Observation of Floquet erratic non-Hermitian skin effect in photonic mesh lattice

This paper experimentally demonstrates the disorder-enabled, bulk-localized erratic non-Hermitian skin effect in a driven photonic mesh lattice, revealing a unique topological transition where wave dynamics self-organize into bulk patterns independent of boundaries, distinct from both conventional skin and Anderson localization.

The Gist

Imagine you are watching a crowd of people trying to walk through a long, narrow hallway.

In a normal hallway (a "Hermitian" system), people spread out evenly. If there's a wall at the end, they stop there, but they don't all pile up against it unless something forces them to.

Now, imagine this hallway has a strange, invisible wind blowing through it. This is the Non-Hermitian Skin Effect (NHSE). In this scenario, the wind pushes everyone toward one specific end of the hallway. No matter where you start, everyone eventually gets squashed against that single wall. It's like a crowd being herded by a strong, one-way breeze. Scientists have known about this "skin effect" for a while, but they always thought it only happened in orderly, predictable hallways.

The Big Surprise: The "Erratic" Crowd

This paper reports a discovery that breaks the rules. The researchers found a way to create a situation where the wind is chaotic and random, yet the crowd still piles up—but not against the walls. Instead, they pile up right in the middle of the hallway, forming a floating island of people that doesn't touch either wall.

They call this the Erratic Non-Hermitian Skin Effect (ENHSE).

Here is how they did it and what it means, using simple analogies:

1. The Playground: A Photonic Mesh Lattice

Think of the experiment as a giant, high-tech pinball machine made of fiber optic cables (glass threads that carry light).

• The Balls: Instead of metal balls, they use pulses of light.
• The Flippers: They use special modulators to speed up, slow down, or change the direction of the light.
• The Chaos: They programmed the machine to introduce "disorder." Imagine the floor of the hallway suddenly having random bumps, dips, and gusts of wind that change every second.

2. The Magic Ingredient: "Imaginary Gauge Fields"

This is the scientific term for the "wind" mentioned earlier. In this experiment, the researchers didn't just blow wind; they programmed the light to experience gain (getting brighter) in some spots and loss (getting dimmer) in others, in a very specific, random pattern.

• The Analogy: Imagine a game of "Red Light, Green Light," but the rules change randomly for every player. Sometimes the light gets a boost (gain), sometimes it gets drained (loss).
• The Trick: The researchers tuned this chaos so that, on average, the wind blowing left equals the wind blowing right (Global Reciprocity). You might think, "If the winds cancel out, the light should just spread out randomly."

3. The Discovery: The "Floating Island"

Here is the magic moment. When the average wind is balanced (zero net wind), but the local wind is chaotic:

• The Old Expectation: The light should scatter everywhere or get stuck in random spots (Anderson Localization).
• The Reality: The light pulses stopped scattering. Instead, they self-organized into a single, stable clump right in the center of the hallway.

This is the Erratic Skin Effect.

• Why "Erratic"? Because the wind is chaotic.
• Why "Skin"? Because the light is still "skinny" or localized, just like the wall-piling effect, but it's not on the wall.
• Why "Bulk"? It happens in the middle of the system, not at the edges.

It's like if you threw a handful of confetti into a room with a chaotic fan, and instead of it hitting the walls or floating everywhere, all the confetti magically stuck together in a floating ball in the exact center of the room, defying gravity and the chaos around it.

4. The Competition: Chaos vs. Static

The researchers also tested what happens if you add more disorder, like making the floor extremely bumpy (adding "on-site disorder").

• Result: The "floating island" of light starts to break apart. The chaotic bumps become so strong that they trap the light in random, tiny spots (this is called Anderson Localization).
• The Lesson: There is a delicate balance. If the "wind" (the imaginary gauge field) is just right, you get the magical floating island. If the "bumps" (scattering) get too strong, the magic disappears, and the light gets stuck in random holes.

Why Does This Matter?

For a long time, scientists thought that to get these special "skin" effects, you needed a perfectly ordered, predictable system. This paper proves that disorder can actually create new, useful structures.

• New Physics: It shows that nature has a hidden layer of order inside chaos. Even when things look random, they can self-organize into stable patterns in the middle of nowhere.
• Future Tech: This could help engineers design better ways to control light and energy. Imagine building a computer chip where data (light) is automatically routed to a specific spot in the middle of the chip, regardless of the noise or interference around it. It opens the door to "disorder-engineered" technology, where we use randomness as a tool rather than a bug.

In short: The researchers built a chaotic light maze and discovered that, under the right conditions, the light doesn't hit the walls or get lost—it forms a perfect, floating island in the center. It's a new kind of magic trick where chaos creates order.

Technical Summary

1. Problem Statement

The paper addresses a fundamental gap in the understanding of non-Hermitian physics, specifically the Non-Hermitian Skin Effect (NHSE).

• Context: In ordered, translationally invariant non-Hermitian systems, the NHSE is a boundary phenomenon where nonreciprocal hopping causes an extensive accumulation of eigenstates at the system's edges, while the periodic-boundary spectrum remains extended.
• The Gap: While disorder is known to generally suppress the conventional NHSE or induce Anderson localization, a recent theoretical prediction proposed a distinct phenomenon: the Erratic Non-Hermitian Skin Effect (ENHSE).
• The Challenge: ENHSE is characterized by bulk-localized eigenstates driven by spatially fluctuating imaginary gauge fields, occurring even when global reciprocity is preserved (zero net bias). Unlike conventional NHSE, it does not require boundaries; unlike Anderson localization, the eigenstates share a common spatial envelope determined by the disorder rather than individual exponential centers. Despite theoretical proposals, the experimental realization of ENHSE had remained elusive due to the difficulty in engineering specific, spatially fluctuating gain/loss profiles in a controlled manner.

