Monte Carlo simulations for phonon transport in silicon nanomaterials

Abstract: In nanostructures phonon transport behaviour is distinctly different to transport in bulk materials such that materials with ultra low thermal conductivities and enhanced thermoelectric performance can be realized. Low thermal conductivities have been achieved in nanocrystalline materials that include hierarchical sizes of inclusions and pores. Nanoporous structures present a promising set of material properties and structures which allow for ultra-low thermal conductivity, even below the amorphous limit. In this paper we outline a semiclassical Monte Carlo code for the study of phonon transport and present an investigation of the thermal conductivity in nanoporous and nanocrystalline silicon. Different disordered geometry configurations are incorporated to investigate the effects of pores and grain boundaries on the phonon flux and the thermal conductivity, including the effects of boundary roughness, pore position and pore diameter. At constant porosity, thermal conductivity reduction is maximized by having a large number of smaller diameter pores as compared to a small number of larger diameter pores. Furthermore, we show that porosity has a greater impact on thermal conductivity than the degree of boundary roughness. Our simulator is validated across multiple simulation and experimental works for both pristine silicon channels and nanoporous structures.