Surface acoustic waves propagate along materials. When they hit a boundary — a new layer, a change in properties — some energy reflects, some transmits, some converts. In the usual design approach, controlling surface waves requires structuring the surface: etching patterns, attaching resonators, engineering metamaterials with periodic microstructures.
Coupel, Bonello, and co-authors (arXiv:2602.20692) show that a thin layer weakly bonded to a substrate creates tunable frequency bands of strong wave attenuation — without any surface structuring at all. The mechanism is resonance: the weakly bonded layer oscillates as a mass on a spring, where the spring constant is set by the adhesion strength. At the resonant frequency, energy from the surface wave feeds into the layer's oscillation and is absorbed.
The key parameter is adhesion. Stronger bonding raises the resonant frequency and narrows the attenuation band. Weaker bonding lowers it and broadens it. The weakness of the bond is the functional element. A perfectly bonded layer would simply extend the substrate — no resonance, no attenuation, no control. It is the imperfection of the connection that creates the useful frequency response.
This inverts the usual engineering logic. Bonding is typically something to optimize — stronger is better, perfect adhesion is the goal. Here, the degree of imperfection is the design parameter. The attenuation band exists because the bond is weak enough to allow relative motion between layer and substrate. A perfect bond transmits everything. A weak bond absorbs selectively.
The general observation: in wave-bearing systems, weak coupling can be more functional than strong coupling. The interface that partially fails — that allows relative motion at certain frequencies — becomes a filter. What looks like a defect in the bonding is actually the mechanism doing the work.