🔬 optics

Non-diffusion transport in decoherent non-Hermitian quasicrystals

This paper demonstrates that, contrary to the conventional Anderson paradigm where decoherence suppresses localization and restores diffusion, non-Hermitian quasicrystals exhibit robust non-diffusive transport phenomena—including dissipation-induced localization and decoherence-induced mobility edges—even in the fully incoherent limit.

Original authors: Yudong Ren, Rui Zhao, Kangpeng Ye, Lu Zhang, Hongsheng Chen, Haoran Xue, Yihao Yang

Published 2026-04-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Yudong Ren, Rui Zhao, Kangpeng Ye, Lu Zhang, Hongsheng Chen, Haoran Xue, Yihao Yang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Idea: When "Noise" Actually Helps You Stay Put

Imagine you are trying to walk through a crowded, chaotic market.

The Old Rule (Hermitian Physics): For decades, scientists believed that if the market gets too noisy and chaotic (decoherence), people stop bumping into each other in a coordinated way. Instead of getting stuck in a specific spot, everyone just wanders aimlessly in all directions until they spread out evenly. In physics terms: Noise kills "localization" (staying put) and forces "diffusion" (spreading out).
The New Discovery (Non-Hermitian Physics): This paper shows that in a very specific type of system (one with "loss" or "dissipation"), the old rule is wrong. Surprisingly, adding noise can actually make things stay put. In fact, if you add enough noise, you can force a system to stop spreading and lock itself into a specific spot.

The researchers built a "traffic jam" for light and proved that in this new world, chaos doesn't just scatter things; it can organize them into a stuck position.

The Analogy: The Two Types of Walkers

To understand the difference, let's imagine two types of people walking through a maze:

1. The "Perfectly Coordinated" Walker (Hermitian System)

Imagine a group of dancers who must move in perfect sync.

How they move: They rely on a complex dance routine where they step in time with each other. If they do this perfectly, they can get "stuck" in a corner of the room because their steps cancel out the movement in other directions. This is Anderson Localization.
What happens when it gets noisy: If you start playing loud, random music (decoherence), the dancers can't hear the rhythm anymore. They stop dancing in sync. They start bumping into walls and wandering randomly. They spread out across the whole room.
The Lesson: In the old world, Noise = Spreading Out.

2. The "One-Way Street" Walker (Non-Hermitian System)

Now, imagine a different group of walkers in a maze where the floor is slippery and has a strong wind blowing in one direction (this is the Non-Hermitian part, representing energy loss or gain).

How they move: Even if they try to walk in a circle, the wind pushes them toward a specific corner. They naturally drift toward a "sink."
What happens when it gets noisy: Here is the magic. When you add the loud, random music (noise) to this windy maze, something weird happens. The noise scrambles their individual steps, but because the wind is so strong, the noise actually helps them settle down faster into that corner. Instead of wandering aimlessly, the noise helps them "lock in" to the spot where the wind pushes them.
The Lesson: In this new world, Noise + Wind = Getting Stuck.

How They Did It: The Light Loop

The scientists didn't use dancers or wind; they used light and fiber optic cables.

The Setup: They created a "synthetic maze" using two loops of optical fiber. One loop is short, and one is long. Light pulses bounce back and forth between them.
The "Maze": They programmed the loops to act like a quasicrystal (a pattern that repeats but never quite matches up, like a Penrose tiling).
The "Wind": They added a special "loss" mechanism (dissipation) to one side of the loop. This makes the system "Non-Hermitian."
The "Noise": They used a modulator to randomly scramble the phase (the timing) of the light pulses. This simulates "decoherence."

The Results: What They Found

They tested what happens when they turn the "noise" dial from zero (perfect silence) to maximum (loud chaos).

In the Old World (Hermitian): When they turned up the noise, the light spread out like ink in water. The "stuck" state disappeared completely.
In the New World (Non-Hermitian):
- Low Noise: The light spreads out, but in a weird, two-step way (it zooms fast, then slows down).
- High Noise: Even with maximum noise, the light did not spread out. Instead, it stayed tightly packed in one spot!
- The "Mobility Edge": They found a sweet spot where adding more noise actually created a stuck state. It's like if you added more traffic to a highway, and instead of a traffic jam, everyone suddenly parked perfectly in a single spot.

Why Does This Matter?

This changes how we think about the universe:

Chaos is a Tool: We usually think of noise (static, interference, loss) as a bad thing that ruins precision. This paper shows that in certain systems, noise is a control knob. You can use it to trap energy or light exactly where you want it.
New Materials: This could lead to new types of lasers, sensors, or computers that work better in noisy environments, rather than failing.
Beyond the Rules: It proves that the "Anderson Localization" rule (which has been a pillar of physics for 60 years) isn't universal. It only applies to "perfect" systems. Once you introduce real-world loss and noise, the rules change completely.

The Takeaway

In a perfect, quiet world, noise makes things scatter.
In a "leaky" (non-Hermitian) world, noise can actually make things stick together.

The researchers discovered a new kind of physics where chaos doesn't destroy order; it creates a different kind of order that is robust against the very noise that usually destroys it.

1. Problem Statement

The paper addresses a fundamental question in wave physics: Does the loss of phase coherence inevitably restore diffusive transport in disordered systems?

