🔬 physics

Nonlocality-induced critical-length hierarchy from non-Hermitian competition

This paper demonstrates that introducing long-range hoppings in non-Hermitian coupled chains fundamentally reorganizes the conventional logarithmic critical-length hierarchy into a diverse set of algebraic and scale-covariant scaling laws by leveraging nonlocal mechanisms like nonanalytic band-edge dispersion and parity-mixing hybridization.

Original authors: Mengjie Yang, Alexander Poddubny, Ching Hua Lee

Published 2026-08-19

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Mengjie Yang, Alexander Poddubny, Ching Hua Lee

Original paper licensed under CC BY 4.0 (https://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

In the hidden world of quantum materials and engineered light, scientists often study how waves travel through a grid of tiny, connected points. Usually, these waves behave predictably, spreading out evenly or getting trapped in specific spots. But when the rules of the game are changed to allow energy to flow more easily in one direction than another, a strange phenomenon occurs. The waves stop behaving like normal ripples and instead pile up, or "skin," against the edges of the material. This behavior, known as the non-Hermitian skin effect, makes the system incredibly sensitive to its boundaries. If you change the size of the material or the distance between its parts, the entire pattern of energy can shift dramatically, sometimes flipping from a stable state into a chaotic, unstable one. Understanding exactly when and how this flip happens is crucial for building future devices that use light or electricity to process information, as engineers need to know the precise limits before a system breaks down.

A team of researchers has now discovered that the way these systems fail is far more complex than previously thought, depending entirely on how far the connections between points reach. In standard models, where points only talk to their immediate neighbors, the system remains stable until it reaches a specific size that grows very slowly as the parts are moved apart. However, the researchers found that if the connections are allowed to stretch across the material, reaching out to distant points, the rules of stability change completely. They studied a setup resembling a ladder made of two parallel chains, where the "rungs" connecting the chains could be either short and local or long and stretching across the gap. By simulating these systems with connections that decay at different rates, they mapped out a new hierarchy of stability that defies the old, simple rules.

The study reveals that the distance between the two chains and the reach of the connections create a competition that determines the system's fate. When the connections are short and local, the system behaves as expected: it takes a very long time for the instability to grow, and the critical size needed to trigger a breakdown increases only logarithmically as the chains are pulled apart. This means you can separate the chains quite far without much trouble. But when the researchers introduced long-range connections on the rungs between the chains, the behavior shifted to an algebraic growth. Here, the system becomes unstable much faster as the chains are separated, following a power-law relationship where the critical size grows steadily with the distance. This change alone was significant, but the most surprising discovery came when they made the connections within the chains themselves long-range as well.

In this fully connected, long-range scenario, the system entered a completely new regime where the concept of a fixed critical size disappeared. Instead of the system failing at a specific length, the point of failure became dependent only on the shape of the ladder, specifically the ratio of its length to its width. Whether the ladder was small or huge, it would remain stable as long as its proportions stayed the same. This scale-covariant behavior means that the physical size of the device no longer matters; only its geometry dictates its stability. The researchers identified two specific mechanisms driving this shift: the way long-range connections alter the energy landscape at the edges of the system, and a unique mixing of wave patterns caused by the one-way flow of energy. These effects combine to remove the usual physical limits that govern how large a system can be before it becomes unstable.

The findings suggest that nonlocality, or the ability of parts of a system to influence each other over long distances, does not just tweak the existing rules but fundamentally reorganizes them. The researchers showed that by tuning how far the connections reach, one can switch between three distinct types of critical behavior: the slow, logarithmic growth of local systems; the faster, algebraic growth of systems with long-range rungs; and the geometric, scale-invariant stability of fully long-range systems. This hierarchy is not a smooth transition but a series of distinct phases where the most stable configuration can change depending on the specific parameters. For instance, a system that is very stable when the chains are close together might become the least stable when they are pulled far apart, overtaken by other configurations. This non-monotonic behavior, where stability rises and falls in a complex pattern, was a key finding that challenges the idea that adding more connections simply makes a system more robust or more fragile in a straightforward way.

