Validity of generalized Gibbs ensemble in a random matrix model with a global $\mathbb{Z}_2$-symmetry

This paper demonstrates that in random symmetric centrosymmetric matrices with a global $\mathbb{Z}_2$ symmetry, thermalization of local observables is violated and accurately described by the generalized Gibbs ensemble, while specific initial states exhibit non-decaying behavior linked to spontaneous symmetry breaking in a measure-zero fraction of the ensemble.

The Gist

The Big Picture: Why Don't Things Always "Settle Down"?

Imagine you have a giant, chaotic room filled with thousands of bouncing balls (these represent particles in a quantum system). Usually, if you shake the room and let it go, the balls will bounce around, collide, and eventually spread out evenly. If you look at any small corner of the room, the balls will behave predictably, as if they have reached a comfortable "temperature." In physics, we call this thermalization. It's how a hot cup of coffee cools down to room temperature; the energy spreads out until everything is the same.

However, this paper asks: What happens if the room has a secret rule that stops the balls from mixing properly?

The author, Adway Kumar Das, investigates a specific type of "room" (a mathematical model called a Random Symmetric Centrosymmetric Matrix) where a hidden rule called $Z_2$ symmetry exists. This symmetry is like a mirror down the middle of the room.

The Mirror Analogy: Two Separate Worlds

Think of the system as a large dance floor with a giant, invisible mirror running down the center.

• The Rule: The dancers (quantum states) can only move in a way that respects this mirror. If a dancer on the left moves, their reflection on the right must move in a perfectly synchronized way.
• The Result: Because of this strict rule, the dance floor is effectively split into two separate, non-communicating rooms: a "Left Room" (Odd sector) and a "Right Room" (Even sector).

In a normal chaotic system, energy flows freely between all parts. But here, the energy is trapped. It's like having two separate parties in the same building where the guests can't talk to each other.

The Experiment: Testing the "Thermostat"

The author tests what happens when we start a "party" (an initial state) in this mirrored room.

1. The "Perfectly Balanced" Party (Symmetric States)
Imagine you start with a dancer exactly in the middle of the mirror, or a pair of dancers moving in perfect sync.

• What happens: Because they are perfectly balanced, they stay in their specific "room." They never mix with the other side.
• The Surprise: The author found that for a tiny, almost non-existent fraction of these systems, the dancers get stuck in a loop. They don't settle down; they keep oscillating back and forth forever. It's like a pendulum that never stops swinging. This is called Spontaneous Symmetry Breaking. Even though the room is chaotic, the dancers act like they are in a frozen, ordered state.

2. The "Messy" Party (Generic States)
Now, imagine you start with a dancer who is not perfectly balanced—maybe they are leaning slightly to the left.

• What happens: They bounce around, but because of the mirror rule, they can't explore the entire building. They are stuck in a specific pattern.
• The Problem: If you try to predict how this dancer will behave using standard physics (the Gibbs Ensemble, which is like a standard thermostat), you will get the wrong answer. The standard thermostat assumes everyone mixes freely. But here, the "mirror" is a conserved quantity that the standard thermostat ignores.

The Solution: The "Super-Thermostat"

Since the standard thermostat fails, the author proposes a new tool: the Generalized Gibbs Ensemble (GGE).

• The Analogy: Imagine you are trying to predict the temperature of a room, but you realize there is a locked door in the middle that no one can cross.
  • Standard Thermostat (Gibbs): Ignores the door. It says, "The whole room is one temperature." (Wrong!)
  • Generalized Thermostat (GGE): Acknowledges the door. It says, "Okay, the left side is this temperature, and the right side is that temperature, and we must respect the rule that no one crosses the door."

The paper proves mathematically that if you use this "Super-Thermostat" (GGE), which takes the mirror symmetry into account, you can perfectly predict the final state of the system.

