Diffusive Dynamics of Nonstabilizerness

Imagine you are trying to bake the most complex, chaotic cake possible in a kitchen. In the world of quantum computers, this "cake" is a special kind of state called a Haar-random state. To make a truly useful quantum computer, you need to bake this cake because it represents the ultimate level of complexity and unpredictability.

However, there's a catch. You can't just throw ingredients together randomly; you have to follow specific rules, like keeping the total number of eggs (a "conserved charge") exactly the same throughout the process. This is what physicists call a symmetry constraint.

This paper, titled "Diffusive Dynamics of Nonstabilizerness," investigates how long it takes to bake this complex cake when you are forced to follow these rules.

The Ingredients: What is "Nonstabilizerness"?

To understand the paper, we need two main ingredients:

Entanglement: Think of this as the "glue" that holds the cake together. It's a well-known quantum resource where parts of the system are deeply connected.
Nonstabilizerness (or "Magic"): This is the paper's main focus. Imagine a standard cake recipe (a "stabilizer state") that a simple, classical computer can easily copy and understand. To make a quantum cake that a classical computer cannot copy, you need to add a secret ingredient called "Magic" (or nonstabilizerness). Without this "Magic," the quantum computer isn't actually doing anything a regular computer couldn't do.

The authors are asking: If we are forced to keep our "egg count" (charge) constant while baking, how does the "Magic" spread through the cake, and how long does it take to reach the perfect, chaotic state?

The Experiment: A Random Kitchen

The researchers simulated a one-dimensional line of quantum bits (qubits) acting like a kitchen line. They applied random "gates" (mixing actions) to pairs of neighbors.

The Rule: Every time they mixed, they had to ensure the total "charge" (like the number of eggs) stayed the same.
The Measurement: They tracked the "Stabilizer Rényi Entropy," which is a fancy way of measuring how much "Magic" is in the system.

The Discovery: The "Diffusive" Spread

The team found that the "Magic" doesn't appear instantly. Instead, it spreads slowly, like a drop of food coloring diffusing through a glass of water.

The Slow Motion: Because the system has to conserve its charge, the "Magic" is dragged down by the slow movement of that charge. The charge moves like a crowd of people shuffling through a hallway; it takes time to get from one end to the other.
The Math of the Wait: The researchers discovered a specific rule for how fast the "Magic" approaches its final, perfect value.
- In the beginning, the gap between the current "Magic" level and the perfect level shrinks slowly.
- Specifically, this gap closes at a rate of 1 over time ( $1/t$ ).
- The Analogy: Imagine you are waiting for a pot of water to boil. If you have no constraints, it boils fast. But if you have to keep adding ice to keep the temperature steady (the symmetry constraint), the water takes much longer to reach the boiling point. The paper shows that this "waiting time" follows a predictable, slow pattern.

The "Thouless Time" Limit

The paper also looked at what happens in a kitchen of a specific, finite size (not an infinite line).

The Diffusive Window: For a while, the "Magic" spreads slowly and predictably (the $1/t$ rule).
The Crossover: Eventually, the "Magic" reaches the very end of the line. Once it hits the wall, the slow diffusion stops, and the system snaps to its final state very quickly (exponentially fast).
The time it takes to hit this wall is called the Thouless time. The paper found that this time gets longer if the kitchen is bigger, growing with the square of the size ( $N^2$ ).

Why This Matters (According to the Paper)

The authors used a powerful computer simulation method (called iTEBD) that allowed them to look at the system as if it were infinitely large, which is usually impossible to do.

They proved that symmetry creates a "traffic jam" for quantum complexity. Even in a chaotic system, if you have a conserved charge, the generation of "Magic" is forced to move at a diffusive speed. This identifies a new "universality class"—a category of behavior that applies not just to their random circuit, but also to a specific type of magnetic chain (the Ising chain) they tested.

Summary in a Nutshell

The Problem: How does quantum "Magic" (complexity) grow when you are forced to keep a specific quantity (charge) constant?
The Method: They simulated random quantum circuits with a conservation law and measured the "Magic" using a new, efficient mathematical trick involving four copies of the system.
The Result: The "Magic" spreads slowly, like a drop of dye in water. The time it takes to reach the final state follows a $1/t$ rule, controlled by how fast the conserved charge can diffuse.
The Conclusion: Symmetry and conservation laws act as a speed limit for generating quantum complexity, forcing it to follow a diffusive path rather than a ballistic (fast) one.

