Electromagnetic resonances are fundamental to nanophotonics: plasmon resonances in metal nanoparticles, Mie resonances in dielectric spheres, Fano resonances in coupled systems. Each resonance appears as a peak or dip in the scattering spectrum. The standard interpretation: each spectral feature corresponds to an eigenmode of the structure.
Binkowski, Betz, Colom, and colleagues show this interpretation is often wrong. They construct a unified framework that separates resonances-as-eigenmodes (properties of the structure in isolation) from resonances-as-spectral-features (properties of the measured scattering cross-section). The two are related but not identical, and the discrepancy is not noise — it's interference.
A single eigenmode can produce multiple spectral features through its interaction with the background scattering. Conversely, multiple eigenmodes can interfere to produce a single spectral feature that looks like one resonance but is actually several. The scattering cross-section is the squared modulus of a sum of complex amplitudes, and the modulus-squared operation creates interference terms that reshape the spectrum.
Fano resonances are the canonical example: the asymmetric line shape arises from interference between a narrow resonance and a broad background, producing a spectral feature that doesn't correspond to any single eigenmode. But the phenomenon is general — any measurement that involves coherent superposition of resonance amplitudes produces spectral features that reflect the measurement setup as much as the structure.
The spectral feature you measure is not the eigenmode you think it is. It's the eigenmode plus the interference plus the background. The observed resonance is an artifact of the observation.