Model and simulation predictions of the thermal conductivity of compact random nanoparticle composites

The lattice thermal conductivity of compact random nanoparticle composites was investigated numerically as well as theoretically. By taking advantage of the three-dimensional Vorogoni diagrams, we were able to generate realistic numerical models of the composites. The phonon motion was then simulated in use of a Monte-Carlo simulator which employed an unstructured grid system (tetrahedron cells) and was developed under the single relaxation time approximation and the grey medium approximation. On the other hand, a new effective medium approximation (EMA) model was proposed in the present work to predict the thermal conductivity of nanoparticle composites which particles are all made of a same material type. This EMA model was next combined with the EMA model proposed by Liang and Ji (2000) for three bond percolation systems to predict the thermal conductivity of nanoparticle composites of two material types. It was found that with an effective particle diameter characterizing either the average particle volume or the average particle surface area, the model predictions agree excellently with the Monte-Carlo simulation results and match well with the measurements of nanostructured bulk alloys.