Date of Award
Thesis - SCU Access Only
Santa Clara : Santa Clara University, 2017.
Master of Science (MS)
The high thermal conductivity and structural strength of carbon nanotubes (CNTs) make them promising materials for use as particle fillers in thermal interface materials (TIMs). It is therefore important to understand the effective thermal conductivity of CNT-matrix composites. Previous work has evaluated composite physical properties such as thermal conductivity using the effective medium approach (EMA). The EMA model predicts the effective thermal conductivity of composites consisting of a matrix and randomly dispersed inclusion for a very low volume fraction (f< 0.01). The EMA model does not account for particle to particle interactions and more detailed calculations are needed to quantify this effect.
The present study compares the EMA model and finite-element (FEA) computations for CNTmatrix compositions for various particle to particle distances. Vertically aligned fibers with identical lengths and various geometric distributions are considered. Nearby particles or nearest neighbors are identified to influence the fiber conductivity. In previous work, only near neighbors based on Voronoi cells were used. That work did not show good performance across the range of configurations. In this thesis the influence of near neighbors has been extended to farther elements. In the FEA simulation, the thermal conductivity of the particle filler is orthotropic because the Kapitza resistance is considered.
The FEA results show that the EMA model overestimates the effective thermal conductivity of the composite. A general correction factor for the effects of inter-particle interaction based on near neighbor distance is proposed and validated against various geometric configurations. Volume fractions up to f<0.1 are evaluated and in most cases, the correction function reduces the EMA over-estimation of the effective conductivity in the CNT-matrix region to within 30% of FEA results. The proposed correction factor extends the applicability of the EMA model from f<0.01 to f<0.10, with less than 30% error. For f<0.01, the correction function reduces the EMA over-estimation of the effective conductivity at the CNT-matrix region to within 10% of FEA results. The correction factor reduces
Truong, Stephanie, "Modeling of Particle to Particle Interactions to Adjust Effective Medium Approach Calculations of Thermal Conductivity of Carbon Nanotube Composites" (2017). Mechanical Engineering Master's Theses. 13.