Dressed Floquet Scars Give Driven Rydberg Atoms a Memory of Order

A new theory paper shows that a periodically driven Rydberg chain can hide unusually ordered many-body states inside an otherwise chaotic Floquet spectrum. For quantum-energy research, the message is practical: not every driven interacting system has to forget its initial structure and heat featurelessly.

The paper, “Dressed Floquet scars from protected zero modes in a Rydberg chain”, was submitted to arXiv on June 14, 2026 by Saptadip Roy, Bhaskar Mukherjee, K. Sengupta and Arnab Sen. The authors study a periodically driven version of the PXP model, the minimal model used to describe arrays of neutral atoms under strong Rydberg blockade. Their main result is an approximate analytic construction of two zero-quasienergy quantum many-body scars: rare states that remain structured even though they live in the middle of a spectrum where ordinary states are expected to look thermal.

That sounds abstract, but it speaks directly to a core problem in Floquet engineering. Periodic driving is the engine behind many proposals for synthetic gauge fields, topological transport, quantum batteries and time crystals. The same drive that creates useful new Hamiltonians can also dump energy into an interacting quantum system until it resembles an infinite-temperature soup. The new work identifies a protected corner of the Floquet problem where special states can retain memory of simple or highly entangled parent states across repeated drive cycles.

The headline is not that the whole driven system avoids heating. It is sharper: symmetry can protect a large zero-quasienergy subspace, and inside that subspace a few “scarred” states can continue to remember where they came from.

The old fear: Floquet heating erases everything

Floquet systems are systems whose Hamiltonian changes periodically in time. Instead of asking for energy eigenstates of a static Hamiltonian, physicists study the one-period evolution operator and its effective Floquet Hamiltonian. This viewpoint has produced some of the most exciting ideas in quantum control: shaken optical lattices, anomalous topological phases, discrete time crystals and engineered dissipative cycles.

The complication is that interacting Floquet systems are often too good at absorbing energy. In a generic isolated many-body system, the eigenstate thermalization hypothesis predicts that individual high-energy eigenstates behave thermally for local observables. Under continued driving, that tendency can push the system toward an effectively infinite-temperature state. For quantum-energy applications, that is a serious design constraint. A heat engine, battery or transport device cannot simply be “driven harder” if the useful quantum structure disappears.

Quantum many-body scars are one of the best-known loopholes. They are rare nonthermal states embedded in a much larger thermal spectrum. They do not prevent most states from thermalizing, but they allow special initial states to show revivals and memory. The modern experimental story began with a 51-atom Rydberg quantum simulator reported by Bernien, Schwartz, Keesling, Levine, Omran and collaborators in Nature in 2017, where a simple initial pattern exhibited unexpected coherent revivals. Turner, Michailidis, Abanin, Serbyn and Papić soon linked that behavior to scarred eigenstates in the PXP chain.

What Roy, Mukherjee, Sengupta and Sen add

The June 2026 paper moves the scar question into a periodically driven setting. The authors consider a ring of Rydberg atoms under the strong blockade condition: no two neighbouring atoms can both be in the Rydberg excited state. In the effective spin language, that constraint turns the Hilbert space into a restricted “no adjacent excitations” space. The drive flips the sign of a detuning-like term halfway through each period, producing a two-step Floquet unitary.

The important structure is symmetry. The model has parity symmetry, and it also has a chiral unitary symmetry that makes the Floquet Hamiltonian anticommute with the chiral operator. That anticommutation forces the quasienergy spectrum to be symmetric around zero. Combined with parity, an index-theorem argument guarantees an exponentially large collection of exact zero modes of the Floquet Hamiltonian. In the authors’ notation, the number of zero modes is bounded below by the square root of the constrained Hilbert-space dimension.

A large protected nullspace, by itself, is not enough. A typical zero mode can still be locally featureless, essentially an infinite-temperature state with zero quasienergy. The surprise is that two special zero modes are not featureless. They look like dressed versions of recognizable parent states: one descends from the simple Rydberg vacuum, and the other from a unitarily rotated version of a volume-law scar state identified by Ivanov and Motrunich in a 2025 Physical Review Letters paper.

“Dressed” means useful memory, not frozen simplicity

The word “dressed” is important. A Floquet Hamiltonian is not merely the time average of the instantaneous Hamiltonian. It contains increasingly nonlocal terms that encode micromotion within each period. So the scarred zero modes in this work are not exact copies of the parent states. They are parent states plus additional components supplied by the drive.

The authors quantify this using the overlap between a chosen parent state and the protected zero-mode subspace. If that overlap remains of order one, the projection into the nullspace can be interpreted as a dressed scar. In plain language: start from a state with a distinctive structure, repeatedly drive the system, and ask how much of that structure survives in the zero-quasienergy sector. For the two states they identify, a surprisingly large fraction can survive over ranges of drive amplitude and frequency.

