In this work, thermally conductive polymer composites containing milled carbon fiber (MCF) were manufactured and characterized using laser flash analysis to measure thermal conductivity and optical microscopy to observe MCF in the matrix. Dicyclopentadiene (DCPD) and epoxy resins (YD-115 with low viscosity and YD-128 with high viscosity) were used as resin matrix to efficiently hold the fiber in place. MCF was aligned by applying a magnetic field with a permanent magnet with a magnetic flux density of 0.47 Tesla to improve thermal conductivity.
Various manufacturing parameters such as distance between specimen and magnet and position of specimen under magnet were examined for DCPD/MCF composites. It revealed that the largest enhancement in thermal conductivity was obtained when the magnetic field was applied at the closest distance between specimen and magnet. The thermal conductivity was increased with increase of MCF content, and more efficiently under a magnetic field. The enhanced thermal conductivity at the MCF content of 20 vol% was about 160 % higher than that of unaligned counterpart.
In the two epoxy resins, thermal conductivity was increased up to 10 vol% and 8 vol% MCF contents for YD-115 and YD-128, respectively, under the magnetic field as for DCPD, but decreased thereafter. The decrease of the thermal conductivity may be due to the higher viscosity of the epoxy resins compared to DCPD, leading to the MCF orientation more difficult under the magnetic field. Aligned YD-115/10 vol% and YD-128/8 vol% MCF specimens under a magnetic field showed 323 % and 325 % enhancement in thermal conductivity. The alignment of MCF in matrix resins as observed by optical microscopy is the major reason for the enhancement of the thermal conductivity.
Theoretical models such as Halpin-Tsai. Nielsen, and modified Nielsen, were applied to the thermal conductivity data as a function of MCF content obtained in this work. Although Halpin-Tsai model was a big difference between experimental and theoretical data, Nielsen and modified Nielsen models fit the experimental data well by adjusting shape factor in the models.