Shrinking phonon spectrometry

Nanomaterials often seem to obey a different set of rules than those governing the bulk materials we encounter in our everyday lives. For example, in a bulk semiconductor or insulator, heat transport properties are determined largely by phonon–phonon scattering; in a nanostructure, phonon–surface interactions play the key role. The precise nature of those interactions, however, is difficult to pin down from first principles. It’s generally agreed that phonons with wavelengths much larger than a surface’s characteristic roughness length scale should reflect specularly, like light off a mirror, and that those with much shorter wavelengths should scatter randomly. Theories differ, however, as to what happens at intermediate ranges. Now Richard Robinson and coworkers at Cornell University have introduced a tool that could help resolve that question. Their microscale phonon spectrometer, pictured here, uses a superconducting tunnel junction (STJ) to produce narrowband phonons whose energies are determined by the applied voltage bias. Phonons that pass through the target sample—silicon nanosheets etched into a raised mesa—generate a tunneling current as they impinge on a superconducting quantum interference device (SQUID). By monitoring that tunneling current as a function of phonon energy, it’s possible to tease out the nanosheets’ wavelength-dependent scattering rates. Preliminary experiments using nanosheets with a characteristic roughness of 1 nm have already produced a surprising result. According to an established theory, phonon spectra with peak wavelengths in the 10- to 15-nm range should yield mostly specular reflection; spectroscopic measurements, however, suggest the great majority of those phonons scattered randomly. (J. B. Hertzberg et al., Nano Lett., in press .)—Ashley G. Smart