Analytical and numerical assessment of the effect of highly conductive inclusions distribution on the thermal conductivity of particulate composites

Highly conductive composites have found applications in thermal management, and the effective thermal conductivity plays a vital role in understanding the thermo-mechanical behavior of advanced composites. Experimental studies show that when highly conductive inclusions embedded in a polymeric matrix the particle forms conductive chain that drastically increase the effective thermal conductivity of two-phase particulate composites. In this study, we introduce a random network three dimensional (3D) percolation model which closely represent the experimentally observed scenario of the formation of the conductive chain by spherical particles. The prediction of the effective thermal conductivity obtained from percolation models is compared with the conventional micromechanical models of particulate composites having the cubical arrangement, the hexagonal arrangement and the random distribution of the spheres. In addition to that, the capabilities of predicting the effective thermal conductivity of a composite by different analytical models, micromechanical models, and, numerical models are also discussed and compared with the experimental data available in the literature. The results showed that random network percolation models give reasonable estimates of the effective thermal conductivity of the highly conductive particulate composites only in some cases. It is found that the developed percolation models perfectly represent the case of conduction through a composite containing randomly suspended interacting spheres and yield effective thermal conductivity results close to Jeffery's model. It is concluded that a more refined random network percolation model with the directional conductive chain of spheres should be developed to predict the effective thermal conductivity of advanced composites containing highly conductive inclusions.