Floquet Thermodynamics Gets a Cleaner Energy Ledger

Floquet engineering promises something deeply practical: use a repeating drive, such as a laser, microwave tone, gate voltage or mechanical modulation, to make a quantum system behave as though it has new energy levels. That idea powers proposals for light-shaped materials, quantum heat engines, thermal diodes and quantum batteries. But there is a quieter problem behind nearly every one of those proposals. When the driven system leaks energy into an environment, how do we keep the bookkeeping honest?

A new preprint posted to arXiv on July 31, 2026 tackles that question directly. Luísa T. Tude, Carlos Ortega-Taberner, Roberta Zambrini and Gonzalo Manzano, from IFISC in Spain and Trinity College Dublin, study “Dissipation in Periodically Driven Quantum Systems: Partial Secularization and Thermodynamic Consistency.” The title is technical. The core message is not: if we smooth away too much of the fast Floquet motion, a model can predict energy currents that look mathematically tidy but physically wrong.

In driven quantum thermodynamics, a beautiful master equation is not enough. It must also give a sensible ledger for heat, work, entropy and the energy supplied by the drive.

The paper matters because many Floquet energy ideas live in the open-system regime. Real quantum devices are not isolated. A qubit is coupled to a resonator, a bath, a readout line, a substrate, a phonon environment or a deliberately engineered reservoir. If the theory used to describe that leakage erases the terms that carry work from the periodic drive, researchers can overstate, understate or misclassify a device’s performance.

The master-equation shortcut under the microscope

Open quantum systems are often described with master equations. Instead of tracking every microscopic atom in a bath, the equation follows the system’s density matrix while summarizing how the environment causes relaxation, excitation and decoherence. The widely used Gorini–Kossakowski–Sudarshan–Lindblad, or GKSL, form is prized because it preserves physical probabilities and complete positivity. In plainer language, it keeps the quantum state from wandering into mathematical nonsense.

For periodically driven systems, researchers commonly combine Floquet theory with a Born-Markov weak-coupling treatment of the environment. Floquet theory rewrites the driven system in terms of quasienergies and sidebands: transitions can occur not only at ordinary energy gaps, but at gaps shifted by integer multiples of the driving frequency. The standard route then applies a secular, or rotating-wave, approximation that discards rapidly oscillating terms.

That approximation is useful. It is also dangerous when used indiscriminately. Tude and colleagues focus on the full secular approximation, which strongly filters the Floquet-Redfield equation until the result has a clean GKSL structure. The authors show that this can remove oscillatory contributions needed to compute steady-state energy currents consistently. In particular, the full secular approximation can force the average mechanical power supplied by the drive to vanish in stationary conditions. For a genuinely driven heat engine, refrigerator or energy-transfer device, that cannot be generally right.

Partial secularization: a middle path

The alternative proposed in the paper is a coarse-grained master equation. Instead of keeping every fast term, as a Redfield equation does, or discarding too much, as the full secular approximation may do, the coarse-grained approach uses a tunable time window. Terms that vary too quickly for the chosen temporal resolution are filtered. Terms that still matter for the resolved dynamics remain.

This is the heart of partial secularization. The coarse-graining time is not merely a mathematical knob. It has a physical interpretation: it sets the time resolution at which a Markovian master equation is allowed to describe the periodically driven system. If the chosen resolution is longer than certain microscopic memory effects, the result can be both manageable and physical. If the relevant system dynamics unfold on comparable or shorter times, the paper’s criterion signals that a non-Markovian treatment may be needed.

That balance is valuable for quantum-energy research. Many proposed devices operate in exactly the uncomfortable zone where a drive is strong enough to do useful work, but not so fast or so weak that every sideband and coherence can be safely averaged away. The paper explicitly notes that quasienergy near-degeneracies and crossings are common, especially as Hilbert-space dimension grows or when the drive is not overwhelmingly fast compared with the system’s natural frequency.

Why energy currents are the hard part

It is easy to say that a driven quantum device exchanges heat with baths and work with a drive. It is harder to compute those quantities without double-counting or hiding energy flows. The new paper emphasizes that the definition of heat and work currents must be consistent with the approximations used to derive the dynamics.

