Sparse Networks Reveal a Hidden Floquet Heating Trap

A new theory paper argues that the stability of driven quantum matter can depend as much on who is connected to whom as on the drive frequency itself. In programmable systems with sparse long-range links, Floquet heating may be dramatically faster than the familiar local-lattice rule of thumb suggests.

The paper, “Anomalous Floquet Heating from Sparse Long-Range Interactions”, was posted to arXiv on June 23, 2026 by Chenyue Guo, Andrea Pizzi and Hongzheng Zhao. It studies periodically driven interacting spins on small-world networks: mostly local lattices with a fraction of links rewired into distant shortcuts. The authors find that even sparse infinite-range connections can change the heating law from the standard high-frequency form, roughly γ ∼ exp[-O(ω)], to a slower stretched-exponential suppression, γ ∼ exp[-O(√ω)].

That mathematical change is not cosmetic. Floquet engineering relies on a practical bargain: drive a system rapidly enough that it behaves for a long time as if governed by a useful effective Hamiltonian, before heating ruins the state. If the lifetime improves exponentially with drive frequency, the bargain can be excellent. If network topology downgrades that protection to a stretched exponential, the same frequency increase buys less stability than expected.

The new warning is simple: in a programmable quantum simulator, a few long-distance shortcuts can create high-connectivity “hot spots” that dominate the entire Floquet heating process.

Why Floquet heating matters for quantum energy

Floquet engineering uses periodic driving—laser pulses, microwave fields, voltage modulation, lattice shaking or gate sequences—to reshape a system’s dynamics. The technique can produce synthetic magnetic fields, dynamical topology, time-crystal order, engineered transport channels and effective interactions that are hard or impossible to obtain statically.

The cost is energy absorption. A closed many-body system under a periodic drive generally tends toward a featureless infinite-temperature state. In plain language: if the drive keeps shaking the system, eventually the shake can become heat. Much of modern Floquet theory is about delaying that failure long enough for the engineered state to be useful.

For systems with local interactions, a powerful result emerged over the past decade. At high drive frequency, energy absorption can be exponentially slow. Foundational work by Abanin, De Roeck and Huveneers; by Mori, Kuwahara and Saito; and by others established rigorous and practical versions of this idea. The system enters a prethermal plateau: a long-lived nonequilibrium regime where it approximately obeys a static effective Hamiltonian before eventual thermalization.

That plateau is central to quantum-energy thinking. Quantum batteries need useful stored ergotropy before decoherence and leakage erase it. Heat engines and refrigerators need controlled cycles before excess dissipation dominates. Floquet materials need coherent band reshaping before the pulse becomes ordinary heat. In each case, the question is not whether heating exists; it is whether the useful window is long enough.

The small-world twist

Guo, Pizzi and Zhao ask what happens when the interaction graph is neither a regular local lattice nor a fully connected mean-field model. Their main example is a small-world network. Start with a square lattice, then rewire each bond with probability p so that some connections jump to distant sites. The average number of bonds per site stays fixed, but the distribution becomes uneven: most sites remain modestly connected, while a few become hubs with unusually high coordination number.

This is close in spirit to what modern quantum platforms are beginning to offer. Rydberg tweezer arrays can be reconfigured; cavity-mediated interactions can couple distant atoms through photonic modes; trapped ions and superconducting devices can implement interaction graphs that are not simple nearest-neighbor chains. For quantum computing, sensing and nonequilibrium materials research, that flexibility is a feature. The new paper shows it can also become a heating liability.

The authors use a periodically driven classical-spin model, chosen because it allows very large simulations over long times. They emphasize that classical and quantum Floquet heating often share qualitative mechanisms, while noting that fully quantum many-body confirmation remains challenging. The Hamiltonian alternates between an Ising interaction step on the graph and a transverse-field step. By tracking stroboscopic energy density, they measure the prethermal lifetime before heating becomes significant.

What they found

On the regular lattice, the expected result returns: the prethermal lifetime grows exponentially with drive frequency. On the small-world network, however, the lifetime follows a different scaling. The paper reports τ ∼ exp[√(ω/J0) ln(ω/J0)] for the thermalization time in the high-frequency regime, corresponding to anomalous heating γ ∼ exp[-O(√ω)] rather than exp[-O(ω)].

The mechanism is intuitive once stated. A site with more neighbors has a larger local energy scale. Larger local energy scales open more ways to absorb energy from the drive. In a regular lattice, every site has the same coordination number, so the high-frequency drive is off-resonant in a fairly uniform way. In a small-world network, rare high-degree sites act differently. As the drive frequency increases, the heating is increasingly dominated by sites with larger coordination numbers. The hubs are statistically rare, but their contribution wins because they absorb energy so much more efficiently.

