⚛️ quantum physics

Independent density and coherence skin effects in adaptive fermion circuits

This paper demonstrates that adaptive fermion circuits exhibit independently tunable density and coherence skin effects, where the stationary density profile is determined solely by measurement strength and feedback displacement while coherence localization is controlled by conditioned gates, enabling a topological transition that can localize density and coherence at opposite edges.

Original authors: Karim Chahine, Markus Siegl, Michael Buchhold

Published 2026-09-30

📖 7 min read🧠 Deep dive

CC BY 4.0

Original authors: Karim Chahine, Markus Siegl, Michael Buchhold

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

In the quantum world, particles do not always behave like the solid objects we see in daily life. They can exist in a state of superposition, where they are not just in one place or another, but in a blend of possibilities at once. This blend is called coherence, and it is the fragile engine behind many of the most powerful technologies scientists hope to build, from ultra-fast computers to unhackable communication networks. However, keeping these particles under control is difficult. Usually, when scientists try to push particles in a specific direction, they lose control over their delicate quantum connections, or when they try to preserve those connections, the particles refuse to move where they are told. For a long time, it seemed that the rules of quantum mechanics forced a trade-off: you could have a flow of matter, or you could have a flow of information, but rarely both in the same system with independent control.

A team of researchers has now shown that this trade-off is not a fundamental law, but a limitation of how we have been building these systems. By using a new kind of quantum circuit that combines measurement with immediate, automated feedback, they have created a setup where the flow of particles and the flow of their quantum connections can be steered in completely different directions. In this system, the particles can pile up at one end of a chain while their quantum connections pile up at the opposite end. This discovery opens a new door for designing quantum devices that can separate and manipulate these two essential properties independently, offering a way to build more robust and versatile quantum machines.

The researchers, working with a model of fermions—particles like electrons that make up ordinary matter—constructed a digital simulation of a quantum circuit. In this circuit, the particles move along a line of sites, but their movement is not governed by a fixed set of rules. Instead, the system constantly checks the state of the particles and adjusts its behavior based on what it finds. This process is known as an adaptive circuit. Imagine a traffic system where a camera watches a car, and depending on whether the car is there or not, a gate opens to let it move forward or backward. In the quantum version, this "camera" is a weak measurement that gently probes the system without destroying its delicate state, and the "gate" is a quantum operation that changes the particle's path based on the result.

The key innovation in this work is that the researchers used two different parts of this feedback loop to control two different things. The first part, the measurement itself, acts like a gentle wind that pushes the particles. If the measurement is strong enough, it creates a bias that makes the particles more likely to move in one direction than the other. This causes the particles to accumulate, or "skin," at one specific edge of the system. The researchers found that the steepness of this pile-up depends entirely on how strong the measurement is and how far the feedback pushes the particle. Crucially, this effect is rigid; it does not care about the details of the quantum gates used to move the particles. The particles will always pile up at the same edge with the same shape, regardless of how the quantum logic is tuned.

The second part of the system controls the quantum connections, or coherences, between the particles. These connections are what allow quantum computers to perform complex calculations. In most systems, if you try to push the particles, you also push the connections in the same direction. However, in this adaptive circuit, the researchers discovered that the phase of the quantum gates—the internal timing of the logic operation—can be tuned to control the connections independently. By adjusting a single angle in the quantum gate, they could change the direction in which the connections accumulate. They could make the connections pile up at the opposite edge from the particles, or even make them pile up at the same edge but with a different steepness.

This separation of effects is what the researchers call independent skin effects. They demonstrated that by tuning the relative phase of the feedback gates, they could drive the system through a topological transition. This is a sudden change in the global structure of the system where the preferred direction for the connections flips. At a specific setting, the connections stop piling up at either edge and spread out evenly, before reversing and piling up at the opposite side. Meanwhile, the particles continue to pile up at their original edge, completely unaffected by this change. The researchers proved that the density of the particles is determined solely by the measurement strength, while the localization of the connections is determined by the interference between the different quantum gates.

To verify these findings, the team ran detailed simulations of the circuit. They observed that the stationary state of the system, where the particles settle down, showed a perfect exponential tilt in density that matched their predictions exactly. This tilt was independent of the gate settings. In contrast, the decaying modes of the system, which represent the quantum connections fading away, showed a different exponential profile that changed as they adjusted the gate phases. They also proposed a practical way to detect this effect in a real experiment. Instead of trying to measure the fragile connections directly, which is difficult, they suggested measuring the difference between two slightly different starting conditions. By comparing how these two states evolve, the hidden information about the connection's direction can be extracted without needing to select only the "successful" outcomes, a process that usually wastes most of the data.

The implications of this work are significant for the future of quantum technology. It shows that the non-reciprocal behavior, where things move differently in one direction than the other, does not have to be a single, monolithic property. By using adaptive feedback, scientists can now design systems where matter and information flow in different ways. This could lead to new types of quantum sensors or memory devices where the storage of data and the flow of data are physically separated. The researchers also noted that this effect persists even when the system is filled with many particles, not just a single one, suggesting that the principle is robust enough for real-world applications.

The study concludes that the adaptive circuit provides a unique dual structure that was missing in previous designs. While traditional systems could only offer one type of directional bias for both particles and connections, this new architecture allows for two distinct biases. The measurement backaction fixes the tilt of the particle density, while the interference of the conditioned gates fixes the winding of the coherence. This allows the two sectors to localize at opposite boundaries, a feat that was previously thought impossible. The researchers have provided the mathematical proof for the one-particle case and shown through simulations that the effect holds for the full system. They have also outlined a clear path for experimentalists to observe this phenomenon using current mid-circuit measurement technologies, turning a theoretical curiosity into a tangible tool for the next generation of quantum engineering.

