Effect of wave versus particle phonon nature in thermal transport through nanostructures
Comprehensive understanding of thermal transport in nanostructured materials needs large scale simulations bridging length scales dictated by different physics related to the wave versus particle nature of phonons. Yet, available computational approaches implicitly treat phonons as either just waves or as particles. In this work, using a full wave-based Non-Equilibrium Green's Function (NEGF) method, and a particle-based ray-tracing Monte Carlo (MC) approach, we investigate the qualitative differences in the wave and particle-based phonon transport at the vicinity of nanoscale features. For the simple example of a nanoporous geometry, we show that phonon transmission agrees very well for both methods with an error margin of approximately 15%, across phonon wavelengths even for features with sizes down to 3-4 nm. For cases where phonons need to squeeze in smaller regions to propagate, we find that MC underestimates the transmission of long wavelength phonons whereas wave treatment within NEGF indicates that those long wavelength phonons can propagate more easily. We also find that particle-based simulation methods are somewhat more sensitive to structural variations compared to the wave-based NEGF method. The insight extracted from comparing wave and particle methods can be used to provide a better and more complete understanding of phonon transport in nanomaterials.