To support the analytic picture, the paper combines Floquet perturbation theory with exact diagonalization. In the perturbative regime, the authors derive an effective Floquet Hamiltonian with terms that respect the chiral selection rule. They then compare this construction with numerical Floquet spectra for finite systems, including chains up to L = 30 sites in the constrained Hilbert space. The point is not that a near-term experiment must exactly match a 30-site calculation; rather, it shows that the mechanism is not just a one- or two-spin curiosity.

Why this matters for quantum energy

At first glance, many-body scars may seem far from energy technology. They are not heat engines, and the paper does not propose a battery. But Floquet quantum-energy devices depend on the same deeper capability: using time-periodic control without losing the quantum structure that made the device valuable.

Consider a Floquet quantum battery. A charging drive must put energy into the system, but useful stored work is associated with structured, non-passive states rather than featureless heat. Or consider a driven quantum thermal machine. Its cycle can rely on coherences, selection rules and engineered quasienergy gaps. If all special structure is rapidly washed out by many-body heating, the machine becomes just another noisy absorber. Scar physics offers a partial counterexample: special states can evade full thermalization even in interacting systems.

The Roy–Mukherjee–Sengupta–Sen result is especially relevant because the protection is formulated in the language of the Floquet Hamiltonian itself. It does not merely say that a static model has scars and then a weak drive perturbs them. It asks how a driven interacting model can host protected zero quasienergy modes, and how certain states remain identifiable after being dressed by the drive. That is much closer to the design logic needed for practical Floquet devices.

For quantum-energy engineers, the broader lesson is that “avoid heating” may be too blunt a design rule. A better rule is: identify the symmetry-protected subspaces where useful nonthermal structure can live.

Connections to time crystals and Floquet codes

This work also sits near two larger threads in driven quantum matter. One is the study of time crystals, where periodic driving leads to robust temporal order. A May 2026 review by Gonzalo Camacho and Benedikt Fauseweh, “Time Crystals on Quantum Devices”, argues that modern quantum processors are now probing regimes beyond the simplest closed-system pictures, including controlled, open, critical and topological realizations. Scars and time crystals are not the same phenomenon, but both ask how a periodically driven many-body system can retain order instead of simply thermalizing.

The second thread is neutral-atom hardware for quantum information. A June 2026 arXiv proposal by Han Wang, Yusheng Zhao, Xiuhao Deng and Jinguo Liu, “Measurement-Free Toric-Code Memory in Array Globally Controlled Rydberg Array”, uses multi-species Rydberg arrays and global pulses to stabilize a topological memory without local addressing or mid-circuit measurement. That is a different goal, but the hardware language overlaps: global control, blockade physics and stroboscopic protocols. Together, these papers suggest that Rydberg arrays are becoming more than quantum simulators of static Hamiltonians. They are becoming laboratories for engineered dynamical phases, memories and protected subspaces.

What to watch next

The immediate next step is experimental clarity. Can a Rydberg platform prepare one of the parent states or a close proxy, apply the two-step drive, and observe stroboscopic memory over many cycles? Can the protected nullspace be diagnosed without full state tomography? And how robust is the effect to noise, imperfect blockade, finite pulse duration and atom loss?

For the quantum-energy community, the more strategic question is whether these protected Floquet scar mechanisms can be coupled to tasks: charging, work extraction, directional transport or low-dissipation sensing. A scarred state that merely revives is scientifically beautiful. A scarred state that stores ordered energy or protects a useful current would be a design principle.

That is why this paper deserves attention at floquet.ca. It is a fundamental Floquet-engineering result, but it points toward a practical energy lesson. Periodic driving does not have to be a reckless source of heat. With the right symmetries and constraints, a driven many-body system can carry memory through the very dynamics that would normally erase it.

Sources and further reading

• Saptadip Roy, Bhaskar Mukherjee, K. Sengupta and Arnab Sen, “Dressed Floquet scars from protected zero modes in a Rydberg chain,” arXiv:2606.15605, submitted June 14, 2026. Read the arXiv abstract.
• H. Bernien et al., “Probing many-body dynamics on a 51-atom quantum simulator,” Nature 551, 579–584 (2017).
• C. J. Turner et al., “Weak ergodicity breaking from quantum many-body scars,” Nature Physics 14, 745–749 (2018).
• Gonzalo Camacho and Benedikt Fauseweh, “Time Crystals on Quantum Devices,” arXiv:2605.27211, submitted May 26, 2026. Read the arXiv abstract.
• Han Wang, Yusheng Zhao, Xiuhao Deng and Jinguo Liu, “Measurement-Free Toric-Code Memory in Array Globally Controlled Rydberg Array,” arXiv:2606.12030, submitted June 10, 2026. Read the arXiv abstract.