Low-dimensional nanostructured materials are promising candidates for high efficiency solid-state cooling devices based on the Peltier effect. Thermal transport in these low-dimensional materials is a key factor for device performance since the thermoelectric figure of merit is inversely proportional to thermal conductivity. Therefore, understanding thermal transport in nanostructured materials is crucial for engineering high performance devices. Thermal transport in semiconductors is dominated by lattice vibrations called phonons, and phonon transport is often markedly different in nanostructures than it is in bulk materials for a number of reasons. First, as the size of a structure decreases, its surface area to volume ratio increases, thereby increasing the importance of boundaries and interfaces. Additionally, at the nanoscale the characteristic length of the structure approaches the phonon wavelength, and other interesting phenomena such as dispersion relation modification and quantum confinement may arise and further alter the thermal transport. In this paper we discuss phonon transport in semiconductor superlattices and nanowires with regards to applications in solid-state cooling devices. Systematic studies on periodic multilayers called superlattices disclose the relative importance of acoustic impedance mismatch, alloy scattering, and crystalline imperfections at the interfaces. Thermal conductivity measurements of mono-crystalline silicon nanowires of different diameters reveal the strong effects of phonon-boundary scattering. Experimental results for Si/SiGe superlattice nanowires indicate that different phonon scattering mechanisms may disrupt phonon transport at different frequencies. These experimental studies provide insight regarding the dominant mechanisms for phonon transport in nanostructures. Finally, we also briefly discuss Peltier coolers made from nanostructured materials that have shown promising cooling performance.
Size-dependent phononic thermal transport in low-dimensional nanomaterials