Technical Summary: Diffusive Dynamics of Nonstabilizerness

Problem Statement
Symmetries and conservation laws are fundamental organizing principles in many-body quantum dynamics, governing the relaxation of entanglement and operator spreading. However, their influence on nonstabilizerness (or "quantum magic")—the resource complementary to entanglement required for quantum advantage—remains poorly understood. While entanglement entropy in symmetry-constrained random circuits is known to grow sub-ballistically ( $\sqrt{t}$ ) due to diffusive hydrodynamics, the dynamical universality class for nonstabilizerness is unclear. Existing measures of nonstabilizerness are often computationally intractable, scaling exponentially with system size $N$ , and direct matrix-product-state (MPS) simulations become inefficient once entanglement reaches the volume law. This paper addresses the question: How does a conserved charge (specifically U(1) symmetry) shape the late-time dynamics of nonstabilizerness generation in chaotic many-body systems?

Methodology
The authors investigate a one-dimensional U(1)-symmetric random circuit on $N$ qubits, where two-qubit gates are drawn from the Haar measure restricted to commute with the total $Z$ -charge. To access the thermodynamic limit ( $N \to \infty$ ) and late-time behavior, they employ a novel computational framework:

Replica Tensor Network: They utilize the stabilizer Rényi entropy ( $M_\alpha$ ) with $\alpha=2$ , which serves as a tractable measure of nonstabilizerness. The disorder-averaged stabilizer purity $\zeta_2(t)$ is expressed as a four-replica tensor network: $\zeta_2(t) = \text{Tr}(Q \rho_t^{\otimes 4})$ , where $Q$ is a specific permutation operator.
Symmetry-Adapted iTEBD: The ensemble average restores translation invariance, allowing the use of infinite time-evolving block decimation (iTEBD) directly in the thermodynamic limit. Crucially, the authors exploit the $S_4$ permutation symmetry of the four replicas. By decomposing the local Hilbert space (reduced from dimension 256 to 70 by charge conservation) into irreducible representations of $S_4$ , they organize the tensor network into block-diagonal multiplets. This non-Abelian symmetry adaptation significantly reduces the computational cost of the non-unitary iTEBD update, enabling convergence with a bond dimension $\chi_{\max} \approx 800$ .
Hydrodynamic Argument: Theoretical scaling is derived by analyzing the relaxation of the conserved U(1) charge density. The authors argue that the slow hydrodynamic mode (diffusive charge) dominates the late-time approach to the Haar-random value, while transverse operators relax exponentially.

Key Results
The study identifies a diffusive universality class for the late-time approach of nonstabilizerness to its random-state plateau:

Thermodynamic Limit Scaling: In the thermodynamic limit, the gap between the stabilizer Rényi entropy density $m_2(t)$ and its Haar-random value ( $m_{\text{Haar}} = 1$ ) closes algebraically as:
$\Delta m_2(t) \equiv 1 - m_2(t) \propto t^{-1}$
This $t^{-1}$ scaling is directly traced to the diffusive relaxation of the conserved charge density, where the variance of the charge fluctuations decays as $t^{-1/2}$ , leading to a $t^{-1}$ contribution in the fourth moment of the Pauli string distribution.
Finite-Size Regimes: Direct simulations of finite chains ( $N \sim 20$ $N \sim 20$ ) using Pauli-string sampling reveal two distinct regimes separated by the Thouless time $t_D \sim N^2/D$ $t_{D} \sim N^{2} / D$ (where $D$ $D$ is the diffusion constant):
1. Diffusive Window ( $\tau \ll t \ll t_D$ ): The gap follows the power law $\Delta M_2(t) \propto t^{-1}$ .
2. Post-Diffusive Regime ( $t \gg t_D$ ): Once the conserved mode becomes spatially uniform, the gap decays exponentially.
Hierarchy of Rényi Gaps: For higher-order stabilizer Rényi entropies ( $M_q$ ), the gaps within the diffusive window follow the hierarchy $\Delta M_q \propto t^{-q/2}$ .
Universality Across Models: The same $t^{-1}$ scaling is verified in an energy-conserving non-integrable mixed-field Ising chain. Here, the local energy density acts as the slow hydrodynamic mode, confirming that the diffusive universality class applies broadly to chaotic systems with a slow conserved density, regardless of whether the symmetry is U(1) charge or energy.

Significance and Claims
The paper claims to establish hydrodynamics as a universal organizing principle for late-time magic dynamics in symmetry-constrained chaotic systems. By combining a symmetry-adapted tensor network method with hydrodynamic arguments, the work provides a direct link between conserved quantities and the generation of nonstabilizerness.

The authors position this work as a complement to the established picture of entanglement entropy and operator spreading. The findings offer practical insights for the design of approximate Haar-random states in Hamiltonian dynamics and quantum simulation protocols that operate within symmetry-constrained Hilbert spaces (e.g., fermionic simulations or gauge theories). The paper modestly notes that while it identifies the late-time diffusive scaling, the early-time growth of nonstabilizerness and the behavior in other integrable models or under different diagnostics (e.g., robustness of magic) remain open questions for future study.