In order to further understand the effect of EVA polarity on resistivity responses to external field, the morphology of both the original and the electric field-induced conductive network has been observed, as shown in Fig. 6. Fig. 6a and c shows CNTs in the original EVA10 and EVA25 samples present random distribution and do not have many obvious contacts to generate the electrical pathways due to the polymer barrier, leading to a high resistivity (see Fig. 1 and Fig. 2). Compared with CNT/EVA25, the dispersion of CNTs in the low-viscosity EVA10 is also more uniform, resulting in the lower initial electrical resistivity. After application of the electric field at elevated temperature (see Fig. 6b and d), however, CNTs seem to overcome the insulating polymer barrier and aligned to overlap each other as a result of a low resistivity (see Fig. 1 and Fig. 2). Observing the variation of the CNTs in the electric field, it is clear that CNTs experience irregular distortion by bending and some distance movement to align in one direction. It should be noted that the dendritic entanglement of CNTs observed from the original sample depresses the efficiency of alignment in electric field, so it is inevitable that some randomly aligned CNTs remain in Fig. 6b and d. Comparing the CNT alignment in Fig. 6b and d, it can be found that, macroscopically, CNTs display a higher degree of alignment in Fig. 6b than that in Fig. 6d. It implies CNTs are subjected to a stronger viscous drag of the surrounding EVA25 melt, which limits CNT aligned motion in the direction of electric field.