Liouvillian Gap in Dissipative Haar-Doped Clifford Circuits

This paper investigates how infinitesimal bulk dissipation induces intrinsic relaxation in Floquet Clifford circuits doped with Haar-random gates, revealing that while subextensive doping preserves a system-size-dependent Liouvillian gap, specific finite-density doping patterns can generate a constant $\mathcal{O}(1)$ gap that remains singular as dissipation vanishes, thereby establishing a circuit-level mechanism for irreversibility in open many-body systems.

The Gist

Imagine you have a giant, chaotic dance floor with thousands of dancers (these are the qubits in a quantum computer). In a perfectly chaotic system, if you whisper a secret to one dancer, that secret spreads instantly to everyone else. This is called scrambling, and it's the hallmark of quantum chaos.

Now, imagine the room isn't perfectly sealed. There are tiny leaks in the walls letting in a gentle breeze (this is dissipation or noise). In a normal chaotic room, this breeze would eventually blow everyone's secrets away, and the room would settle into a calm, uniform state. The speed at which this happens is called the Liouvillian gap.

Here is the big puzzle the paper solves: How much "chaos" do you need to make the room settle down quickly, even when the breeze is incredibly weak?

The Setup: The "Clifford" Dance Floor

The authors start with a very specific type of dance floor called a Clifford circuit.

• The Analogy: Think of this as a dance floor where the choreography is rigid and predictable. It's like a game of "Simon Says" where the moves are strictly defined.
• The Problem: Even though this dance looks orderly (and is actually easy for a computer to simulate), the authors found something surprising: if you introduce even a tiny breeze (dissipation), the secrets spread so fast that the whole room settles down instantly as the room gets bigger. The "gap" (relaxation speed) grows with the size of the room. It's like the breeze becomes a hurricane simply because the room is huge.

The Experiment: Adding "Haar" Doping

The researchers asked: What if we break the rigid rules just a little bit?
They introduced Haar-doping.

• The Analogy: Imagine that every now and then, instead of following the strict "Simon Says" choreography, one dancer is allowed to spin completely randomly (a Haar-random gate). They call this "doping" because it's like adding a pinch of spice to a bland dish.

They wanted to know: How much random spice do we need to stop the "hurricane" effect and make the relaxation speed stay constant, no matter how big the room gets?

The Findings: The "Traffic Jam" vs. The "Highway"

The paper discovers a fascinating crossover based on where you put the random dancers.

1. The Undoped Case (The Highway)

If you have zero random dancers, the rigid choreography acts like a perfect highway. A secret (a Pauli string) travels down the line, gets bigger and bigger, and eventually covers the whole room. Because it covers the whole room, the tiny breeze hits it everywhere at once, causing a massive relaxation rate.

• Result: The relaxation speed explodes as the room gets bigger.

2. The Fully Doped Case (The Traffic Jam)

If every dancer is random, the secrets get mixed up so thoroughly that they never grow large enough to cover the whole room. They stay small and local.

• Result: The relaxation speed is constant and finite, regardless of room size.

3. The "Just Right" Zone (The Block-Staggered Pattern)

This is the most exciting part. The authors found you don't need everyone to be random.

• The Analogy: Imagine a long line of dancers. If you have a block of 100 rigid dancers followed by 1 random dancer, followed by another 100 rigid dancers... the secret travels fast through the rigid block, but when it hits the random dancer, it gets "scrambled" and shrinks back down before it can grow too big.
• The Discovery: As long as you have a non-zero density of random dancers (meaning the ratio of random to total dancers doesn't vanish as the room gets huge), you can create "return cycles."
  • Think of these cycles as a conveyor belt. The secret travels a short distance, gets scrambled, and comes back to where it started, but it never gets big enough to trigger the "hurricane" effect.
  • Even if the random dancers are sparse (e.g., one random dancer for every 10 rigid ones), the system still settles down at a constant, finite speed.

The "Liouvillian Gap" in Plain English

The Liouvillian Gap is essentially the speed limit of forgetting.

• Large Gap: The system forgets its past very fast (strong chaos).
• Small Gap: The system remembers its past for a long time (integrable or ordered).

