The Quantum Magnets That Forgot How to Forget
12 May 2026, Yanjiang
A chain of Rydberg atoms inside an optical cavity exhibits both blockaded ferromagnetism and persistent quantum many-body scars that resist thermalization.
What if a quantum system’s destiny was not to forget everything about its past?
This is the question that hums beneath a new theoretical framework from a team led by Hossein Hosseinabadi at Johannes Gutenberg University Mainz. Their preprint (arXiv:2510.02246) — co-authored with collaborators spanning ICFO in Barcelona, the Max Planck Institute for the Physics of Complex Systems in Dresden, and beyond — introduces a minimal model for what happens when you trap atoms inside an optical cavity, tickle them with lasers into Rydberg states, and then watch as two fundamentally different kinds of interaction compete for control.
The surprise is not just that the system finds exotic ways to order itself. It is that, even when it should be dissolving into the featureless soup of thermal equilibrium, it stubbornly remembers where it started. This is the physics of quantum many-body scars — and here, in this hybrid platform of light and matter, those scars turn out to be something new.
But let us start not with the scars, but with the tension that creates them.
The Traffic Rules of the Quantum World
Picture a chain of atoms, each one a two-level system — a quantum bit — suspended in a line by optical tweezers. The whole ensemble sits inside an optical cavity, a hall of mirrors where a single mode of light can bounce back and forth, touching every atom with equal intimacy. This is the first interaction: a long-range, democratic coupling. Every atom talks to every other atom through the cavity photon, like guests at a dinner party who can all hear the same toast.
A chain of atoms trapped in a cavity feels both long-range and short-range forces. Their competition creates exotic quantum states not seen in ordinary matter. (Source: arXiv:2510.02246)
Now add a second, much ruder rule. When you drive these atoms into their excited Rydberg states — those bloated, hypersensitive electronic configurations where a single atom can be microns across — they develop a violent aversion to their neighbors. Two adjacent atoms cannot both be excited. The energy cost is simply too high. This is the Rydberg blockade, and it is a short-range, dictatorial constraint: you shall not excite next to an already excited neighbor.
Hosseinabadi and colleagues built a model that takes both rules seriously. They worked in the regime where the blockade is the strongest energy scale in the problem — strong enough that the Hilbert space, the mathematical arena of all possible quantum states, gets carved down to only those configurations that obey the blockade. This is not a passive filter. It is an active, kinetic constraint, one that shapes everything that follows.
The result is something they call the (PXP)² model — a kinetically constrained, long-range spin model in one dimension. Its name is technical, but its behavior is anything but dry.
An important question, however, lingers from experimental reality. Recent work by Marsh and colleagues (arXiv:2505.22658) has shown that real cavities host many modes, not just one, and that this multimode structure can generate disordered, glassy behavior rather than the clean all-to-all coupling assumed here. The single-mode approximation captures the essential physics, but it is a simplification — one that experimentalists will need to navigate carefully.
The Magnet That Shouldn’t Be
At equilibrium, when the system settles into its lowest-energy state, the model reveals three distinct phases. The first two are familiar enough: a paramagnetic phase where spins all point down in a featureless, unmagnetized state, and a Néel-ordered phase where they alternate up-down-up-down in a crystalline pattern of opposition.
The third phase is the one that should give you pause. It is a ferromagnetic phase — spins aligned, pointing together — but it arises under conditions where the blockade forbids the classical configuration that would normally produce ferromagnetism. You cannot simply align all spins, because the blockade prevents nearest-neighbor excitations. And yet the system finds a way.
Hosseinabadi and colleagues call this a blockaded ferromagnetic phase, and it is distinct from the conventional superradiant phase that cavity-QED systems are known for. The superradiance is still there — the cavity photon still mediates a collective polarization — but it is warped, constrained, forced to express itself through the limited vocabulary of blockade-compatible states.
The spectrum of low-energy excitations tells the story in mathematical form. At a particular value of the tuning parameter — what the team calls Delta, a detuning that controls the balance between the two interactions — the energy gap vanishes. The system goes through a second-order phase transition, and the ground state reorganizes itself into this new, blockaded magnetic order.
Think of it like a crowded room where everyone is told they must stand, but no two adjacent people can stand at the same time. The natural solution — everyone standing — is forbidden. And yet, through the mediating influence of a shared resource (the cavity photon), the crowd finds a collective stance that approximates unanimous agreement without violating the local rule. It is a compromise that nature negotiates without anyone at the table.
This is not a democratic process in any conscious sense; it is simply the ground state of a Hamiltonian that balances two incompatible demands.
The Scars That Remember
Now we arrive at the heart of the matter — and the reason this paper will be talked about in the quantum many-body community.
Out of equilibrium, most quantum systems follow the eigenstate thermalization hypothesis (ETH). The name is a mouthful, but the idea is intuitive: given enough time, a closed quantum system will explore all the states available to it, and any memory of its initial condition will be washed away. Entanglement grows, information scrambles, and local observables settle into thermal values. The system forgets.
Except when it doesn’t.
In 2017, physicists discovered that certain kinetically constrained models — most famously the PXP model describing Rydberg atoms — harbor a set of atypical eigenstates. These states, dubbed quantum many-body scars, have anomalously low entanglement and anomalously high overlap with simple, ordered initial states. They are the immune system of the quantum world: a small set of configurations that resist the infection of thermalization.
Bright lines in the energy spectrum reveal quantum scars that keep the system from fully thermalizing.
