(350o) Deciphering the Effect of Morphology of Fillers in Thermal Conductivity of Nanofluid Using Diffusion-Based 3-D Modeling

In this work, a Finite Element Analysis (FEA) model has been developed using COMSOL Multiphysics® to estimate the effective thermal conductivity of nanofluids. Representative Volume Element (RVE) containing copper (Cu) nanofillers in ethylene glycol (EG) was created using MATLAB. Different morphologies such as spheres, cylinders, and triangular plates have been used as fillers in this study.

First, the required number of non-overlapping filler elements are placed inside the RVE randomly, and the domain is discretized. Then the steady-state heat conduction equation was solved for the composite using a cubic geometry, as shown in Figure-1. Temperature boundary condition was used in one pair of opposite surfaces, and the lateral surfaces were insulated. The effective thermal conductivity was obtained by using the steady state heat flux through the composite.

This model predicts around 26% increase in thermal conductivity in compared to the base fluid when rod-like filler was used at 0.15 vol% loading. In contrast, only 0.5% improvement is obtained for the particle for the same loading. The enhancements obtained for the spherical copper nanoparticles are in close agreement with the Hashin Shtrikman (HS) lower limit or the Maxwell model. Nanoplates show enhancements in between those of particles and wires. These simulation results suggest that elongated morphologies provide better enhancement than spherical geometries.

Since the morphology effect is very strong, significant effect of particle aggregation is also expected. For a given volume fraction, aggregated particles may form elongated aggregates which can increase the effective thermal conductivity significantly. This point has been confirmed by the simulation, and it is also found that the degree of enhancement is dependent on the nature of aggregate: a more elongated structure will produce more enhancement. For identical volume fraction (0.6 vol%), merely changing the state of aggregation changes the enhancement from 1.8% to 3.8%. All these shreds of evidence clearly indicate that the 'anomaly' usually reported for nanofluid is a morphology effect created by unstable colloids. Whenever a well-dispersed colloid is formed, such anomaly vanished producing the enhancement predicted by Maxwell’s theory. Possibly this feature has created much of the confusion on the subject.