Fig. 1

SEM images of the a RGO(OOH), b 3DGNs, c pristine ER, d RGO(OOH)-ER, e RGO(OH)-ER, f RGO(O)-ER, g 3DGNs-ER, and h 3DGNs-RGO(O)-ER. The digital photos of the ER, RGO filler, and RGO-3DGNs-ER are supplied in the insets of e–g, and all the scale bars represent 2 cm. The cross-sectional view of the SEM images is shown in the insets of h. SEM images of the pristine RGO, 3DGNs, and resulting TIMs are shown in figure, and the as-prepared composite TIMs display the smooth appearances (the digital photos of the ER, RGO filler, and RGO-3DGNs-ER are supplied in e–g). Different from that of the RGO, the size of wrinkles on the 3DGNs surface is much bigger (a, b). As for the RGO sample, the presence of wrinkles is spontaneous to enhance its stability, while the discrepancy between the thermal expansion coefficients of the graphene and nickel substrate leads to the wrinkles of the 3DGNs. A rough surface with obvious pores and cracks can be seen from the pristine ER, implying a poor thermal conductivity (c, the change of force constant resulting from the vacancies of the ER brings about a poor thermal conductivity) [11]. Contrarily, these cracks (forming during the solidification process) disappear after adding the graphene filler, which is in line with our previous reports [10, 12]. Moreover, partial RGO fillers can be seen on the surface of the RGO-ER specimens (d–f), while some obvious concave-convex (induced by the inner 3DGNs) appear on the surface of the 3DGNs-ER (g). Both these characteristics can be seen from the RGO and 3DGNs co-modified sample (h). The presence of the 3DGNs can be seen clearly from the cross-sectional view of the SEM images (insets of h)