2. Methodology

The authors utilized a time-multiplexed photonic mesh lattice to simulate a 1D Floquet non-Hermitian system with programmable disorder.

• Experimental Platform:
  • The setup consists of two fiber loops of unequal lengths connected by a 50:50 fiber coupler.
  • Light pulses circulate in the loops, with the time-bin encoding creating an effective 1D lattice.
  • Control Mechanism: The system employs three Amplitude Modulators (AMs) and one Phase Modulator (PM).
    • The AMs control the coupling coefficients ($G_{uu}, G_{uv}, G_{vu}, G_{vv}$) between the loops, allowing for independent programming of gain and loss (attenuation/amplification) at each lattice site.
    • The PM introduces tunable on-site potential disorder (real part of the Hamiltonian).
• Theoretical Model:
  • The system simulates a disordered Hatano-Nelson model with a non-Hermitian tight-binding Hamiltonian:
    $$\hat{H} = \sum_{n=1}^{N-1} \left( J_n^R \hat{c}_{n+1}^\dagger \hat{c}_n + J_n^L \hat{c}_n^\dagger \hat{c}_{n+1} \right)$$
  • Disorder is introduced via imaginary gauge fields $\{h_n\}$, where hopping amplitudes are $J_n^L = J e^{-h_n}$ and $J_n^R = J e^{h_n}$.
  • The key parameter is the global reciprocity $\bar{h}$ (the statistical mean of $h_n$).
    • $\bar{h} \neq 0$: Global nonreciprocity (conventional NHSE).
    • $\bar{h} = 0$: Global reciprocity (target for ENHSE).
• Implementation of Disorder:
  • The authors engineered the gain/loss profile by programming the AMs to follow a specific sequence of random variables $\{h_n\}$ (Bernoulli distribution in the main experiment).
  • The time-evolution operator was derived via an inverse similarity transformation, mapping the uniform quantum walk to a system with disordered imaginary gauge fields.

3. Key Contributions

1. First Experimental Observation of ENHSE: The paper provides the first direct experimental evidence of the Erratic Non-Hermitian Skin Effect, confirming the theoretical prediction that disorder can induce bulk localization without boundaries.
2. Engineering Global Reciprocity with Local Nonreciprocity: The team successfully demonstrated a system where strong local nonreciprocity (fluctuating gain/loss) coexists with global reciprocity ($\bar{h}=0$), a condition necessary for ENHSE.
3. Topological Phase Transition: They identified and observed a non-Hermitian topological phase transition driven by the tuning of global reciprocity ($\bar{h}$), characterized by a change in the real-space winding number.
4. Disentangling Localization Mechanisms: The study clarifies the competition between ENHSE and Anderson localization, showing how increasing on-site disorder suppresses the erratic skin dynamics.

4. Results

• Phase Transition via Winding Number:
  • The authors calculated the real-space winding number ($W$).
  • For $\bar{h} < 0$ and $\bar{h} > 0$, $W = -1$ and $W = +1$ respectively, indicating conventional skin accumulation at the left and right boundaries.
  • At the critical point $\bar{h} = 0$, the winding number vanishes ($W=0$), signaling the transition to the ENHSE phase.
• Observation of Bulk Localization:
  • $\bar{h} \neq 0$: Eigenstates accumulated at the physical boundaries (Left or Right), consistent with standard NHSE.
  • $\bar{h} = 0$ (ENHSE): Eigenstates did not accumulate at the boundaries. Instead, they formed stable, localized patterns within the bulk of the lattice.
  • Distinctive Feature: Unlike Anderson localization where different eigenstates localize at different random centers, all ENHSE eigenstates shared a common spatial envelope determined by the integrated imaginary gauge field ($X_n$). This envelope was non-exponential and specific to the disorder realization.
• Dynamics Verification:
  • Time-evolution simulations and experimental measurements showed that an injected wave packet evolved into a stable bulk-localized pattern after several round trips, matching the predicted eigenstate profiles.
• Competition with Anderson Localization:
  • When on-site disorder (phase disorder $\phi_n$) was introduced:
    • Weak disorder: ENHSE dynamics persisted, though with some spreading.
    • Strong disorder ($\phi_{max} = \pi$): The system transitioned to Anderson localization. The multiple scattering from strong on-site disorder suppressed the integration of the link variables required for the global reciprocity effect, effectively destroying the ENHSE.

5. Significance

• New Paradigm for Localization: This work establishes ENHSE as a unique, disorder-induced non-Hermitian phenomenon that defies both the conventional Bloch band theory (which assumes translation invariance) and the standard boundary-based skin effect framework.
• Topological Control: It demonstrates that topological properties (winding numbers) and localization behaviors can be tuned continuously by adjusting the global balance of nonreciprocity, even in the presence of strong disorder.
• Technological Implications: The ability to engineer "bulk-localized" states without interfaces opens new avenues for controlling wave transport, designing robust photonic devices, and exploring non-Hermitian physics in higher dimensions or with nonlinearities.
• Experimental Benchmark: The successful implementation using programmable photonic mesh lattices sets a new standard for simulating complex non-Hermitian Hamiltonians with spatially varying gain and loss.