🔬 mesoscale physics

Bypass Transition via the Non-Hermitian Skin Effect

This paper demonstrates that the interplay between non-Hermitian skin effect amplification and nonlinear feedback can trigger a bypass transition to self-sustained periodic or chaotic motion in linearly stable lattices, establishing a new framework for predicting and controlling nonlinear wave dynamics through return-map theory.

Original author: Konghao Sun, Haiping Hu

Published 2026-11-05

📖 6 min read🧠 Deep dive

CC BY 4.0

Original author: Konghao Sun, Haiping Hu

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 author. For technical accuracy, refer to the original paper. Read full disclaimer

In the world of physics, there is a constant tension between order and chaos. In many systems, from the flow of water in a river to the vibration of a guitar string, a state of rest is naturally stable. If you nudge the system slightly, it wobbles and then settles back down. To make such a system move on its own, you usually need to push it hard enough to break that stability, forcing it into a new, active state. This is the standard rule: if a system is stable, it stays stable unless you apply a massive force. However, there is a special class of materials and devices where this rule seems to bend. These are systems that do not conserve energy in the usual way; they are designed to absorb energy from their surroundings or lose it in specific directions. In these environments, waves can behave strangely, piling up at the edges rather than spreading out evenly. This phenomenon, known as the non-Hermitian skin effect, allows signals to be amplified as they travel in one direction, acting like a one-way street for energy. The question researchers have long asked is whether this directional amplification, combined with the messy, unpredictable nature of real-world materials, can trigger a system to start moving on its own, even when a tiny nudge should theoretically be enough to stop it.

A team of physicists in Beijing has now shown that the answer is yes, and they have mapped out exactly how it happens. They studied a theoretical model consisting of two long, open chains of sites, like two parallel rows of stepping stones. In this setup, waves travel along the first row, gathering strength as they move toward one end, much like a crowd of people rushing toward a single exit. When they reach the end, a specific mechanism converts some of their energy into the second row, where the waves travel in the opposite direction, gathering strength again as they rush toward the other end. Finally, they are converted back into the first row, completing a full loop. The researchers found that if the initial push is strong enough, this loop creates a self-sustaining cycle. The waves amplify themselves with every trip around the loop, eventually growing into a permanent, rhythmic motion or even a chaotic, unpredictable dance that never stops.

What makes this discovery surprising is the size of the push required to start it. In a normal system, you would expect the threshold to be a fixed, manageable amount. Here, the researchers found that the amount of energy needed to trigger this self-sustaining motion drops dramatically as the chains get longer. In fact, the required energy decreases exponentially with the length of the chain. This means that for a sufficiently long system, even a minuscule, almost invisible nudge can set off a massive, permanent reaction. The resting state of the system remains mathematically stable, meaning that if you look at it with standard tools, it appears perfectly safe and quiet. Yet, a tiny disturbance can bypass this stability entirely, pushing the system into a state of perpetual motion. This is what physicists call a "bypass transition," a route to chaos that skips the usual warning signs of instability.

To understand this process, the researchers developed a new way of thinking about the system, which they call a return-map theory. Instead of trying to track every single wave as it moves, they focused on the most important moments: when a wave packet arrives at the end of a chain and when it returns after being converted. They calculated how much the wave grows during one trip and how much it grows when it comes back. By comparing these two numbers, they could predict exactly how strong the initial push needed to be to keep the cycle going. Their theory matched their computer simulations perfectly, showing that the system behaves like a feedback loop where the output of one trip becomes the input for the next. If the loop is strong enough, the signal grows; if it is too weak, the signal fades away. The theory also revealed that the system has a critical, unstable orbit—a specific pattern of motion that acts as a boundary. If the system is nudged just slightly below this boundary, it fades back to rest. If it is nudged just above, it explodes into sustained motion.

The researchers also explored what happens when the system is not perfectly quiet but is instead jostled by random noise, which is always present in real-world environments. They found that even without a deliberate push, random fluctuations can eventually trigger the same self-sustaining motion. However, the system needs to be long enough for the noise to be amplified sufficiently to cross the threshold. They discovered a clear rule for this: the length of the chain required to start the motion grows logarithmically as the noise gets weaker. In other words, if the background noise is very faint, you need a much longer chain to catch enough of it to start the engine. This finding suggests that in real devices, such as sensors or amplifiers, the length of the device is a critical factor in determining whether it will stay quiet or start oscillating on its own.

The work also clarifies where the energy for this motion comes from. The system does not create energy out of nothing; it draws it from the directional amplification built into the material. The non-reciprocal hopping, which causes the waves to prefer one direction, acts as a pump, feeding energy into the loop. The conversion between the two chains simply routes this energy back into the pump, allowing it to be used again and again. Without this specific routing, the energy would simply dissipate. The researchers showed that if the chains were arranged to amplify in the same direction, the motion would not sustain itself, proving that the opposing directions are essential for closing the loop. This mechanism is robust, surviving even when the chains are not identical or when the material properties are slightly altered.

