Work Statistics Turn Floquet Heating into an Energy Diagnostic

Floquet engineering is often described as a way to make new quantum phases by shaking a system at the right rhythm. But every driven system also has an energy accounting problem: where does the drive’s work go, how fast is it absorbed, and when does useful coherent motion turn into featureless heat?

A new arXiv preprint by Feng-Li Lin and Ching-Yu Huang, “Subsystem Thermalization and Work Statistical Characterizations of Floquet Dynamics”, posted July 1, 2026, addresses that problem directly. The authors study a periodically driven non-integrable Ising spin chain and compare two diagnostics that are usually discussed in different languages. One is subsystem thermalization: do small parts of the system look thermal after many drive cycles? The other is quantum work statistics: what distribution of work is performed during a Floquet cycle, and does it obey fluctuation-theorem tests associated with thermal behavior?

The key contribution is a bridge: local thermalization and cycle-by-cycle work statistics encode the same crossover from heating, to prethermal plateaus, to high-frequency finite-temperature behavior.

Why this matters for quantum-energy research

For quantum engines, quantum batteries and Floquet materials, “heating” is not merely a nuisance. It is the shadow side of control. A periodic drive can open topological gaps, dress bands, stabilize synthetic phases or charge a many-body battery. The same drive can also dump energy into unwanted degrees of freedom. Practical devices need more than a phase diagram; they need diagnostics that say whether the drive is storing useful work-like energy, maintaining coherent response, or simply heating the system toward an infinite-temperature state.

The Lin–Huang paper is theoretical, not an experiment. Still, its diagnostic logic is experimentally relevant because modern quantum simulators increasingly measure both local reduced states and energy-transfer statistics. Superconducting qubits, trapped ions, Rydberg arrays and cold-atom platforms can all implement periodic pulses, reconstruct small-subsystem density matrices, and in some cases perform two-point-measurement-style protocols or interferometric measurements of characteristic functions.

The model: a driven spin chain as a clean test bed

The paper considers an open Ising spin chain with two non-commuting pieces. One part contains Ising interactions and a longitudinal field; the other is a transverse-field term. During each period, the Hamiltonian is driven in a bang-bang fashion, alternating between these pieces. Repeating the same pulse sequence defines a Floquet operator, the one-cycle map that advances the quantum state stroboscopically.

This is a useful model because it is simple enough for exact diagonalization at modest sizes, but complicated enough to be non-integrable and thermally interesting. The authors examine how an initial state evolves after many cycles and ask what thermal state, if any, best represents it. Depending on the frequency, the same system can look like three different machines. At low frequency, it absorbs energy strongly and tends toward an infinite-temperature state. At high frequency, the drive is too fast to be absorbed efficiently, so an effective Hamiltonian can describe a finite-temperature state. In between, the system develops a long-lived prethermal plateau: not fully thermal, not fully cold, and very important for applications because it is the window in which engineered dynamics can survive.

Local thermalization: what a small subsystem sees

Subsystem thermalization asks a humble question: if we look at only one spin, or a small block of spins, does it resemble the corresponding part of a thermal state? This avoids demanding that the entire closed quantum system become a mixed thermal ensemble. Globally, unitary evolution preserves information. Locally, however, entanglement can make a small piece look thermal because information about its initial condition is spread across the rest of the chain.

Lin and Huang quantify this using reduced density matrices and relative entropy. They compare the state of a subsystem after many Floquet cycles with the reduced state of an appropriate thermal ensemble. When the relative entropy becomes small, the subsystem is effectively thermal with respect to that chosen Hamiltonian and temperature. In their calculations, a single-site subsystem and a half-chain subsystem both expose how the answer depends on drive frequency and on which effective Hamiltonian is used as the thermal reference.

That last detail is important for energy technology. A driven device does not always thermalize with respect to its bare Hamiltonian. In a Floquet system, the relevant energy bookkeeper may be the average Hamiltonian or a more subtle effective Hamiltonian generated by the whole pulse sequence. Choosing the wrong bookkeeper can make a device appear wasteful, coherent or thermal for the wrong reason.

Work statistics: the global energy ledger

The second diagnostic comes from quantum thermodynamics. In classical thermodynamics, work is a number. In quantum mechanics, work is not a simple observable that can be read at one instant. The standard operational method is the two-point measurement protocol: measure energy before the process, drive the system, measure energy again, and define work as the energy difference. Repeating the protocol builds a probability distribution.