The paper shows that in these specific quantum circuits:

1. Purely Ordered (Clifford): The "speed limit" is actually infinite (it forgets instantly) because the system is so big.
2. Mixed (Doped): By adding just enough randomness in the right pattern, you can "tame" the system. You create a scenario where the system forgets at a steady, manageable pace, even as the system grows to infinity.

Why Does This Matter?

Usually, we think "Chaos = Randomness." But this paper shows that Chaos = Irreversibility (the ability to forget).

• You can have a system that looks very ordered (Clifford) but is actually more irreversible (faster relaxation) than a system with some randomness.
• However, by adding a specific "doping" of randomness, you can tune the system to have a stable, finite relaxation rate.

The Takeaway Metaphor

Imagine a river (the quantum system).

• Undoped: The river is a straight, smooth canal. If you drop a leaf in, it shoots downstream at lightning speed, hitting the ocean instantly. The "relaxation" is too fast to measure.
• Fully Doped: The river is a swamp with random whirlpools everywhere. The leaf spins in place and sinks slowly.
• Doped (The Paper's Discovery): You build a series of small, random whirlpools spaced out along the river. The leaf travels fast between them, but every time it hits a whirlpool, it gets slowed down and reset. The result? The leaf drifts downstream at a steady, constant speed, no matter how long the river is.

The paper proves that you don't need a swamp to get a steady drift; you just need the right pattern of whirlpools. This gives us a new way to understand how quantum systems lose their information and become irreversible.

Technical Summary

1. Problem Statement and Motivation

The paper addresses the nature of quantum chaos and irreversibility in open many-body systems. While classical chaos is defined by trajectory sensitivity, quantum chaos is typically diagnosed via spectral statistics, Out-of-Time-Order Correlators (OTOCs), or entanglement growth. However, these probes can yield conflicting results.

A recent proposal suggests a dissipative signature for chaos in Floquet systems: the behavior of the Liouvillian gap ($\Delta$) in the limit of weak dissipation ($\gamma \to 0^+$) after taking the thermodynamic limit ($N \to \infty$).

• Chaotic systems: Exhibit a finite intrinsic relaxation rate (finite $\Delta$) even as $\gamma \to 0^+$, because rapid operator spreading makes local dissipation act on an extensive support.
• Integrable/Non-chaotic systems: The gap typically closes as $\gamma \to 0$ or scales differently.

The specific puzzle this paper tackles is the behavior of Clifford circuits. Clifford circuits are classically simulable and lack random-matrix spectral statistics (often a hallmark of chaos), yet they exhibit maximal operator spreading. The authors ask: What minimal departure from pure Clifford dynamics (via "doping" with non-Clifford gates) is required to generate a finite intrinsic relaxation rate (a finite Liouvillian gap) in the thermodynamic limit?

2. Model and Methodology

The authors study a Dissipative Haar-Doped Floquet Circuit (DHFC):

• Architecture: A 1D periodic chain of $N$ qubits.
• Unitary Dynamics: A two-layer "brickwork" circuit built from a fixed iSWAP-class two-qubit Clifford gate ($u = iSWAP(H \otimes H)$).
• Doping: Single-qubit Haar-random unitary gates are inserted immediately after each Clifford layer on a specific subset of $n_h$ qubit lines.
• Dissipation: A local depolarizing channel with strength $\gamma$ acts on all qubits at the end of each Floquet period.
• Diagnostic: The Liouvillian gap $\Delta = -\max_{a \neq 0} \text{Re}(\lambda_a)$, where $\lambda_a$ are eigenvalues of the Floquet channel.

Analytical Framework: Weight Truncation
To analyze the strong dissipation regime ($\gamma \gg 1$), the authors employ a weight-truncation method.

• The depolarizing channel suppresses Pauli strings exponentially with their weight ($w(S)$).
• The dynamics are projected onto a subspace of Pauli strings with weight $w \leq w_t$.
• Lower Bound Strategy: If the truncated dynamics are nilpotent (all low-weight operators eventually grow beyond the cutoff and are removed), the gap must be large (proportional to the cutoff weight).
• Upper Bound Strategy: If there exist return cycles (operators that return to themselves within the truncated subspace), the gap is bounded by the decay rate of these cycles.

