- Email: [email protected]
Synergistic enhancement of thermal conductivity in polymer composites filled with self-hybrid expanded graphite fillers
Journal of Non-Crystalline Solids 450 (2016) 75–81
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/locate/jnoncrysol
Synergistic enhancement of thermal conductivity in polymer composites filled with self-hybrid expanded graphite fillers Hyun Su Kim, Jung Hyun Na, Yong Chae Jung, Seong Yun Kim ⁎ Multifunctional Structural Composite Research Center, Institute of Advanced Composite Materials, Korea Institute of Science and Technology (KIST), 92 Chudong-ro, Bongdong-eup, Wanju-gun, Jeonbuk 55324, Republic of Korea
a r t i c l e
i n f o
Article history: Received 4 May 2016 Received in revised form 27 July 2016 Accepted 28 July 2016 Available online 4 August 2016 Keywords: Polymer composite Thermal conductivity X-ray micro computed tomography Expanded graphite
a b s t r a c t The size of expanded graphite (EG) was controlled using a high-speed crusher because a hybrid of differently sized fillers can induce a synergistic enhancement of the thermal conductivity in polymer composites. We found that the thermal conductivity of a polymer composite filled with both 10 wt% EG and 10 wt% highspeed crusher treated EG (wEG) was synergistically improved by 12.0 and 20.7% compared to that of polymer composites filled with 20 wt% EG and 20 wt% wEG alone, respectively. A three-dimensional (3D) non-destructive analysis using X-ray micro-computed tomography (micro-CT) was applied to explain the synergistic enhancement and to identify the dispersion and 3D network of EG fillers in the composites accurately. According to the non-destructive analysis results, the synergistic enhancement was caused by the formation of efficient thermally conductive pathways due to the hybrid of the differently sized EG and wEG fillers. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Electronic devices have become increasingly integrated into ever smaller spaces. It is therefore becoming more critical to dissipate the heat generated efficiently from the insides of devices, as failure adequately to discharge residual heat to the outside environment can cause heat damage or even a fire [1–3]. Particularly, when applying heat-dissipating structures to light-emitting diodes and electronic device housing units, it is necessary to develop polymer composites with high thermal conductivity; and which also have relatively simple processing conditions and a good degree of product flexibility [4]. These advanced thermally conductive polymer composites are typically fabricated by mixing thermally conductive fillers such as metals, ceramics and carbons into a polymer resin [5]. Recently, nano-carbon fillers such as graphene and carbon nanotubes are receiving more attention due to their excellent thermal conductivity between 1950 and 7000 W/m·K [6–11]. However, when a polymer matrix is mixed with nano-carbon fillers, incomplete contact can be produced at the interface between the nano-filler and the polymer matrix, or between the nano-fillers themselves [6,12]. As a result, it has been reported that this incomplete contact can produce interfacial thermal resistance and thermal contact resistance, resulting in phonon scattering and low thermal conductivity near the lower bound of the rule of mixture [6,12–14]. In addition, the high price of nano-carbon fillers compared to other types of thermal conductive fillers is an ⁎ Corresponding author. E-mail address: [email protected] (S.Y. Kim).
http://dx.doi.org/10.1016/j.jnoncrysol.2016.07.038 0022-3093/© 2016 Elsevier B.V. All rights reserved.
obstacle to the commercial use of thermally conductive polymer composites containing nano-carbon fillers. Expanded graphite (EG) is a carbon material with the unique structure of graphite sheets with interlayer spacings. When filled within polymers, it can become distributed inside the composite in an efficient manner, forming a thermally conductive network. Noh and Kim [15] evaluated the thermal conductivity of thermally conductive polymer composites filled with various carbon fillers such as EG, pitch-based carbon fiber, graphite, graphene nanoplatelet, multi-walled carbon nanotube and carbon black. They reported that the polymer composite filled with EG fillers showed the best performance in terms of thermal conductivity [15]. In addition, given that EG is one of the most cost-effective carbon fillers, it is highly advantageous to use from a commercialization perspective. It is known that the electrical conductivity of polymer composites rapidly increases at a specific range of filler content, according to percolation theory [16,17]. However, the thermal conductivity of composites filled with EG fillers show a monotonous increase with respect to the filler loading amount [18,19]. Yung and Liem [20] reported that the introduction of hybrid fillers consisting of different particle sizes contributed to the synergistic enhancement of the thermal conductivity of thermally conductive polymer composites and that this synergistic effect was related to the internal structure of the composites. Accordingly, to induce the synergistic enhancement of the thermal conductivity of a composite filled with EG fillers, it can be an effective approach to mix differently sized EG fillers into a polymer matrix. In this study, inexpensive commercial EG was treated by a highspeed crusher, which is an advantageous method in terms of the
