Programming the Quantum Connection: Floquet Control Between Distant Qubits

Two quantum components can share a communication channel without being easy to control independently. Connect superconducting qubits through a microwave transmission line and the same electromagnetic environment that links them can also make them exchange energy or radiate together. The challenge is not simply to establish a connection. It is to decide when that connection should be strong, weak, coherent or dissipative.

A new experimental preprint, submitted to arXiv on October 12, 2026, reports a Floquet approach to that problem. Y.-H. Huang and colleagues periodically modulate two superconducting artificial atoms in a mirror-terminated waveguide, changing their photon-mediated interaction through the drive frequency, amplitude and relative phase. The result is a programmable connection between distant quantum emitters, rather than a new source of energy. [1]

For quantum-energy research, the significance is practical but preliminary: future devices must control not only where energy is stored, but also when it moves and when it escapes. This experiment investigates hardware-level control of precisely those interactions. It does not demonstrate a quantum battery, a heat engine or beyond-Carnot efficiency.

The advance is control over the route energy can take—not evidence that more energy comes out than goes in.

A fixed circuit with a changeable interaction

The device contains two superconducting qubits connected through a one-dimensional transmission line. A capacitive termination acts as a mirror for the electromagnetic field, producing a standing-wave pattern. One qubit sits at the mirror; the other is located approximately 60 millimetres away along the waveguide. Both belong to a single-chip device, so “distant” here does not mean separate buildings or a metropolitan quantum network. [1]

A standing wave has nodes, where its field vanishes, and antinodes, where it is strongest. Think of the stationary pattern on a vibrating string, except that the relevant wave is electromagnetic. The mirror qubit sits at an antinode. The distant qubit can effectively occupy a node or an antinode at its operating frequency.

No component has to slide across the chip. Changing the qubit transition frequency changes the wavelength associated with that transition, and therefore its position within the standing-wave pattern. This selects between two qualitatively different ways for the qubits to interact.

Coherent exchange versus collective loss

In the node configuration, the distant qubit is weakly coupled to the resonant field, yet an interaction mediated by off-resonant, virtual-photon processes survives. The two qubits acquire a collective frequency shift. In spectroscopy, their interaction produces an avoided crossing: rather than passing directly through one another, the resonances split apart.

In the antinode configuration, both qubits couple strongly to the resonant waveguide field. Their emission becomes correlated, creating collective decay channels. Some collective states radiate more strongly, while others radiate less strongly than one would expect from independent emitters. These are the ideas behind superradiance and subradiance, established concepts reviewed extensively in waveguide quantum electrodynamics. [2]

Floquet driving adds a control dial

The static geometry establishes the starting interaction. Floquet engineering adds a repeated modulation on top. In this experiment, a sinusoidal magnetic-flux drive changes a qubit’s transition frequency in time. The frequency is therefore not simply moved to a new constant value; it oscillates around its working point. [1]

Periodically driven systems develop a ladder of spectral features called Floquet sidebands. These are separated by the modulation frequency and describe additional drive-assisted channels through which the system can respond. Sidebands alone would not prove useful interaction control: a device could display extra spectral lines while leaving the connection of interest largely unchanged.

The important claim in the new work is stronger. The periodic drive changes the effective coupling between the qubits themselves. Its strength follows a Bessel-function dependence on modulation parameters. A Bessel function is an oscillating mathematical function with special zero crossings; here, those zeros identify settings where a particular effective interaction can be strongly suppressed.

When only one qubit is driven, the authors track collapses and revivals of the spectral splitting as the modulation is varied. Their interpretation is destructive interference between different Floquet-assisted exchange pathways. The qubits remain physically connected to the same waveguide, but the pathways can cancel in the effective interaction. The paper identifies this as coherent destruction of tunnelling in a waveguide-QED setting.

The cancellation should not be read as a universal disappearance of every possible loss channel. It targets the interaction described by the model and the measured spectral response. Independent decoherence, imperfections and other environmental couplings still matter when turning such a control point into a useful operation.

Why the timing between two drives matters

Driving both qubits introduces an especially useful variable: relative phase. Two sinusoidal signals can have the same frequency and amplitude yet reach their peaks at different times. That timing offset changes the relative modulation experienced by the qubits.

For the matched driving conditions discussed in the paper, in-phase modulation restores the static coupling. Moving away from that timing changes the effective modulation difference and produces suppression and revival of the interaction. The researchers report phase-dependent spectral responses consistent with their theoretical description. [1]

The attraction is that the interaction can be varied without changing the qubits’ mean frequencies, their physical separation or the pump power. Frequency tuning is already an important control technique, but it can disturb other resonances in a larger circuit. Phase offers another handle when a device designer wants to preserve those operating points.

Keeping pump power fixed while changing an interaction is a useful control advantage. It is not a measurement of lower total energy consumption.