This research provides a new blueprint for understanding how complex, self-sustaining behaviors can emerge in systems that are technically stable. It shows that the combination of directional amplification and nonlinear feedback can create a pathway to chaos that is distinct from the traditional routes. The findings have potential implications for the design of active devices, such as lasers or sensors, where controlling the onset of oscillation is crucial. By understanding the precise thresholds and the role of system size, engineers might be able to design devices that are immune to accidental triggering or, conversely, that can be switched on with the smallest possible signal. The study confirms that in the realm of non-Hermitian physics, the rules of stability are more flexible than previously thought, opening up new possibilities for controlling how waves move and interact in engineered materials.

Technical Summary: Bypass Transition via the Non-Hermitian Skin Effect

Problem Statement
The non-Hermitian skin effect (NHSE) is known to cause extensive eigenmode accumulation at boundaries and directional amplification of propagating waves. While transient amplification can occur even when all eigenmodes decay, such growth typically yields to decay without external driving. The paper addresses a fundamental gap in understanding how nonlinear feedback interacts with NHSE-induced amplification to produce self-sustained motion. Specifically, it investigates whether a finite local excitation can trigger a "bypass transition"—a route to sustained periodic or chaotic motion while the resting state remains linearly stable—analogous to the bypass transition from laminar flow to turbulence in shear flows. A predictive theory linking skin amplification to this onset has been lacking.

Methodology
The authors propose a minimal model consisting of two open chains (A and B), each with $L$ sites, possessing opposite skin directions. The chains are coupled via an intensity-dependent conversion rate $q_n = \chi I_n$ , where $I_n = |a_n|^2 + |b_n|^2$ is the local intensity. The system is governed by coupled differential equations featuring asymmetric hopping (controlled by $\gamma$ ) and uniform loss ( $\Gamma$ ).

The study employs a combination of direct numerical simulation and analytical theory:

Numerical Simulations: The system is initialized with a single-site excitation $a_1(0) = A_0$ while all other amplitudes are zero. The threshold amplitude $A_c$ separating decay from sustained motion is determined via bisection. Stability is analyzed using Lyapunov exponents and Floquet multipliers.
Return-Map Theory: To predict the triggering threshold analytically, the authors develop a return-map theory. They expand the fields in powers of the seed amplitude $A_0$ . The linear propagation of a unit seed along chain A generates a field of order $A_0$ . The nonlinear conversion creates a field of order $A_0^3$ in chain B, which propagates back to chain A, generating a feedback term of order $A_0^5$ .
Noise-Induced Ignition: The study also examines the effect of weak complex Gaussian white noise added to the resting state to determine the characteristic ignition length required for noise to trigger sustained motion.

Key Contributions and Results

Bypass Transition Mechanism: The paper demonstrates that a finite local excitation can trigger sustained periodic or chaotic motion even when the resting state is linearly stable (all eigenmodes decay at rate $\Gamma$ ). This occurs because the NHSE provides directional amplification, and the nonlinear conversion closes a feedback loop. A packet amplified along one chain is converted to the other, amplified again on its return, and re-injected.
Exponential Scaling of Threshold: The triggering threshold $A_c$ decreases exponentially with the chain length $L$ . This is a counter-intuitive result given that the linear decay rate $\Gamma$ remains constant regardless of $L$ . The threshold scales as $A_c \propto G^{-3/2}$ , where $G$ is the directional gain factor, which itself scales exponentially with $L$ ( $G \sim e^{r_\infty L}$ ).
Return-Map Prediction: The authors derive an analytical expression for the threshold: $\hat{A}_c = 1/(G\sqrt{K})$ , where $G$ is the peak gain of the first arrival and $K$ is the nonlinear return coefficient. This theory predicts the threshold and its scaling with system parameters without fitting to the threshold data. The theory accurately reproduces numerical results across various chain lengths and parameter sets.
Critical Orbit and Near-Critical Dynamics: The return map predicts a nonzero fixed point $x^*$ corresponding to a critical circulating orbit. Trajectories near the threshold approach this unstable orbit before either decaying or diverging into sustained motion. The instability of this orbit is quantified by a Floquet multiplier $\Lambda_u$ , which approaches the value 9 (derived from the cubic nonlinearity) as chain length increases. The departure time from the critical orbit scales logarithmically with the distance from the threshold ( $\ln(1/\epsilon)$ ).
Noise-Induced Ignition: Weak noise can also trigger sustained motion. The characteristic "ignition length" $L_{1/2}$ (the chain length where 50% of realizations ignite) grows logarithmically as the noise strength $D$ decreases. This scaling arises because the directional gain amplifies noise fluctuations exponentially, compensating for the reduced noise amplitude.
Power Balance: In the sustained regime, the energy source is the nonreciprocal hopping, which supplies power to balance the uniform loss. The initial seed triggers the motion but does not power it; the system becomes autonomous once the feedback loop is established.

Significance
The paper establishes a "skin-effect route" to bypass transitions, revealing that NHSE amplification acts not merely as a transient response but as a dynamical resource that nonlinear feedback can repeatedly harness. The work provides a tractable framework for predicting and controlling nonlinear wave dynamics in non-Hermitian systems. The authors suggest these findings are relevant to the stability and performance of active non-Hermitian devices and could inform the design of switches and locally triggered oscillators. The mechanism is argued to be structural, surviving in variations such as unequal propagation channels, dimerized lattices, and different nonlinear orders, pointing to a broader class of nonreciprocal systems where transient amplification converts into autonomous dynamics.

More like this