3. Key Contributions and Results

A. The Undoped Clifford Limit (Baseline)

• Result: For a pure iSWAP-class Clifford circuit ($n_h = 0$), the Liouvillian gap scales linearly with system size: $\Delta \propto N\gamma$.
• Mechanism: The iSWAP class ensures that Pauli strings spread ballistically without getting "stuck" in low-weight orbits. In the thermodynamic limit, typical operators become extensive ($w \sim N$), so local dissipation damps them at a rate $\sim N\gamma$.
• Significance: This is counter-intuitive because Clifford circuits are not "chaotic" in the random-matrix sense, yet they exhibit the strongest possible intrinsic relaxation (fastest decay) under this dissipative diagnostic.

B. Weak-Dissipation Singularity with Doping

• Numerical Findings: Introducing Haar doping ($n_h > 0$) preserves the $N$-scaling of the gap for small system sizes but leads to a crossover.
• Singularity: The paper confirms that for any non-zero doping density, the gap exhibits a singularity where $\lim_{N \to \infty} \lim_{\gamma \to 0} \Delta \neq \lim_{\gamma \to 0} \lim_{N \to \infty} \Delta$. The system retains chaotic-like relaxation properties.

C. Necessary Condition for a Finite Gap (Theorem 2)

• Main Theorem: For the Liouvillian gap to remain finite (independent of $N$) in the thermodynamic limit, the number of doped sites $n_h$ must scale linearly with $N$ (i.e., doping density $p_h = n_h/N$ must be non-zero).
• Sublinear Doping: If $n_h$ grows sublinearly (e.g., $n_h \sim \sqrt{N}$), the gap diverges as $N \to \infty$ ($\Delta \to \infty$).
• Proof: Based on the "pigeonhole principle," if doping is sparse, there exist long contiguous blocks of undoped sites. In these blocks, the deterministic Clifford dynamics drive Pauli weight to grow extensively, forcing the gap to scale with the block length.

D. Sufficient Conditions and Upper Bounds (Structured Doping)

The authors derive explicit upper bounds for specific doping patterns in the strong dissipation regime:

1. Fully Doped ($n_h = N$): $\Delta = \gamma + 2$. The gap is finite and independent of $N$.
2. Dense Doping (No adjacent undoped sites): $\Delta \leq 3\gamma + 3$. The gap remains finite.
3. Block-Staggered Doping: Doped sites separated by $k$ undoped sites.
   • The gap is bounded by $\Delta \leq A\gamma + B$, where $A, B$ depend on $k$ but not on $N$.
   • This implies that even with sparse doping (as long as the spacing $k$ is finite), the system exhibits a finite intrinsic relaxation rate.

E. Mechanism of Finite Gap

In doped circuits, the random Haar rotations mix Pauli components, preventing the deterministic Clifford growth from pushing all operators to high weights. Instead, localized return cycles form within small contiguous blocks. These cycles involve operators that translate along the chain but return to their original configuration (up to translation) after $N/2$ steps, staying within a low-weight sector. These cycles dominate the slowest decay modes, fixing the gap to an $O(1)$ value.

4. Significance and Implications

• Redefining Chaos via Dissipation: The paper demonstrates that the Liouvillian gap is a sensitive probe of irreversibility that can distinguish between "integrable-like" (finite gap) and "chaotic-like" (divergent gap) behaviors in open systems, even when unitary diagnostics (like spectral statistics) might be ambiguous for Clifford circuits.
• Role of Doping Density: It establishes a sharp threshold: linear doping density is required to transition from a system with diverging relaxation rates (undoped Clifford) to one with finite intrinsic relaxation (doped/chaotic).
• Universality of Bounds: The derived bounds depend only on the spatial pattern of doping, not on whether the Haar rotations are fixed (quenched) or resampled every period. This makes the results robust for experimental implementations.
• Connection to Scrambling: The doping scale required for a finite Liouvillian gap (linear in $N$) aligns with the scale required for scrambling diagnostics (like OTOCs) to become Haar-like, suggesting a unified picture of quantum chaos where relaxation and scrambling emerge at similar parameter scales.

Conclusion

The authors resolve the tension between the classical simulability of Clifford circuits and their rapid relaxation in open systems. They show that while undoped iSWAP circuits exhibit maximal relaxation (divergent gap), the introduction of non-Clifford Haar doping at a linear density creates a "finite-gap" phase characterized by localized return cycles. This provides a circuit-level understanding of how intrinsic irreversibility emerges in open quantum many-body systems.