Thermalization with Gaussian Quantum Cellular Automata

This paper establishes conditions under which translation-invariant Gaussian quantum cellular automata drive locally normal many-body bosonic lattice states with bounded particle density toward thermalization at infinite temperature, utilizing a novel many-body generalization of the Riemann-Lebesgue lemma to bound local Weyl operator expectations.

The Gist

The Big Picture: Why Does Everything Heat Up?

Imagine you have a perfectly isolated room full of gas molecules. If you start with all the molecules in one corner (a very ordered state), physics tells us that eventually, they will spread out evenly to fill the whole room. This is thermalization: the process of a system losing its specific initial order and settling into a "hot," random, equilibrium state (often called the "infinite temperature" state in this context).

For decades, physicists have struggled to prove exactly when and why this happens in complex quantum systems. This paper takes a specific type of quantum system and proves that, under certain rules, it always thermalizes.

The Setup: A Quantum Grid of Springs

The authors study a grid (like a chessboard that goes on forever in every direction). On every square of this grid, there is a "bosonic mode."

• The Analogy: Think of each square as having a tiny, invisible spring attached to it. These springs can vibrate.
• The Rules: The system evolves in discrete time steps (like a video game updating frame by frame). The rules for how these springs move are governed by Gaussian Quantum Cellular Automata (GQCA).
  • Cellular Automata: The rule is local. The vibration of a spring at one spot only affects its neighbors within a fixed distance in the next step. Information cannot travel faster than a certain speed (like a wave spreading through a crowd).
  • Gaussian: The rules are "linear" and preserve the fundamental quantum relationships (like the balance between position and momentum).

The Goal: Proving the System "Forgets"

The researchers want to know: If we start with a specific, ordered pattern of vibrations (a specific "state"), will the system eventually look like a random mess where every local measurement gives a zero average?

They prove that if the system follows two specific conditions, the answer is yes. The system will "forget" its initial shape, and any local measurement will eventually read zero (which represents the random, thermal state).

The Two Magic Ingredients

To make the system thermalize, the authors identify two "recipes" (sets of conditions) that work.

Recipe 1: The "Everyday" Hyperbolic System

• The Concept: Imagine the grid has two types of directions for vibrations: "Stable" (directions that shrink or die out) and "Unstable" (directions that explode or grow).
• The Condition: The system is "Everyday" if no local pattern of vibrations sits entirely in the "Stable" direction. Every single local pattern you can make must have at least a tiny bit of "Unstable" energy in it.
• The Result: Because every pattern has some unstable energy, the system stretches that energy out exponentially fast. It's like pulling a piece of taffy; the more you pull, the thinner and more spread out it gets. Eventually, the "taffy" (the information about the initial state) is stretched so thin across the infinite grid that any local observer can't see it anymore. It has thermalized.

Recipe 2: The "Regular" Locally Hyperbolic System

• The Concept: Sometimes, the system isn't hyperbolic (stretching) everywhere, but it is hyperbolic in some specific regions or frequencies.
• The Condition: The system must be "Regular." This means no local pattern can just copy itself and move to a neighbor (like a "glider" in a game of Life) without changing its shape or growing.
• The Result: If a pattern tries to just slide along without growing, the "Regular" rule stops it. The system forces the pattern to eventually hit an "Unstable" region where it gets stretched out and diluted, just like in the first recipe.

The Secret Weapon: The Quantum Riemann-Lebesgue Lemma

How do they prove the stretching actually makes the system forget? They use a mathematical tool they call a "Many-Body Quantum Riemann-Lebesgue Lemma."

• The Classic Analogy: In regular math, the Riemann-Lebesgue lemma says that if you take a smooth wave and make its frequency go to infinity (wiggle it super fast), its average value over a region goes to zero.
• The Quantum Twist: In this paper, the "frequency" is the size of the vibration pattern (how much energy/momentum it has), and the "region" is the area the pattern covers.
• The Trade-off:
  1. The system stretches the pattern, making its "frequency" (energy) grow exponentially (very fast).
  2. But, because the rules are local, the pattern also spreads out, making its "size" (support) grow only polynomially (slowly, like a square or a cube).
• The Conclusion: The exponential growth of the energy wins the race against the slow growth of the size. The "wiggle" becomes so intense and spread out that the average value of any local measurement drops to zero. The system has thermalized.

Summary of Findings

The paper proves that for these specific quantum grids:

1. If the system stretches every local pattern (Recipe 1) OR if it prevents patterns from just sliding around without growing (Recipe 2)...
2. ...then the system will inevitably lose all memory of its starting point.
3. It will settle into a state where local measurements look completely random (thermalized).

The authors emphasize that this works for any starting state that isn't infinitely dense with particles, and it doesn't matter if the starting state was perfectly ordered or messy. As long as the "stretching" rules are in place, the system will eventually heat up and forget its past.