Key Takeaways in Plain English

1. Symmetry is a Traffic Cop: In quantum systems, symmetries (like the mirror rule) act like traffic cops. They stop energy from flowing freely, preventing the system from "thermalizing" (settling into a standard equilibrium).
2. The "Measure Zero" Mystery: The author found a weird case where, for a vanishingly small number of specific systems, the symmetry breaks so badly that the system gets stuck in a long-lived loop. It's like finding a coin that lands on its edge and never falls over, but only if you flip it in a very specific, rare way.
3. We Need Better Tools: If you have a system with a global symmetry (like this mirror rule), you cannot use standard statistical mechanics to predict its behavior. You must use the Generalized Gibbs Ensemble, which adds an extra "rule" to the math to account for the symmetry.

Why Does This Matter?

This isn't just about math puzzles. These "mirror rules" appear in real-world physics, from how light moves through photosynthetic plants to how information is transferred in quantum computers.

If we build a quantum computer and don't account for these hidden symmetries, our predictions about how the computer behaves will be wrong. This paper gives us the correct "instruction manual" (the GGE) to understand and predict how these special, symmetrical quantum systems behave, ensuring we don't get fooled by their hidden rules.

Technical Summary

1. Problem Statement

The paper investigates the role of global symmetries, specifically the discrete $\mathbb{Z}_2$ symmetry, in the thermalization of isolated quantum systems. While the Eigenstate Thermalization Hypothesis (ETH) suggests that generic chaotic quantum systems thermalize to a Gibbs ensemble, the presence of conserved quantities (integrals of motion) can prevent this. The author focuses on Random Symmetric Centrosymmetric (SC) matrices, which commute with an exchange symmetry operator $J$.

• Core Question: Does a system with a global $\mathbb{Z}_2$ symmetry (splitting the Hilbert space into decoupled even and odd sectors) thermalize in the standard sense? If not, what statistical ensemble correctly describes its equilibrium state?
• Context: The study bridges random matrix theory (RMT) and non-equilibrium quantum dynamics, addressing how symmetry breaking (or conservation) affects the relaxation of local observables and the emergence of statistical mechanics.

2. Methodology

The author employs a combination of analytical random matrix theory and numerical simulations:

• Model Definition: The system is defined by $N \times N$ random SC matrices ($H$) where $[H, J] = 0$. The exchange matrix $J$ has eigenvalues $\pm 1$, splitting the Hilbert space into two decoupled subspaces (even and odd sectors). The energy spectrum is a superposition of two independent Gaussian Orthogonal Ensemble (GOE) spectra.
• Dynamical Analysis:
  • Initial States: The study considers initial states $|\Psi_\omega\rangle = \cos\omega |k\rangle + \sin\omega |k'\rangle$, where $|k\rangle$ and $|k'\rangle$ are exchange doublets. Specific cases include symmetric/antisymmetric states ($\omega = \pm \pi/4$) and superpositions of ground states from different sectors ($|\Psi_{o-e}\rangle$).
  • Survival Probability: The time evolution of the survival probability $R(t) = |\langle \Psi_\omega(t) | \Psi_\omega \rangle|^2$ is calculated analytically using the spectral form factor and level spacing distributions.
• Thermalization Analysis:
  • The expectation values of local observables $O$ are compared across three ensembles:
    1. Diagonal Ensemble (DE): The infinite-time average of the specific initial state.
    2. Micro-canonical Ensemble (ME): Average over energy shells.
    3. Gibbs Ensemble (GE): Standard thermal ensemble based on energy conservation only.
    4. Generalized Gibbs Ensemble (GGE): An ensemble incorporating both energy and the conserved symmetry charge ($J$).
  • Analytical expressions for energy correlations (level spacing, number variance) and survival probability timescales are derived.