These results were derived through rigorous mathematical analysis and confirmed by numerical simulations on finite-sized lattices. The team calculated the exact point where the energy spectrum of the system would split from real numbers into complex numbers, signaling the onset of instability. They found that for systems with very long-range connections, the critical threshold is determined by a balance between the energy gaps of the individual chains and the collective strength of the long-range couplings. This balance leads to the unique scale-covariant regime where the aspect ratio is the only deciding factor. The researchers also noted that this behavior is bounded by a physical limit related to the skin depth, beyond which the system's behavior reverts to a different pattern, but within a wide range of parameters, the new geometric rule holds true.

The implications of this work extend beyond theoretical physics, offering a blueprint for designing robust systems in photonics, electrical circuits, and quantum simulators. Since these platforms can be programmed to have specific connection patterns, engineers could potentially use these findings to create devices that are inherently stable against size variations or, conversely, highly sensitive to small geometric changes. The ability to switch between different stability regimes by simply adjusting the range of interactions suggests a new way to control amplification and signal processing. The researchers emphasize that this framework is testable in programmable topoelectrical circuits and photonic lattices, where the connections can be reconfigured electronically. By understanding that the critical length is not a fixed number but a function of geometry and connection range, designers can build systems that avoid unexpected failures or harness these transitions for sensing applications.

Ultimately, this paper demonstrates that the boundary sensitivity of non-Hermitian systems is not an intrinsic property of the amplification mechanism itself, but a result of how far the underlying couplings reach. The study rules out the idea that long-range interactions merely rescale the existing critical behavior; instead, it shows they qualitatively reshape the instability thresholds. The hierarchy of critical lengths, ranging from logarithmic to algebraic to scale-invariant, provides a comprehensive map of how nonlocality reorganizes the physics of these systems. This work offers a platform-independent framework for understanding and controlling the stability of complex networks, proving that the geometry of connections is just as important as the strength of the interactions within them.

Technical Summary: Nonlocality-induced critical-length hierarchy from non-Hermitian competition

Problem Statement
Non-Hermitian lattices exhibit intrinsic nonlocal responses, most notably the non-Hermitian skin effect (NHSE), where eigenmodes accumulate at boundaries. In short-range systems composed of two coupled chains with antagonistic (opposite) non-reciprocity, a competition exists between skin accumulation and inter-chain hybridization. This competition typically leads to a "critical NHSE" phenomenon: a transition from a real to a complex energy spectrum occurs only when the system length $N$ exceeds a critical threshold $N_c$ . Conventionally, for short-range models, this threshold scales logarithmically with the inter-chain separation $D$ ( $N_c \sim \ln D$ ).

The central problem addressed is how this critical behavior is reshaped when long-range (power-law decaying) hoppings are introduced. Specifically, the authors investigate whether long-range couplings merely rescale the existing logarithmic instability or fundamentally reorganize the spectral transition into a new hierarchy of scaling laws.

Methodology
The study employs a theoretical framework based on a minimal ladder model consisting of two coupled non-Hermitian chains (legs) separated by a distance $D$ . The authors analyze three distinct geometric configurations to isolate the effects of nonlocality:

Model I (NN–NN): Both intra-chain (leg) and inter-chain (rung) hoppings are local (nearest-neighbor).
Model II (NN–LR): Legs are local, but inter-chain rungs exhibit power-law decay ( $\propto |n-m|^{-\alpha}$ ).
Model III (LR–LR): Both legs and rungs exhibit power-law decay.

The analysis proceeds by:

Finite-Size Bloch Symbol Analysis: Deriving the band-edge eigenenergies ( $E_1, E_2$ ) for single chains with open boundary conditions (OBCs) using truncated polylogarithm functions to capture non-analytic dispersion relations arising from long-range hoppings.
Exceptional Point (EP) Projection: Projecting the full ladder Hamiltonian onto the subspace of the two single-chain modes ( $|a\rangle, |b\rangle$ ) that first collide to form an EP. This reduces the problem to a $4 \times 4$ effective Hamiltonian.
Critical Threshold Derivation: Solving the EP discriminant condition $\Lambda^2_\eta + 4u_{ab}v_{ba} < 0$ , where $\Lambda_\eta$ involves the single-chain energy spacing and projected rung shifts, and $u_{ab}v_{ba}$ represents the hybridization product. The critical length $N_c$ is defined as the minimum system size where this condition is met.
Analytical and Numerical Validation: Deriving closed-form scaling laws for $N_c$ in various regimes of the power-law exponent $\alpha$ and comparing them against exact diagonalization of finite-size lattice Hamiltonians.