Heating is no longer governed by the average lattice site. It is governed by rare, highly connected sites whose local energy scale rises with coordination number.

A crucial control test strengthens the claim. The authors report that random regular graphs, which include long-range links but keep a fixed coordination number, heat similarly to regular lattices. So the culprit is not nonlocality by itself. It is the broad distribution of coordination numbers: the coexistence of ordinary sites and hubs.

Not all bad news: topology as a control knob

The paper is not merely pessimistic. It also shows that long-range shortcuts can stabilize useful order during the prethermal window. In their model, small-world connectivity can raise the effective ordering tendency of the Floquet Hamiltonian. The same initial state that loses magnetization on a regular lattice can retain finite magnetization on the rewired network.

The authors then demonstrate a prethermal period-4 discrete time crystal by adding a π/2 spin rotation after each drive period. In the small-world case, the magnetization follows a robust four-cycle pattern—z, y, −z, −y, then back to z—whereas the regular lattice quickly loses the needed order. The lifetime of this time-crystalline response increases with drive frequency, even though the heating law is anomalous.

Why this connects to practical energy devices

Quantum energy technologies are unlikely to be perfect textbook lattices. They will use chips, resonators, photonic buses, cavities, gates, defects, synthetic dimensions and feedback channels. These ingredients create structured interaction networks. Some links will be local; others will be effectively long range. Some elements will be ordinary; others will become hubs.

The new result says that an energy budget based only on average coupling strength can miss the dominant loss channel. In a driven quantum battery, a strongly connected module might charge quickly but also become the first place where drive energy turns into unwanted heat. In a Floquet simulator, a hub qubit or cavity mode might destabilize an engineered phase. In a light-driven material, a small population of strongly coupled modes might determine whether the pulse produces coherent work-like control or broad heating.

That insight points toward new engineering questions. Can we design interaction graphs that keep the beneficial ordering effects of shortcuts while suppressing high-degree hot spots? Can dissipation be added locally to cool hubs without destroying the prethermal state? Can drive frequencies, amplitudes and pulse shapes be adapted to the actual degree distribution rather than to a single average energy scale?

The open questions

Several cautions are important. The main simulations are classical, not fully quantum. The model is deliberately minimal, not a device blueprint. The stretched-exponential law is an asymptotic statement, and finite systems may show crossovers. Real platforms also include noise, decoherence and control imperfections that can either worsen or mask the mechanism.

Still, the paper identifies a mechanism that is difficult to ignore: network topology can reshape the frequency dependence of Floquet heating. That makes it relevant for any platform where researchers celebrate programmability of interactions. Programmability means the graph is a design variable. If the graph changes the heating law, it belongs in the energy accounting.

For Floquet.ca’s broader theme, the message is disciplined optimism. Beyond-Carnot and quantum-energy research does not succeed by pretending dissipation disappears. It succeeds by learning where energy flows, which channels are useful, which channels are wasteful, and how coherent driving changes those tradeoffs. Sparse-network heating is a reminder that the “wiring diagram” of a quantum system can be as important as the waveform driving it.

The takeaway

Guo, Pizzi and Zhao have added a new variable to the Floquet stability playbook. Frequency still matters. Coupling strength still matters. Disorder, dimensionality and drive protocol still matter. But for programmable long-range systems, degree distribution may decide whether high-frequency driving buys a long exponential lifetime or a more fragile stretched-exponential one.

If future quantum simulators, quantum batteries and driven materials are built from reconfigurable networks, this is exactly the kind of result engineers will need. It does not say “avoid long-range links.” It says: use them knowingly. The same sparse shortcuts that help create exotic nonequilibrium order can also mark the places where Floquet energy leaks first.

Sources and further reading

• Chenyue Guo, Andrea Pizzi and Hongzheng Zhao, “Anomalous Floquet Heating from Sparse Long-Range Interactions”, arXiv:2606.24541, submitted June 23, 2026.
• Marin Bukov, Luca D’Alessio and Anatoli Polkovnikov, “Universal high-frequency behavior of periodically driven systems: from dynamical stabilization to Floquet engineering,” Advances in Physics 64, 139–226 (2015).
• Dmitry A. Abanin, Wojciech De Roeck and François Huveneers, “Exponentially Slow Heating in Periodically Driven Many-Body Systems”, Physical Review Letters 115, 256803 (2015).
• Takashi Mori, Tomotaka Kuwahara and Keiji Saito, “Rigorous Bound on Energy Absorption and Generic Relaxation in Periodically Driven Quantum Systems”, Physical Review Letters 116, 120401 (2016).
• A. Pizzi, J. Knolle and A. Nunnenkamp, “Higher-order and fractional discrete time crystals in clean long-range interacting systems,” Nature Communications 12, 2341 (2021).