3. Key Contributions and Results

A. Energy Correlations and Spectral Statistics

• The energy spectrum of a random SC matrix is a superposition of two GOE spectra.
• Level Spacing: The distribution $P(s)$ exhibits intermediate level repulsion between GOE (Wigner-Dyson) and Poisson statistics. The analytical form is derived as a superposition of two GOE distributions.
• Long-Range Correlations: The number variance $\Sigma^2(E)$ and the two-level form factor $b_2(\tau)$ are shown to follow specific scaling laws: $\Sigma^2_{SC}(E) = 2\Sigma^2_{GOE}(E/2)$ and $b^{SC}_2(\tau) = b^{GOE}_2(2\tau)$. This confirms the "mixed" nature of the spectrum.

B. Dynamical Timescales and Survival Probability

• Universal Decay: For generic initial states, the survival probability exhibits a quadratic decay ($1 - N t^2/4$) at short times (Zeno regime), followed by a power-law decay ($t^{-3}$) leading to a "correlation hole."
• Symmetry Dependence:
  • Symmetric ($\omega = \pi/4$) and antisymmetric ($\omega = -\pi/4$) states remain confined to their respective subspaces. They reach the correlation hole faster and have minimal spread in energy space.
  • Generic states ($\omega = 0$) couple both subspaces and spread over the full Hilbert space.
• Spontaneous Symmetry Breaking (Measure Zero):
  • The paper identifies a specific initial state $|\Psi_{o-e}\rangle$, a superposition of the ground states of the even and odd sectors.
  • Due to the clustering of the lowest energy levels in the two sectors, there is a non-zero probability (though vanishing as $N \to \infty$) that the energy gap between these states is extremely small.
  • In these rare cases, the system exhibits Rabi oscillations with a period $t_{Rabi} \gg t_H$ (Heisenberg time). The system retains memory of the initial configuration for timescales much longer than the relaxation time, effectively behaving as a symmetry-broken state. This occurs for a measure-zero fraction of the ensemble.

C. Violation of Standard Thermalization and Validity of GGE

• Failure of ETH: For local observables that commute with the symmetry operator $J$ (i.e., $[O, J] = 0$), the standard Gibbs ensemble fails. The diagonal ensemble average $\langle O \rangle_{DE}$ does not coincide with the micro-canonical or canonical averages.
• Success of GGE: The equilibrium expectation values are accurately captured by the Generalized Gibbs Ensemble (GGE).
  • The GGE density matrix is defined as $\hat{\rho}_{GGE} \propto e^{-(H - \mu J)/T}$, where $\mu$ is the Lagrange multiplier associated with the conserved charge $J$.
  • The Lagrange multipliers $T$ (temperature) and $\mu$ are determined by the initial energy $\langle E \rangle$ and the initial symmetry expectation value $\langle J \rangle = \sin(2\omega)$.
• Conclusion: The presence of the global $\mathbb{Z}_2$ symmetry acts as a constraint that prevents full thermalization to the standard Gibbs state. The system equilibrates to a state determined by both energy and the symmetry charge.

4. Significance

• Theoretical Insight: The paper provides a rigorous analytical framework for understanding how global symmetries in disordered systems modify thermalization. It demonstrates that even in chaotic (GOE-like) systems, global symmetries can lead to non-ergodic behavior for specific observables.
• Experimental Relevance: The results are relevant for systems exhibiting exchange symmetry, such as:
  • Quantum wires and networks designed for perfect state transfer.
  • Photosynthetic complexes where exciton transfer relies on symmetry.
  • Systems with parity or isospin mixing.
• Generalization: It reinforces the necessity of the Generalized Gibbs Ensemble in describing the equilibrium of systems with conserved quantities beyond just energy, extending the applicability of GGE beyond strictly integrable models to systems with specific global symmetries.
• Spontaneous Symmetry Breaking: The identification of a measure-zero set of matrices exhibiting long-lived symmetry-broken states offers a new perspective on pre-thermalization and metastability in random matrix models.

In summary, the paper establishes that while random SC matrices exhibit chaotic spectral statistics, the conservation of $\mathbb{Z}_2$ symmetry prevents standard thermalization for symmetry-respecting observables. The equilibrium state is correctly described not by the Gibbs ensemble, but by the Generalized Gibbs Ensemble, which accounts for the conserved exchange symmetry.