|_2$, quantifying non-Hermiticity and causal influence.
4. Kernel Negativity ($N_H(t)$): A new diagnostic defined as the sum of the magnitudes of the negative eigenvalues of the Hermitian part of the kernel, $H_{AB}(t) = (T_{AB}(t) + T^\dagger_{AB}(t))/2$. This quantity equals the trace-norm distance to the set of positive semidefinite operators and represents the maximal witnessable negative quasiprobability.

These measures are compared against the Haar-averaged two-point function $F_2(t, \beta)$, which serves as a standard spectral diagnostic analogous to the spectral form factor (SFF).

Key Results
The numerical analysis (using $M=6$ qubits with symmetric bipartitions) reveals distinct behaviors across the three phases:

• Ergodic Regime ($\gamma \lesssim 1$):

  • The Frobenius norm exhibits rapid initial growth and saturates at a high late-time plateau.
  • The 2-imagitivity grows rapidly, indicating strong non-Hermiticity and causal influence between $A$ and $B$.
  • The real part of the second Tsallis entropy, $\text{Re } Z_2(t)$, displays a clear dip-ramp-plateau structure, mirroring the behavior of the spectral form factor $F_2(t, 0)$. This indicates robust spectral rigidity and level repulsion.
  • Kernel negativity $N_H(t)$ develops rapidly to a large plateau, signifying strong non-classical temporal correlations.

• Fractal Regime ($1 \lesssim \gamma \lesssim 2$):

  • All measures exhibit intermediate behavior. The Frobenius norm and imagitivity grow more slowly and saturate at lower values compared to the ergodic phase.
  • The dip-ramp-plateau structure in $\text{Re } Z_2(t)$ persists but is systematically weakened; the ramp is stretched and less sharp, reflecting intermediate spectral rigidity.
  • Kernel negativity is present but reduced.

• Localized Regime ($\gamma \gtrsim 2$):

  • The Frobenius norm and imagitivity are strongly suppressed, with the kernel remaining close to an effectively classical, positive-semidefinite object.
  • The ramp in $\text{Re } Z_2(t)$ is strongly suppressed, and the curves approach the plateau directly, consistent with the loss of long-range spectral rigidity (Poisson statistics).
  • Kernel negativity is minimal, remaining close to zero.

• Correlation with Fractal Dimension:
  The time-averaged kernel negativity and the Frobenius norm decrease monotonically as $\gamma$ increases from the ergodic to the localized regime. Their trends closely track the fractal dimension $D_2$ (where $D_2 \approx 1$ in ergodic, $0 < D_2 < 1$ in fractal, and $D_2 \approx 0$ in localized phases). The approach to the late-time plateau for these observables follows a Kohlrausch–Williams–Watts (stretched/compressed exponential) function, consistent with sub-diffusive physics in the multifractal phase.

Significance and Claims
The paper claims to provide a unified framework that bridges quantum information and quantum chaos diagnostics. The primary contributions are:

1. Unified Diagnosis: The spacetime density kernel simultaneously captures eigenvector ergodicity (via norm growth and imagitivity) and spectral chaos (via the dip-ramp-plateau structure in Tsallis entropy).
2. New Diagnostic (Kernel Negativity): The introduction of kernel negativity as a quantitative marker for nonclassical temporal correlations. The authors demonstrate that this quantity is highly sensitive to the phase of the RP model and serves as a dynamical witness for the fractal phase, complementing static measures like IPR.
3. Operational Meaning: The kernel negativity is shown to equal the trace-norm distance to positive semidefinite operators and the maximal witnessable negative quasiprobability, providing a concrete operational interpretation.
4. Dynamical vs. Static: The work establishes that time-averaged dynamical observables derived from real-time evolution can track the static fractal dimension of eigenstates, offering a time-domain counterpart to the fractal phase characterization.

The authors conclude that these timelike entanglement-based observables offer a robust method to probe the transition between integrable and chaotic dynamics, with potential applications in understanding horizon formation and interior physics in holographic settings, though they note that precise measurement protocols and extensions to other universality classes are directions for future work.