This explains why a simple magnetic pattern can persist for long times, offering a new handle for quantum control. (Source: arXiv:2510.02246)
Hosseinabadi’s team found scars in their (PXP)² model — but not the familiar kind. In the original PXP model, scars produce a characteristic linear growth of entanglement entropy: the system forgets, but slowly, like ice melting at a constant rate. Here, in the presence of the long-range cavity interaction, the entanglement grows logarithmically.
That is a dramatic difference. Logarithmic growth means the system’s memory decays slower and slower as time goes on — it is not just resisting thermalization, it is settling into a kind of asymptotic stubbornness. Even for large systems, the entanglement remains far below the thermal value for astronomically long times.
The team’s calculations show this for chains of up to 26 spins, starting from the Néel state — that alternating up-down pattern. The entanglement entropy curves bend only gently upward, tracing a logarithmic shape rather than a straight line. When they tune the cavity detuning Delta to specific values, they find that the entanglement growth rate peaks, suggesting resonant processes that connect the initial state to the scarred eigenstates.
This result sits in an interesting tension with the broader class of short-range scarred models studied in recent years. A separate line of work by Santis and colleagues (arXiv:2602.12152) has demonstrated that realistic cavity-coupled Rydberg arrays can be built in the laboratory — and that the blockade constraint is experimentally achievable. The theoretical prediction of logarithmic scar dynamics therefore lands on fertile ground, awaiting experimental confirmation.
Why Logarithmic Growth Is a Big Deal
To understand why logarithmic entanglement growth matters, we need to appreciate what it implies about the structure of the scarred eigenstates.
In short-range models, scars are fragile. They exist because of a subtle algebraic structure — often traced to an approximate SU(2) representation — that protects certain states from thermalizing. But the protection is imperfect, and the scars eventually dissolve into the continuum. The linear entanglement growth is a sign of this slow dissolution.
Long-range interactions change the game. When every spin can talk to every other spin through the cavity mode, the system has a kind of global awareness that short-range models lack. The scars are not just protected by a local algebraic structure; they are reinforced by the collective coupling. The logarithmic growth suggests that the system finds a way to store information in increasingly nonlocal degrees of freedom, spreading it so thin across the entire chain that local measurements cannot detect the approach to equilibrium.
The team also examined what happens when they deform the model, adding terms that break the idealized blockade constraint. Even with a modest deformation, the logarithmic signature persists — though the entanglement growth rate shows a clear maximum at a particular detuning, consistent with the resonant excitation of magnons, the quantized spin-wave excitations of the ordered phases.
This is not a fragile phenomenon. It appears to be robust to realistic imperfections, which is exactly what you want if you hope to see it in an experiment.
A natural question arises from the study’s design: does the single-mode approximation, which captures the dominant cavity effect, miss multimode phenomena that could either enhance or suppress the scars? The answer, from the broader cavity-QED literature, is that the scars are likely to survive moderate multimode effects, but the detailed entanglement dynamics could change. This is a calculation for another day — and another preprint.
The Deeper Truth
Quantum many-body scars are not just a curiosity. They challenge the foundational assumption of statistical mechanics: that a closed quantum system, left to its own devices, will thermalize.
The ETH is not a theorem in the strict mathematical sense. It is a hypothesis about what typical quantum systems do — one that has been proven for certain classes of models but remains an open question in full generality. Scars are the exceptions that probe the boundaries of typicality. They ask: under what conditions does a quantum system retain a coherent memory of its past? And what does that memory cost in terms of entanglement?
Hosseinabadi’s work provides a new answer: when interactions are long-range enough, the cost can be logarithmic — far slower than the linear growth seen in short-range scarred models. The system learns to hide its memories in the correlations between distant spins, where no local observer can easily find them.
This is not a metaphor. It is a precise mathematical statement about the growth of the von Neumann entropy with system size. But it is also — and here the Dialectic style earns its keep — a statement about the nature of quantum information. In a world where entanglement is the currency of computation, logarithmic growth is a tantalizing hint that some forms of quantum memory might be more robust than we thought.
The blockaded ferromagnet, meanwhile, reminds us that constraints — the very things that seem to limit a system’s freedom — can generate new kinds of order. Deprived of the obvious way to align, the spins find a subtler path. It is a lesson that applies far beyond this specific model.
Perhaps, in the coming years, when experimentalists in cavity-QED labs around the world finally build the Rydberg arrays that this model describes, they will not simply be confirming a phase diagram. They will be watching a quantum system walk the knife-edge between order and chaos, constrained by rules it cannot break, yet finding — through the mediating touch of light — a way to remember.
And that memory, written in the logarithmic curve of entanglement, is a whisper from the frontier of quantum statistical mechanics. A whisper that says: not all forgetting is inevitable. Some systems, it seems, were born to remember.
How This Article Was Reviewed: To provide a critical perspective, we identified the most relevant preprints cited by the authors and examined how those earlier works relate to and sometimes challenge the claims made here. No interviews were conducted. The final narrative is the editor’s own synthesis.
Yanjiang is an online editor of LoomSci
References
- Hosseinabadi et al., Kinetically constrained cavity QED: from blockaded ferromagnetism to long-range quantum scars, arXiv:2510.02246
- Marsh et al., A multimode cavity QED Ising spin glass, arXiv:2505.22658
- Santis et al., Realization of a cavity-coupled Rydberg array, arXiv:2602.12152