The Kondo effect is a many-body phenomenon: a localized magnetic impurity becomes screened by a cloud of conduction electrons that collectively conspire to form a singlet ground state. The Kondo cloud — the spatial extent of this entangled screening — can span thousands of lattice sites. Periodic driving (a Floquet perturbation) pumps energy into the system, creating inelastic scattering channels that should disrupt the delicate many-body correlations.
Herre, Karrasch, and Kennes (arXiv:2603.12844) develop a frequency-resolved functional renormalization group for periodically driven quantum impurities and find something unexpected: the Kondo cloud survives. The Kondo resonance — the sharp spectral feature at the Fermi level that signals the screening — broadens significantly under driving. But the underlying correlations that build the cloud persist. The cloud is more robust than its spectral signature.
The mechanism: periodic driving opens inelastic scattering channels at integer multiples of the drive frequency (Floquet replicas). These channels redistribute spectral weight away from the main Kondo resonance, smearing it out. But the Kondo cloud is a real-space correlation — it's about how far the screening extends, not how sharp the resonance is. The driving degrades the spectral feature without unwinding the spatial correlation.
This separation between spectral and real-space signatures matters. Most experimental probes of the Kondo effect measure spectral features — tunneling conductance, photoemission spectra. If the spectral feature broadens, the natural conclusion is that the effect is weakened. But the cloud — the actual many-body state — remains. The probe is more fragile than the phenomenon it measures. Kondo physics under driving is hidden, not absent.