The Paradigm: In traditional Hermitian systems (e.g., electrons, photons in lossless media), Anderson localization relies on phase-coherent interference. The established intuition is that decoherence (dephasing) destroys this interference, suppressing localization and forcing the system back into a featureless diffusive state.
The Gap: It remains unclear whether this intuition holds in non-Hermitian systems, where transport is governed not only by interference but also by dissipation, gain/loss, and mode competition. Specifically, can decoherence cooperate with non-Hermitian effects to create new non-diffusive phases, or does it still universally drive the system toward diffusion?

2. Methodology

The authors employed a combined theoretical and experimental approach using a programmable photonic lattice realized via coupled optical fiber loops.

Experimental Platform:
- Synthetic Dimensions: Two optical fiber loops of unequal lengths create a 1+1D mesh lattice. The loop length difference creates a spatial dimension (synthetic), while the round-trip time creates the temporal dimension.
- Non-Hermitian Control: A Mach-Zehnder Modulator (MZM) imposes a site-dependent, incommensurate gain-and-loss profile $g(x) = \lambda \cos[2\pi\alpha(x+1)]$ , realizing an imaginary quasiperiodic potential (non-Hermitian Aubry–André model).
- Decoherence Control: A Phase Modulator (PM) introduces random phase shifts $\phi(x,t)$ that fluctuate in both space and time. The strength of this dephasing ( $W$ ) is tunable from $0$ (fully coherent) to $2\pi$ (fully incoherent).
- Independent Tuning: The setup allows independent real-time control over the coupling strength ( $\theta$ ), the non-Hermitian potential amplitude ( $\lambda$ ), and the dephasing strength ( $W$ ).
Theoretical Framework:
- The authors developed a unified theoretical framework based on a hierarchy of correlation functions.
- They derived a recursive equation for the two-point correlation tensor that describes the exact ensemble-averaged dynamics.
- This framework bridges the gap between the unitary evolution of pure states (coherent limit) and classical random walks (incoherent limit) via a coherence retention factor $\gamma(W)$ .

3. Key Contributions

Experimental Realization: First experimental demonstration of a non-Hermitian quasicrystal with fully programmable dephasing, spanning the entire coherence landscape.
Unified Theory: Development of a theoretical model that analytically captures ensemble-averaged dynamics across all dephasing strengths, connecting coherent quantum walks to incoherent classical diffusion.
Paradigm Shift: Demonstration that the "decoherence $\to$ diffusion" rule is not universal and breaks down in non-Hermitian systems.

4. Key Results

A. Failure of the Hermitian Paradigm

In Hermitian quasicrystals, the introduction of even weak dephasing destroys the Anderson localization transition. The system transitions from a localized phase (at low coupling) to a purely diffusive phase regardless of coupling strength once dephasing is present.

B. Non-Diffusive Transport in Decoherent Non-Hermitian Systems

Contrary to Hermitian systems, decoherent non-Hermitian quasicrystals retain structured, non-diffusive transport even in the fully incoherent limit ( $W=2\pi$ ). The authors identified three distinct phenomena:

Diffusion-Localization Transitions:
- Even with zero phase coherence, the non-Hermitian system exhibits a sharp transition from diffusive to localized transport as the coupling strength $\theta$ is varied.
- This localization is driven by the competition between diffraction and non-reciprocal energy flow induced by the imaginary potential, rather than interference.
Decoherence-Induced Mobility Edges:
- By fixing the coupling strength and increasing dephasing ( $W$ ), the system undergoes a transition from an extended phase to a localized phase.
- Dephasing acts as a control parameter that sweeps a mobility edge across the spectrum, progressively localizing eigenstates. At full dephasing, all eigenstates become localized, a phenomenon impossible in Hermitian systems where dephasing only delocalizes states.
Distinct Transport Mechanisms:
- Coherent Non-Hermitian: Transport shows a two-step evolution (ballistic $\to$ diffusive) due to mode competition.
- Incoherent Non-Hermitian: Transport is purely diffusive initially but can transition to localization. The mechanism is a memoryless Markov process governed by on-site densities, yet it retains the ability to localize due to the non-Hermitian potential.

C. Phase Diagrams

The constructed phase diagrams in the $(\theta, W)$ plane reveal a robust phase boundary for non-Hermitian systems that persists up to maximum dephasing ( $W=2\pi$ ).
In contrast, the Hermitian phase boundary vanishes immediately upon introducing any dephasing ( $W > 0$ ).

5. Significance

Redefining Decoherence: The work challenges the conventional view of decoherence as merely a destructive force that restores diffusion. Instead, it establishes decoherence as an independent dimension of control that can be harnessed to stabilize localized phases in open systems.
Beyond Anderson: It demonstrates that non-Hermitian physics offers transport mechanisms fundamentally distinct from the Anderson paradigm, where dissipation and disorder cooperate rather than compete.
Applications: These findings provide a conceptual foundation for engineering wave dynamics in realistic, noisy environments (e.g., biological systems, optical networks, and quantum devices) where dissipation and decoherence are unavoidable. It suggests new strategies for wave control, such as using noise to induce localization or protect information in non-Hermitian settings.

In summary, the paper proves that non-Hermitian quasicrystals can sustain non-diffusive transport phases even in the complete absence of phase coherence, fundamentally altering our understanding of wave transport in complex, open media.