| To prepare polymer composites with high thermal conductivity, polypropylene (PP) and polymethylmethacrylate (PMMA) were employed as matrix, aluminum nitride (AlN) and multi-welled carbon nanotubes (MWNT) were used as filler. The PP based composites were prepared by melt blend and for PMMA based composites solution mixing was employed. In the case of AlN/PP composites, AlN with different diameter (1-3μm and 110nm) were used to compare the effect of filler size on the thermal conductivity of the composites. The result shows that AlN/PP composites contain larger particles process higher thermal conductivity, as reported by many works.To investigate the effect surface modification of filler on the thermal conductivity of composites, surface modified AlN by coupling agents was added into PP matrix to prepare m-AlN/PP composites. Many papers have proved that surface modification of filler have positive effect on improving thermal conductivity of polymer based composites. However, in this work thermal conductivity of m-AlN/PP composites are lower than AlN/PP composites. The reason for surface modification showing a negative effect should be attributed to the poor dispersion of the filler in the PP matrix and too much coupling agent which is acting as thermal barrier around AlN particles.MWNT were also added into A1N/PP composites as second filler in order to study the thermal conductivity of hybrid filler system. Two methods were employed to prepare MWNT/A1N/PP composites:1) mix the MWNT and AIN powder at first and then add the mixed filler into the melt PP matrix; 2) blend the AIN and PP firstly then add MWNT into the A1N/PP mixture.In the case of PMMA based composites, which were prepared via solution mixing, AIN and MWNT were also used as sole or hybrid filler as in part one.Theoretical models for the prediction of thermal conductivity of composite were briefly reviewed and Maxwell model, Bruggeman model and Agari model were compared to the experiment result of this work. It was found that, Maxwell model and Bruggeman model could predict the thermal conductivity of composites with low filler volume fraction well, but when the filler volume fraction were higher, especially more than 30%, the predicted value were much lower than the experiment result. This is because Maxwell model and Bruggeman model didn't take interact of filler particles into consideration at the high filler volume fraction. When Agari model were employed, the coefficient of Cp, which stand for the ease of filler to form conductive chain and Cf, which strand for the crystallinity and the crystal size of the polymer matrix were calculated and compared among different composites system. Filler's critical volume fraction-CVF, above which the thermal conductivity of composites increased faster, also calculated. It turns out that the CVF predicted by Agari model were lower than experiment result. This should be attributed to the neglect of interface thermal resistance, which severely affect the thermal conductivity of the composites, for Agari model. |