This distinction is particularly important for Floquet technologies. A timing adjustment may be inexpensive relative to rebuilding a circuit, yet the drive electronics still consume power. Whether a phase-controlled protocol is energetically better than an alternative requires a separate, measured comparison.

What the experiment actually establishes

The researchers probe the device using a weak coherent microwave field and measure its reflection spectrum. Resonance positions, splittings and linewidths provide evidence about how the qubits interact. They compare these responses with a theoretical model and numerical simulations, including additional measurements in the supplemental material.

The evidence is not equally clean for every observable. The paper explicitly reports a weak stray resonance that obscures the direct linewidth signal in part of the dissipative-coupling data. Numerical spectra with that background removed help reveal the predicted modulation of collective decay. That caveat matters: a clear collapse of coherent splitting and a model-supported change in collective linewidth should not be presented as identical measurements.

• Reported experiment: microwave spectroscopy of periodically driven superconducting qubits in a mirror-terminated waveguide.
• Central result: tunable photon-mediated interactions, including phase-dependent control, with coherent and dissipative configurations selected by the operating point.
• Evidence boundary: some dissipative features require interpretation alongside numerical modelling because of a stray resonance.
• Publication status: an arXiv experimental preprint; the cited version is not being presented here as a peer-reviewed journal publication.

The distinction between selecting a configuration and modulating it is also important. The standing-wave operating point determines the coherent or dissipative character of the baseline coupling. The periodic drive then dresses that interaction. It would overstate the result to say that changing only the pump phase arbitrarily converts any coherent connection into any desired dissipative one.

Where quantum-energy applications could begin

The following connections are research possibilities, not applications demonstrated in this paper. They help explain why a quantum-energy hub should follow an experiment whose immediate emphasis is quantum control and networking.

1. Connect a storage element only when needed

A quantum battery needs a charging stage, a holding stage and a way to deliver useful energy. A permanently open connection can complicate storage by allowing unwanted exchange. A controllable interaction could potentially connect a charger to a storage element during one stage and suppress their exchange during another.

Testing that idea would require measuring stored energy and extractable work, not just spectral splitting. Researchers would also need to show that switching does not introduce losses larger than those it prevents. The present two-qubit experiment supplies a possible control ingredient, not that full storage-and-release demonstration.

2. Treat emission as something to schedule

Collective decay is usually discussed as a problem when the goal is preserving a quantum state. In other tasks, emission is the desired output: a device may need to release a microwave excitation into a transmission line at a particular time.

Floquet control of collective radiative behaviour could help investigate the balance between retaining and releasing an excitation. A convincing energy application would measure what reaches a load or receiver, together with losses into other channels. A narrower resonance or longer-lived mode by itself is not a delivered-energy metric.

3. Build cleaner tests of open-system thermodynamics

Quantum thermodynamics studies systems that exchange energy with drives and environments. Having experimental control over coherent interactions and correlated decay can make it easier to test models of such exchanges. It may also support simulations in which dissipation is a designed ingredient rather than uncontrolled noise.

But a waveguide is not automatically a hot or cold thermal reservoir merely because energy can flow into it. Demonstrating a heat engine requires identifying reservoirs and their states, measuring heat and work consistently, and defining a cycle or steady-state operation. Those are additional experiments.

Why this is not a beyond-Carnot result

Carnot’s bound concerns heat-to-work conversion between equilibrium thermal reservoirs under its specified assumptions. This study concerns driven quantum interactions and their spectroscopic signatures. It reports neither a heat-engine efficiency nor a comparison against a Carnot benchmark.

Periodic modulation is an external resource. In any proposed energy-conversion device derived from this platform, the drive’s contribution would have to be included alongside energy entering and leaving through the waveguide. State preparation, readout and the supporting cryogenic apparatus would also matter for a system-level engineering assessment.

Likewise, references to vacuum-mediated coupling do not imply extraction of unlimited vacuum energy. They identify the quantum-electrodynamic mechanism behind the interaction. A change in a spectral line or an exchange pathway is not evidence of net work obtained without supplying a resource.

The next convincing milestones

For this approach to move from interaction spectroscopy toward a useful quantum-energy component, several tests would be especially informative:

• Time-resolved switching measurements showing how rapidly and reproducibly a connection can be suppressed and restored.
• Direct excitation-transfer or release measurements, with a receiver or load rather than only a reflected probe.
• Tests of robustness against phase noise, calibration drift and unwanted resonances.
• Energy accounting that includes the modulation source and separates device-level performance from supporting infrastructure.
• Scaling beyond two qubits without losing selective control or introducing excessive cross-talk.

The near-term achievement is already meaningful without turning it into a power-generation claim. Fixed hardware no longer has to imply a fixed photon-mediated interaction. By programming a rhythm—and the timing between two rhythms—researchers can reshape how quantum components communicate, exchange excitations and radiate. Whether that flexibility produces better energy devices is now a question for measured protocols, not a conclusion supplied by Floquet sidebands alone.