Technical Summary

Technical Summary: Thermalization with Gaussian Quantum Cellular Automata

Problem Statement
The paper addresses the emergence of thermodynamic behavior from the unitary evolution of quantum systems, specifically focusing on the "approach to equilibrium" or thermalization. While significant progress has been made in understanding thermalization for finite-dimensional systems (e.g., spin chains), a rigorous understanding for systems with infinite-dimensional degrees of freedom (bosonic lattice systems) in infinite volume remains elusive. The authors investigate whether translation-invariant Gaussian Quantum Cellular Automata (GQCAs) acting on such systems can drive a broad class of initial states toward the infinite-temperature state.

Methodology and Framework
The authors analyze discrete-time dynamics on a lattice $\Lambda = \mathbb{Z}^d$ with one bosonic mode per site. The framework relies on the algebraic approach to quantum statistical mechanics:

1. Algebraic Setting: The system is described by the local Weyl algebra $\mathcal{A}_{loc}$, generated by Weyl operators $W(\xi)$ labeled by finitely supported phase-space vectors $\xi \in V = c_{00}(\Lambda, \mathbb{R}^2)$.
2. Dynamics: The evolution is given by a Gaussian QCA, $\alpha$, which acts on Weyl operators as $\alpha(W(\xi)) = e^{i\chi(\xi)}W(\Phi\xi)$. Here, $\Phi$ is a translation-invariant, finite-spread symplectic map on the phase space labels.
3. Initial States: The study considers "locally normal" states with uniformly bounded particle density. This means the expectation value of the particle number operator $N_\Gamma$ in any finite region $\Gamma$ scales linearly with the volume $|\Gamma|$.
4. Thermalization Criterion: Thermalization is defined as the convergence of the expectation value of any non-trivial local Weyl operator to zero, i.e., $\lim_{k\to\infty} \omega(\alpha^k(W(\xi))) = 0$. This corresponds to convergence to the infinite-temperature Weyl state $\omega_\infty$.

Key Contributions and Intermediate Results
The central technical achievement is a many-body generalization of the Riemann-Lebesgue lemma (Theorem 1.5).

• The Lemma: For a locally normal state with finite particle density $\nu$, the expectation value of a Weyl operator $W(\xi)$ is bounded by:
  $$|\omega(W(\xi))| \leq \frac{\pi \sqrt{2\nu |\Gamma| + 1}}{\|\xi\|_2}$$
  where $\Gamma$ is the support of $\xi$ and $\|\xi\|_2$ is the $\ell^2$-norm of the phase-space label.
• Mechanism: This bound establishes a competition between two factors under time evolution $\xi_k = \Phi^k \xi$:
  1. Support Growth: Due to the finite spread of the QCA, the support size $|\Gamma_k|$ grows polynomially with time ($O(k^d)$).
  2. Norm Growth: Under hyperbolic dynamics, the phase-space norm $\|\xi_k\|_2$ grows exponentially.
     Thermalization occurs if the exponential growth of the norm dominates the polynomial growth of the support, causing the bound to vanish as $k \to \infty$.

Main Results
The authors formulate two sets of sufficient conditions on the GQCA $\Phi$ to guarantee thermalization for any locally normal state with finite particle density:

1. Uniform Hyperbolicity + Everydayness (Theorem 1.1):

   • Uniform Hyperbolicity: The spectrum of the symbol $A(\theta)$ (the Fourier transform of $\Phi$) is uniformly separated from the unit circle for all $\theta$ in the Brillouin torus. This ensures a stable/unstable splitting of the Hilbert phase space.
   • Everydayness: No non-zero strictly local label lies entirely within the stable subspace ($V \cap P_- = \{0\}$). This ensures that every local observable has an unstable component that expands exponentially.
   • Result: Under these conditions, $\|\Phi^k \xi\|_2$ grows exponentially for all non-zero local $\xi$, leading to thermalization.

2. Local Hyperbolicity + Regularity (Theorem 1.2):

   • Local Hyperbolicity: Hyperbolicity is visible only on an open subset of the Brillouin torus (i.e., $A(\theta_0)$ is hyperbolic for some $\theta_0$).
   • Regularity: No non-zero local label is mapped by $\Phi$ to a scalar multiple of a lattice translate of itself (excluding "gliders" or eigen-labels that avoid expansion).
   • Result: These conditions are sufficient to ensure that the norm of any local label grows exponentially, even if hyperbolicity is not global.

Significance and Scope
The paper claims to provide a rigorous proof of thermalization for a specific class of infinite-dimensional quantum lattice systems.

• Generality of Initial States: The results apply to a broad class of initial states that need not be translation-invariant, quasi-free, or Gibbs states; they only require local normality and bounded particle density.
• Limitations: The convergence is established strictly at the level of local Weyl observables (finite linear combinations of Weyl operators). The authors explicitly note that the theorems do not imply thermalization for arbitrary bounded operators in the full algebra or arbitrary unbounded local observables (though the number operator enters via the density condition).
• Connection to Classical Dynamics: The work utilizes the correspondence between GQCAs and classical symplectic cellular automata on phase space. The authors highlight that while classical results on non-compact phase spaces are rare, their approach successfully transfers these dynamical properties to the quantum many-body setting.

The paper concludes by noting that the "everydayness" condition is algebraic and checkable, and it identifies the presence of eigenvalues on the unit circle as a potential obstruction to thermalization, suggesting a link to localization phenomena.