One clear example is a periodically driven two-level system coupled to a thermal reservoir. In the full secular approximation, diagonal and off-diagonal density-matrix elements decouple in the Floquet basis. Coherences vanish in the long-time limit. But the drive’s mechanical power can depend precisely on coherences in the relevant rotating frame. If those coherences are erased, the model can miss the work being exchanged with the drive.

The authors compare the full secular master equation, the coarse-grained approach, the Floquet-Redfield equation and numerically exact non-Markovian simulations. That comparison is important: it turns the paper from a formal complaint into a practical diagnostic. The point is not that every old Floquet master equation is useless. The point is that researchers should know when a convenient approximation is outside its thermodynamic comfort zone.

For a Floquet heat machine, losing the wrong oscillating term is like removing a line item from an accounting sheet: the final balance may still add up, but the business is no longer being described honestly.

The three-level maser test

The paper’s second major example is a three-level maser coupled to hot and cold thermal reservoirs. This model traces back to the famous 1959 Scovil and Schulz-DuBois maser heat-engine proposal, one of the conceptual ancestors of quantum heat-engine theory. It is a natural stress test because it links thermal reservoirs, coherent driving and work extraction in a compact setting.

In the heat-engine regime, the device absorbs heat from the hot bath, dumps heat into the cold bath and delivers work through the driven transition. The authors examine how different master-equation treatments predict steady-state populations, coherences, heat currents, power and efficiency. They report that the full secular master equation can reproduce some population trends while still predicting vanishing coherences. Those missing coherences affect the computed power.

The efficiency discussion is especially useful for the beyond-Carnot conversation. The paper does not claim to beat Carnot efficiency. Instead, it shows that the entropy-production expression remains consistent and that efficiencies are bounded by the Carnot value when heat and work are treated properly. The more subtle result is that different approximations can give different power-efficiency curves. Near maximum output power or near resonance, the convenient full secular picture can overestimate power or miss the correct curve even when the final efficiency appears superficially similar.

What this means for practical Floquet energy devices

For non-specialists, the practical analogy is a camera shutter. If the shutter is open too long, fast motion blurs out. If it is too short, the image may be noisy and hard to interpret. Partial secularization chooses a physically meaningful exposure time for the driven open quantum system. The resulting picture is not infinitely detailed, but it keeps the motion needed to account for energy exchange.

That matters across Floquet.ca’s core themes:

• Quantum heat engines: power and efficiency estimates depend on how drive work and bath heat currents are separated.
• Floquet thermal diodes and transistors: rectification, amplification and noise can be distorted if relevant coherences are averaged away.
• Quantum batteries: charging protocols often rely on periodic driving and open-system effects, so injected work and dissipated heat must be tracked consistently.
• Light-driven materials: dissipation is not an afterthought; it influences whether a driven phase is stable, useful or simply heated away.

The broader lesson is methodological. The closer Floquet energy research moves toward devices, the less acceptable it becomes to report only elegant quasienergy spectra or idealized closed-system dynamics. Engineers will ask where the heat goes, how much work the drive supplies, what entropy is produced and whether the model remains physical over the relevant time scale. This paper gives theorists a sharper tool for answering those questions.

A sober advance, not a magic loophole

Because Floquet systems can create sidebands and effective Hamiltonians that do not exist in equilibrium, it is tempting to frame every new result as a route to “beyond Carnot” performance. The responsible reading is different. Floquet driving can open new control channels, but the drive is itself an energy and information resource. A thermodynamically consistent model must include that resource rather than pretending it is free.

That is why the new partial-secularization work is important. It does not sell a spectacular device claim. It repairs the measurement stick. Better thermodynamic accounting may sound less glamorous than a record efficiency or a new topological phase, but it is essential if quantum-energy devices are to graduate from diagrams to hardware. A theory that knows when it is allowed to be Markovian, when it must keep coherences and when it should defer to non-Markovian simulation is a theory ready for harder engineering questions.