The authors analyze the characteristic function of work both with and without the two-point-measurement construction. This distinction matters because projective energy measurements remove certain quantum coherences. Comparing the two versions reveals how much coherent structure is still present in the driven state. If the two descriptions converge and the relevant fluctuation theorem behaves as expected, the system is behaving thermally in a stronger sense. If they differ, the energy ledger is still carrying quantum coherence.

In this framing, work statistics are not just about average absorbed energy. They are a spectroscopy of coherence, thermality and the thermodynamic cost of the drive.

The paper also uses fluctuation-theorem logic, including the Jarzynski relation for cyclic processes, as a global test of thermal behavior. The Jarzynski relation states, in one common form, that a suitable exponential average of work equals a free-energy factor; for a cyclic process that begins in a thermal state, the relevant benchmark reduces to a sharp constraint. Deviations from that constraint reveal that the system is not globally thermal in the assumed sense, even if small subsystems look close to thermal.

The crossover map

Putting the two diagnostics side by side gives the article its value. Subsystem thermalization is local and state-based. Work statistics are global and process-based. The preprint shows that both track the same frequency-dependent story. At low frequency, strong absorption drives the system toward infinite-temperature behavior. In the intermediate regime, coherent quantities and thermal-fit measures reveal a prethermal crossover. At high frequency, the system can settle into a finite-temperature description tied to the effective Floquet Hamiltonian.

This unified picture helps researchers avoid a common trap: declaring victory from only one diagnostic. A local block of spins may look thermal even while global coherence or fluctuation-theorem tests reveal nontrivial structure. Conversely, work statistics may show coherent energy absorption that is invisible to a single local observable. Energy devices need both views because useful performance lives in the gap between microscopic control and macroscopic thermodynamic accounting.

How this connects to Floquet materials and batteries

The immediate study is a spin-chain simulation, but the message is broader. Floquet materials researchers want drives that reshape bands without uncontrolled heating. Quantum-battery researchers want drives that deposit extractable energy rather than random heat. Quantum heat-engine researchers want cycle protocols where work, heat and coherence can be separated experimentally. The Lin–Huang framework gives these communities a shared measurement vocabulary.

For example, a periodically driven quantum battery may show a high average stored energy after many cycles. That alone does not prove useful charging if the state is passive or nearly infinite-temperature. Work statistics and fluctuation-theorem deviations can help distinguish coherent, potentially extractable energy from mere heating. Similarly, a Floquet topological material may exhibit the desired band dressing over a prethermal window. Subsystem thermalization can help estimate how long the engineered state remains locally meaningful, while work statistics reveal whether the drive is quietly eroding the state through global energy absorption.

What is new—and what remains open

The paper does not claim to beat Carnot, build a working engine, or stop Floquet heating. Its contribution is more foundational: it shows that two operationally meaningful diagnostics can be used together to characterize how driven many-body systems absorb, store and thermalize energy. That is a necessary step toward reliable driven quantum machines.

Several open questions follow naturally. Can the framework be scaled to larger systems with tensor-network or quantum-simulator data? Can experimental platforms measure the relevant work characteristic functions with sufficient accuracy over many cycles? Can the diagnostic identify not just when heating occurs, but which control modifications reduce it? And can similar methods distinguish useful ergotropy in quantum batteries from heat-like stored energy?

The broader trend is clear. Floquet engineering is moving from “make the phase appear” toward “quantify the energy flow that makes the phase possible.” That shift is essential for any practical quantum energy research hub. A drive is not only a clock; it is a power supply. Work statistics tell us what the power supply is really doing.

Sources and further reading

• Feng-Li Lin and Ching-Yu Huang, “Subsystem Thermalization and Work Statistical Characterizations of Floquet Dynamics”, arXiv:2607.00381 (2026).
• Achilleas Lazarides, Arnab Das and Roderich Moessner, “Equilibrium states of generic quantum systems subject to periodic driving”, Physical Review E 90, 012110 (2014).
• Luca D’Alessio and Marcos Rigol, “Long-time behavior of isolated periodically driven interacting lattice systems”, Physical Review X 4, 041048 (2014).
• Dmitry A. Abanin, Wojciech De Roeck, Wen Wei Ho and François Huveneers, “Effective Hamiltonians, prethermalization, and slow energy absorption in periodically driven many-body systems”, Physical Review B 95, 014112 (2017).
• Christopher Jarzynski, “Nonequilibrium Equality for Free Energy Differences”, Physical Review Letters 78, 2690 (1997).
• Peter Talkner, Eric Lutz and Peter Hänggi, “Fluctuation theorems: Work is not an observable”, Physical Review E 75, 050102 (2007).