Generic ETH: Eigenstate Thermalization beyond the Microcanonical

This paper extends the Eigenstate Thermalization Hypothesis beyond its traditional microcanonical limits by demonstrating "generic ETH" in a qutrit lattice system with conserved quasilocal charge, where thermalization signatures persist even in states far outside standard energy and charge windows.

The Gist

The Big Picture: How Chaos Makes Things "Settle Down"

Imagine you have a jar full of marbles. If you shake the jar violently (adding energy), the marbles fly around chaotically. Eventually, if you stop shaking, the marbles settle into a specific, predictable pattern. In physics, this process is called thermalization. It's why your hot coffee cools down to room temperature, or why a gas spreads out evenly in a room.

For a long time, physicists believed that for a system to settle into this "thermal" state, it had to be in a very specific, narrow condition (like having a very precise amount of energy). This was the standard rulebook, known as the Eigenstate Thermalization Hypothesis (ETH).

The Big Discovery:
This paper says, "Wait a minute! Systems can settle down even if they are not in that narrow condition." They call this new phenomenon "Generic ETH." It turns out that even if a system is messy, spread out, and far from the "perfect" starting point, it can still behave like a thermal system.

---

The Cast of Characters

To understand how they proved this, let's look at the "toys" they used in their experiment:

1. The Qubit Spin Chain (The Simple Toy)

First, they looked at a simple line of coins (called qubits).

• The Analogy: Imagine a row of coins on a table. Some are Heads, some are Tails.
• The Experiment: They flipped the coins around using magnetic fields.
• The Result: When the coins interact in a chaotic way (like a messy game of pinball), they eventually settle into a state where you can't tell which coin was which. They become "thermal." This confirmed the old rules.

2. The Qutrit Spin Chain (The Complex Toy)

Next, they built a more complex toy called a qutrit.

• The Analogy: Instead of a coin with just two sides (Heads/Tails), imagine a coin with three sides: Heads, Tails, and a Third Side (let's call it "The Charge").
• The New Rule: In this new toy, the "Third Side" represents a special property called Charge. The total amount of "Charge" in the whole line must stay the same (conserved), but the charge can move from one coin to its neighbor.
• The Challenge: Usually, if you have a conserved rule like this, the system gets "stuck" and refuses to thermalize. It's like a traffic jam where cars can't move past each other.

3. The "Charge Spreader" (The Traffic Cop)

To fix the traffic jam, the authors added a special mechanism called the Charge Spreading Term.

• The Analogy: Imagine a traffic cop who forces cars to swap lanes. Even though the total number of cars on the highway stays the same, they can move around freely.
• The Result: This "cop" made the system chaotic again. The coins started mixing up, and the system began to thermalize within specific groups of charge.

---

The "Generic" Breakthrough

Here is the most exciting part of the paper.

The Old Way (Microcanonical):
Physicists usually study systems that start with a very precise amount of energy and charge. It's like trying to bake a cake with exactly 200 grams of flour and 100 grams of sugar. If you get it right, the cake turns out perfect (thermal).

The New Way (Generic ETH):
The authors asked: "What if we don't care about the exact amounts? What if we just throw in a random handful of flour and sugar, as long as the average is right?"

They created states where the "charge" was spread out wildly across the whole line of coins, not concentrated in a neat little box.

• The Surprise: Even though these starting states were messy and spread out (far from the "perfect" condition), they still thermalized!
• The Metaphor: Imagine you throw a handful of confetti into a wind tunnel. You might expect it to fly everywhere randomly. But instead, the wind (chaos) organizes it into a beautiful, predictable pattern, even though you didn't aim perfectly.

Why Does This Matter?

1. It's More Robust: We used to think thermalization was a fragile thing that only happened under perfect conditions. This paper shows it's actually very robust. It happens even when the system is "generic" (messy and spread out).
2. It Explains More: This helps us understand how real-world systems (like the early universe or complex materials) settle down, even if they started in very messy, non-ideal states.
3. The "Smoothness" Factor: The authors explain that because the relationship between energy/charge and temperature is "smooth" (like a gentle hill rather than a jagged cliff), the system doesn't care if you start slightly off-center. It just rolls down the hill to the same spot.

Summary in One Sentence

This paper proves that chaotic quantum systems are so good at "forgetting" their messy beginnings that they will settle into a predictable, thermal state even if you start them in a very messy, spread-out way, not just in the perfect, narrow conditions we thought were required.

Technical Summary

1. Problem Statement

The Eigenstate Thermalization Hypothesis (ETH) is the prevailing framework for explaining how isolated quantum systems thermalize. Standard ETH posits that for chaotic systems, the matrix elements of local observables in the energy eigenbasis take a specific form: diagonal elements are smooth functions of energy, and off-diagonal elements are exponentially suppressed. Consequently, states with narrow energy support (microcanonical windows) evolve to thermal equilibrium described by standard ensembles (e.g., Gibbs or microcanonical).

However, the paper addresses two key limitations in the current understanding of ETH:

1. Conserved Charges: How does ETH behave in systems with conserved charges (specifically quasilocal ones), and does it apply within individual charge sectors?
2. State Generality: Does thermalization occur only for states with narrow energy support, or can it occur for "generic" states with broad distributions in energy and charge? The authors investigate whether thermalization persists for states that do not fit the prototypical microcanonical window, a phenomenon they term "Generic ETH."

2. Methodology

The authors employ a combination of analytical arguments and numerical exact diagonalization on lattice spin models.