Key Contributions and Results

Reorganization of Critical Scaling: The paper demonstrates that long-range hoppings fundamentally alter the critical-length hierarchy, moving beyond the conventional logarithmic law ( $N_c \sim \ln D$ ).
- Model I (Local): Confirms the standard logarithmic scaling $N_c \sim \ln D$ , arising from the balance between exponentially small skin overlap and algebraically weak local coupling.
- Model II (Nonlocal Rungs): When only inter-chain couplings are long-range, the scaling becomes algebraic: $N_c \sim D^{\alpha/3}$ . This arises because long-range rungs aggregate contributions from many nodes, creating a collective hybridization that grows linearly with $N$ , balancing against the quadratic band-edge spacing of the local legs ( $\Delta E \sim N^{-2}$ ).
- Model III (Fully Nonlocal): When both legs and rungs are long-range, the system enters a scale-covariant regime for $\alpha < 2$ . Here, the critical threshold depends solely on the aspect ratio $N_c/D$ , independent of the absolute system size. The scaling is linear ( $N_c \propto D$ ).
Hierarchy of Regimes in Model III: The authors identify a complex hierarchy of scaling behaviors in the fully nonlocal limit depending on $\alpha$ :
- $\alpha < 2$ : Scale-covariant linear growth ( $N_c/D = \text{const}$ ).
- $\alpha = 2$ : Marginal regime with logarithmic corrections.
- $2 < \alpha < 3$ : Algebraically corrected regime where the aspect ratio is no longer strictly fixed.
- $\alpha \ge 3$ : Transition back to algebraic scaling ( $N_c \sim D^{\alpha/3}$ ) before saturating at a finite-size crossover scale $N_{sat}$ .
Mechanisms Identified: Two new non-local mechanisms are identified as the drivers of this hierarchy:
- Non-analytic Band-Edge Dispersion: Long-range intra-chain hoppings modify the single-chain energy gap scaling from $N^{-2}$ (local) to $N^{1-\alpha}$ (long-range).
- Parity-Mixing Hybridization: Non-reciprocity in long-range chains induces a mixing of band-edge modes that is distinct from the skin deformation in local chains. This mixing amplitude $\chi_N(\alpha)$ exhibits unique size-dependence (e.g., constant for $\alpha < 2$ , logarithmic for $\alpha=2$ ), which dictates the critical threshold.
Stability Phase Diagram: By comparing $N_c$ across the three models, the authors construct a phase diagram in the $(D, \alpha)$ plane. They show that the "most stable" configuration (largest $N_c$ ) is not fixed but varies non-monotonically with $D$ and $\alpha$ . For instance, at small $D$ and specific $\alpha$ , the fully nonlocal model may be the most stable, while at larger $D$ , the fully local model may become more stable, illustrating a hierarchy of first-order-like crossings between different instability mechanisms.

Significance and Claims
The paper claims to provide a platform-independent framework for understanding finite-size criticality in non-Hermitian systems. The primary significance lies in demonstrating that nonlocality systematically removes the physical length scales (specifically the skin depth) that underpin conventional critical NHSE behavior.

Theoretical Impact: The work challenges the view that nonlocality is merely "more connectivity." Instead, it shows that nonlocality creates interference pathways across multiple length scales, reshaping relevant gaps and renormalizing finite-size crossovers. The transition from logarithmic to algebraic to scale-covariant scaling represents a fundamental reorganization of spectral transitions.
Experimental Relevance: The authors assert that these phenomena are testable in existing programmable platforms, specifically citing topoelectrical circuits, photonic lattices, and digital quantum simulators. They propose that mapping the critical length $N_c$ by scanning system size at fixed parameters ( $D, \alpha, \gamma$ ) can experimentally verify these scaling laws.
Sensing Applications: The sensitivity of the critical threshold to the geometric aspect ratio ( $N_c/D$ ) suggests potential applications in sensing, where small structural changes could trigger a switch across an exceptional-point threshold, resulting in measurable spectral splitting.

The authors maintain a modest tone regarding future implications, focusing on the rigorous derivation of the scaling laws and their verification in the proposed platforms, rather than speculating on broader technological applications beyond the immediate scope of non-Hermitian physics and sensing.

More like this