With the continued miniaturization of electronic device components comes the problem of heat dissipation. This issue means that it is constantly necessary to improve the properties of thermal interface materials (TIMs) in modern chip packaging.^1–3 For this reason, elastomeric TIMs are usually fabricated by filling silicone rubber (SiR) with ceramic conductive fillers, i.e., made of alumina (Al₂O₃), aluminum nitride, boron nitride, or silicon carbide. Of these filler materials, Al₂O₃ is the most commonly used because of its high performance-to-price ratio. The thermal conductivity of Al₂O₃/SiR composites, however, is not very high (even with high filler loadings).

A large research effort has thus been dedicated to the use of graphite nanosheets (GNSs) to improve the thermal conductivity of Al₂O₃/SiR composites.^4–10 These materials are excellent heat conductors at room temperature and could thus potentially be used as thermally conducting fillers. Previous studies on the use of GNSs in Al₂O₃/SiR composites indicate that a high aspect ratio of the dispersed GNSs is an important factor for achieving high thermal conductivity enhancement (TCE) in the nanocomposites. Despite the success of these studies, they involved the use of several mixing techniques and organic solvents (i.e., which are not environmentally friendly). Large-scale fabrication of thermally conductive composites has also been restricted because of the complicated nature of the preparation procedure.

In this work,¹¹ we have therefore developed a more efficient method for the fabrication of thermally conductive SiR composites. In our approach, we use a double planetary mixer—and the direct mechanical compounding method—to achieve nanoscale dispersion of GNSs in liquid SiR. The mixer provides moderate shear forces to ensure good distribution and dispersion of the GNSs, but without severely reducing their aspect ratio. With the use of this solvent-free mixing technique, we are thus able to incorporate small amounts of GNSs (1–5 parts per hundred rubber, phr) into our highly filled S-Al₂O₃/SiR composites. The spatial distribution of the GNSs and S-Al₂O₃ in the hybrid filler composites is illustrated schematically in Figure 1. The lateral size of the dispersed GNSs is comparable to the diameter of the S-Al₂O₃ particles (see Figure 2), which means that the GNSs can act as bridges between the S-Al₂O₃ particles. Extra networks can also be formed if the distance between the surfaces of neighboring S-Al₂O₃ particles is less than the lateral size of the GNSs.

We find that the incorporation of only a small amount of GNSs can dramatically improve the thermal conductivity of both the SiR and the highly filled S-Al₂O₃/SiR composite samples (see Figure 3). For example, the thermal conductivity of SiR increased from 0.203 to 0.599W/(m·K) after the incorporation of 5 phr of GNSs. This increase corresponds to a TCE of 195%. In addition, we measured an increase in thermal conductivity of the 300phr S-Al₂O₃/SiR composite from 1.103 to 2.162W/(m·K) upon addition of 5phr GNS (i.e., an improvement of about 96%).

The strain dependence of the storage modulus (G ′) for filled rubber is generally known as the ‘Payne effect,’ which can be used to qualitatively characterize filler–filler networks. We therefore investigated the Payne effect for our composite samples (with different loadings of S-Al₂O₃ and GNSs) to indirectly characterize the filler–filler network strength. In particular, we have determined the dynamic storage modulus difference (ΔG ′)—i.e., between the dynamic storage modulus at a very small strain and at a high final strain—which is considered to be a qualitative indication of filler–filler network strength.¹² Our results (see Figure 4) show that ΔG ′ (i.e., the filler network strength) of the SiR/S-Al₂O₃ composites increases with increasing GNS loading (between 0 and 5phr). Furthermore, the change in ΔG ′ with GNS loading varies with S-Al₂O₃ content. As a result, the filler–filler network is stronger for the 300phr S-Al₂O₃ sample than for the 100 and 200phr S-Al₂O₃ composites.

We also calculated the thermal conductivity for each of the hybrid filler systems by summing the individual effects of the GNS and S-Al₂O₃ fillers. In this way, we were able—see Figure 5—to determine the extent of synergy (i.e., the interaction of multiple elements in a system to produce an effect different from, or greater than, the sum of their individual effects) in the composites. We find that the extent of synergy increases with GNS loading and is at the maximum for the 300phr S-Al₂O₃ composite. We ascribe this synergetic effect to the extra filler–filler networks that are formed by the graphite nanoflakes bridging neighboring S-Al₂O₃ particles. These bridges thus act as an effective pathway for phonon movement in the insulator matrix.¹³ In addition, our results indicate that the extent of synergy depends on both the composition and loading of the fillers.

In summary, we have demonstrated a new and efficient method for producing thermally conductive SiR composites. In this approach, small amounts of GNSs (1–5phr) are added to highly filled S-Al₂O₃/SiR composites via direct mechanical compounding. Nanoscale dispersion of the GNSs was successfully achieved and the addition of these small GNS loadings produced substantial increases in the thermal conductivity of the samples. The results of the study also indicate that there is a synergetic effect on the thermal conductivity improvement. In our future work, we will try to optimize the blending conditions so that we can further improve the dispersion of the GNSs, and thus reduce their loading.