• Model 1: Qubit Spin Chain (Pedagogical Baseline)

  • They utilize a 1D transverse-field Ising model with parallel fields ($H_{qubit}$).
  • Purpose: To demonstrate the transition from integrability (Poisson level statistics) to chaos (Wigner-Dyson statistics) and verify standard ETH behavior. They analyze entanglement entropy growth and operator expectation values.
  • Key Insight: They show that random unentangled states, which naturally possess broad energy distributions mimicking a Gibbs state (rather than a microcanonical one), still thermalize to Gibbs predictions, not microcanonical ones.

• Model 2: Qutrit Spin Chain (Novel Contribution)

  • They design a new $L$-site qutrit Hamiltonian ($H$) embedding the chaotic qubit model into an enlarged Hilbert space ($3^L$).
  • Conserved Charge: The model includes a global conserved charge $Q$ (sum of local charges).
  • Charge Spreading: To ensure chaos within charge sectors, they introduce a "charge-spreading" term ($\Delta Q$) that allows local charge diffusion while preserving global conservation. This breaks local charge conservation but maintains global $U(1)$ symmetry.
  • Tunability: The model allows independent control over total energy, total charge, and local entanglement entropy.
  • Analysis: They analyze level-spacing statistics within fixed charge sectors to confirm chaos. They then study the dynamics of:
    1. Random Microcanonical States: Superpositions within a narrow energy and fixed charge window.
    2. Generic States: Completely unentangled product states with a fixed expectation value of charge ($\langle Q \rangle$) but broad distributions across both energy and charge sectors.

3. Key Contributions

• Definition of "Generic ETH": The authors coin the term "Generic ETH" to describe thermalization in states that are far outside the standard microcanonical window (broad support in energy and/or charge). They argue that if the diagonal matrix elements of observables are smooth functions of the conserved quantities, states with broad distributions will still thermalize, provided the fluctuations are averaged out or the observable is insensitive to higher moments.
• Generalized ETH in Qutrits: They demonstrate that in the presence of a conserved quasilocal charge, ETH holds within individual charge sectors (Generalized ETH).
• Ensemble Distinction: They rigorously distinguish between predictions made by the Microcanonical ensemble (fixed energy/charge) and the Gibbs/Generalized Gibbs ensemble (fixed average energy/charge). They show that for states with broad distributions, the Gibbs ensemble provides a more accurate thermal prediction than the microcanonical one.
• Improved Thermalization: Counter-intuitively, they find that "Generic ETH" states (broad support) often exhibit smaller temporal fluctuations and better thermalization in finite systems compared to strict microcanonical states, due to the averaging over a larger number of eigenstates.

4. Key Results

• Chaos and Level Statistics:

  • In the qubit model, the system transitions from Poisson (integrable) to Wigner-Dyson (chaotic) statistics as magnetic field parameters are tuned.
  • In the qutrit model, without the charge-spreading term, the spectrum is highly degenerate (integrable). With the charge-spreading term ($a=1$), individual charge sectors (except for the extreme edges) exhibit Wigner-Dyson statistics, confirming chaos within sectors.

• Thermalization within Charge Sectors (Generalized ETH):

  • For states with narrow support in both energy and charge (fixed $Q$), local observables and entanglement entropy converge to smooth functions of energy and charge, matching Generalized ETH predictions.
  • Off-diagonal matrix elements are exponentially suppressed, satisfying the ETH ansatz.

• Thermalization of Generic States (Generic ETH):

  • Broad Support: The authors prepared unentangled product states with a fixed average charge $\langle Q \rangle$. These states have a binomial distribution of charge across sectors and a broad energy distribution.
  • Convergence: Despite being far from a microcanonical window, these states thermalize. Their long-time averages match the predictions of a Generalized Gibbs Ensemble (GGE) defined by $\beta$ and $\mu$ (chemical potential), rather than a microcanonical ensemble.
  • Fluctuations: The temporal fluctuations of observables in these generic states are significantly smaller than in microcanonical states of the same system size. This is attributed to the "self-averaging" effect of the broad superposition of eigenstates.
  • Deviation from Microcanonical: For non-linear observables (e.g., two-site operators), the thermalized values of generic states deviate from microcanonical predictions. This is explained by the concavity of the smooth function describing the observable; the broad distribution samples the curvature, shifting the average value, whereas the microcanonical window is too narrow to detect this.

• Breakdown at Edges: Thermalization and ETH behavior break down in charge sectors with very small Hilbert space dimensions (e.g., $Q \approx L$ or $Q \approx 0$), where the density of states is too low to support the statistical assumptions of ETH.

5. Significance

• Broadening the Scope of ETH: The paper challenges the notion that thermalization requires strict confinement to a microcanonical window. It establishes that "Generic ETH" is a robust mechanism where thermalization occurs for a much wider class of initial states, provided the underlying eigenstate structure is smooth.
• Practical Implications: In experimental settings (e.g., cold atoms or quantum simulators), preparing states with exact microcanonical windows is difficult. This work suggests that states with broad distributions (which are easier to prepare) will still thermalize effectively, often with better stability (less fluctuation) than idealized microcanonical states.
• Conserved Charges: It provides a concrete numerical verification of Generalized ETH in systems with diffusive conserved charges, bridging the gap between integrable systems (Generalized Gibbs Ensembles) and fully chaotic systems.
• Finite-Size Effects: The results highlight that in finite systems, the choice of ensemble (Microcanonical vs. Gibbs) matters significantly for states with large fluctuations, and the "Generic ETH" behavior naturally selects the ensemble that matches the state's higher moments.

In summary, the authors successfully extend the Eigenstate Thermalization Hypothesis to demonstrate that thermalization is a generic feature of chaotic quantum systems, persisting even when initial states are far from the traditional microcanonical regime, provided the system possesses sufficient chaos and the observables are sufficiently local.