Technical Summary: Independent Density and Coherence Skin Effects in Adaptive Fermion Circuits

Problem Statement
Non-Hermitian skin effects (NHSE) describe the accumulation of extensive eigenstates at a boundary due to non-reciprocal dynamics. While traditionally studied in non-Hermitian Hamiltonians (often representing post-selected, no-jump evolution), physical quantum systems evolve under trace-preserving Liouvillians. A fundamental distinction arises here: a non-Hermitian Hamiltonian provides a single spatial generator for both particle occupations (diagonal elements of the density matrix) and coherences (off-diagonal elements). In contrast, standard asymmetric-hopping Lindbladians typically offer only a single control parameter, causing occupations and coherences to either decay in place or translate with the same bias. The central question addressed is whether trace-preserving dynamics can endow density and coherence with independently tunable skin localization, potentially allowing them to accumulate at opposite boundaries.

Methodology
The authors investigate an adaptive fermion circuit architecture consisting of a chain of $L$ sites with spinless fermions. The dynamics are driven by a measurement-feedback loop applied stochastically at each site $\ell$ with rate $\eta$ :

Weak Measurement: A local occupation $\hat{n}_\ell$ is weakly measured with strength $\tau$ , yielding outcomes $\sigma = \pm 1$ with Born probabilities determined by measurement operators $\hat{M}_{\sigma, \ell}$ .
Conditioned Unitary Gate: Based on the outcome $\sigma$ , a unitary gate $\hat{U}_{\sigma, \ell}$ is applied, coupling site $\ell$ to $\ell + \sigma d$ . These gates are specified by $U(2)$ matrices $u_\sigma$ , containing SU(2) mixing parameters and a relative occupation phase $\gamma_\sigma$ .
Ensemble Dynamics: The system evolves according to the Liouvillian $\mathcal{L}\hat{\rho} = \eta \sum_\ell \sum_{\sigma} (\hat{K}_{\sigma, \ell} \hat{\rho} \hat{K}^\dagger_{\sigma, \ell} - \hat{\rho})$ , where $\hat{K}_{\sigma, \ell} = \hat{U}_{\sigma, \ell}\hat{M}_{\sigma, \ell}$ are the Kraus operators.

The study focuses on the one-particle sector and arbitrary fillings, analyzing the stationary state and the decaying eigenmodes of the Liouvillian. Two specific gate configurations are compared: a "conditioned hop" where outcomes bias transport in the same direction, and a "balanced pair" where the mixing probabilities are fixed, but a relative phase $\theta = \gamma_+ - \gamma_-$ is tunable.

Key Contributions and Results

Rigid Density Localization: The authors derive an exact stationary state for the diagonal density elements $n_R$ . The profile exhibits an exponential tilt $n_R \propto e^{\kappa_{\text{dens}} R}$ with a localization exponent:
$\kappa_{\text{dens}} = \frac{2\tau}{d}$
Crucially, this exponent depends only on the measurement strength $\tau$ and the feedback displacement $d$ . It is independent of the specific details of the conditioned $U(2)$ gates. This result holds for arbitrary filling, where the stationary state is a tilted ensemble of configuration weights.
Tunable Coherence Localization: In contrast to the density, the decaying single-particle coherence modes (off-diagonal elements) exhibit a skin effect governed by a Hatano–Nelson generator. The localization exponent $\kappa_{\text{coh}}$ is derived as:
$\kappa_{\text{coh}} = \frac{1}{d} \ln \left| \frac{t_R}{t_L} \right| = \frac{1}{d} \left[ \tau + \frac{1}{2} \ln \frac{|B_+|^2}{|B_-|^2} \right]$
where $B_\sigma$ depends on the gate parameters. Unlike the density, $\kappa_{\text{coh}}$ retains dependence on the conditioned-gate phases. By tuning the relative phase $\theta$ in the "balanced pair" configuration, $\kappa_{\text{coh}}$ can be varied continuously through zero without affecting $\kappa_{\text{dens}}$ .
Opposite-Edge Localization and Topological Transition: The ability to tune $\kappa_{\text{coh}}$ independently allows for a regime where density and coherence localize at opposite edges of the chain. The sign reversal of $\kappa_{\text{coh}}$ corresponds to a point-gap topological transition in the effective Hatano–Nelson matrix governing the coherences. At the transition point ( $\kappa_{\text{coh}} = 0$ ), the localization length diverges.
Detection Protocol: Since coherences vanish in the stationary state, the authors propose a post-selection-free protocol to extract $\kappa_{\text{coh}}$ from transient, ensemble-averaged correlators. By preparing two states with specific initial coherences and measuring the ratio of correlators at symmetric positions, the exponential factor $e^{2\kappa_{\text{coh}}\Delta R}$ can be isolated, allowing direct extraction of the coherence exponent.

Significance
The paper demonstrates that adaptive circuits provide a dual structure capable of separating non-reciprocity into two distinct sectors. The measurement backaction fixes the stationary density tilt, while the interference between conditioned gates determines the coherence winding. This architecture enables the realization of a "coherence skin effect" that can be tuned independently of the "density skin effect," a feature absent in canonical asymmetric-hopping Lindbladians. The work establishes that trace-preserving dynamics can support a point-gap topological transition for coherences that leaves the stationary density profile completely unaffected, offering a new mechanism for controlling quantum transport and localization in open quantum systems.

More like this