76
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
Table 1 Sample composition Sample
EG (wt%)
wEG (wt%)
PC (wt%)
E5 E10 E15 E20 w5 w10 w15 w20 E10w10
5 10 15 20 0 0 0 0 10
0 0 0 0 5 10 15 20 10
95 90 85 80 95 90 85 80 80
processing time and cost, to control the size of the EG. The resulting selfhybrid composites filled with different sizes of the same fillers (raw EG and wEG) were fabricated using a typical and cost-effective melt mixing method. Their thermal conductivity values were measured in an effort to investigate the possible synergistic enhancement of the thermal conductivity by the self-hybrid. In addition, a non-destructive X-ray microCT method was used to observe the internal structure of the composite to analyze the synergistic effect and to optimize the thermal conductivity of the composites. 2. Experimental 2.1. Materials EG powder was provided by Hana Chemtech (ES 250 B5D, Hana Chemtech, Seoul, Korea). The EG has a unique structure consisting of graphite sheets with interlayer spacings which effectively enhances the thermal conductivity of composites filled with EG filler. The EG had a carbon content of 95% and an expansion coefficient of 250 cm3/g. Average size of the EG was larger than 300 μm. To control the size of the EG, a high-speed crusher (WB-1, Osaka Chemical Co. Ltd., Osaka, Japan) was used to treat it at a speed of 25,000 rpm for 30 s at room temperature. The crusher crushes samples instantly from hard to soft, such as ceramics, cereals, chips of wood, clothes, crude drugs, medicinal herbs and minerals, based on its super-high-speed rotation and strong motor power of 700 W. A linear polycarbonate (PC) resin (LUPOY PC 1300-03, LG Chemistry Co., Gyeonggi-do, Korea) which was designed for extrusion or injection, was used as the matrix.
The Vicat softening point was 151 °, as measured under conditions of 50 °/hr and a 50 N load according to ASTM D 696. The density of the resin was 1200 kg/m3, as measured according to ASTM D 792. The melt flow rate was 3 g/10 min, as measured according to ASTM D 1238. 2.2. Composite fabrication The prepared EG and/or wEG and PC resin samples were weighed to the target contents as shown in Table 1. They were put into a HAAKE Rheomix internal Mixer (HAAKE™ Rheomix OS Lab Mixers, Thermo Scientific Inc., Marietta, GA, USA) heated to 260 °C and mixed at a constant screw speed of 60 rpm for 30 min, and then pelletized. The fabricated pellets were put into a mold (2.5 cm2 × 2 mm) for the thermal conductivity tests. The pellets were pressed using a heating press (D3P-30J, Daheung Science, Incheon, Korea) at a pressure of 15 MPa under a temperature of 260 °C for 15 min and were then quenched to 30 °C using cold water.
2.3. Characterization 2.3.1. EG filler For an analysis of the surface functional groups of the EG fillers, Fourier transform infrared spectroscopy (FT-IR, Nicolet 6700, Thermo Scientific, MA, USA) and X-ray photoelectron spectroscopy (XPS, KAlpha, Thermo Scientific, Massachusetts, USA) were used. The FT-IR spectra were measured in the range of 1000–4000 cm−1 at a resolution of 16 cm−1. The XPS spectra were obtained with an Al X-ray source at a pressure of 1 × 10−8 Pa. To evaluate the defect level of the EG fillers, Raman spectroscopy (LabRAM HR 800, HORIBA Jobin Yvon, Japan) was used along with a 514 nm Ar ion laser. 2.3.2. Morphology The fabricated composites were fractured using liquid nitrogen to obtain the samples. The EG, wEG and composite samples were coated with platinum for 120 s under a vacuum using a sputter-coating machine (Ion Sputter E-1030, Hitachi High Technologies, Tokyo, Japan). In this case, 10 kV of electricity was applied to the coated samples under a nitrogen vacuum using a field emission scanning microscope (FE-SEM, Nova NanoSEM 450, FEI Corp., OR, USA).
Fig. 1. SEM images of EG fillers with respect to treatment time of the high-speed crusher. (a) Raw EG, (b) 10 s, (c) 20 s, and (d) 30 s.
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
2.3.3. Thermal conductivity The thermal conductivity of the composites was measured at ambient temperature and normal pressure using a thermal conductivity measurement instrument (TPS 2500S, Hot Disk ab, Gothenburg, Sweden) according to the ISO 22007-2 standard. The sensor used was constructed with dual spirals made of thin nickel wires and was operated by a sequential plane heat source. In order to measure the change in the temperature of the sensor itself, using the change in the sensor's resistance, the sensor was supplied with a certain level of electric power (P), which induced a temperature rise (ΔT). The thermal conductivity of the samples was obtained from the Fourier thermal conductivity formula based on the supplied power and induced temperature changes. The reproducibility and accuracy of the instrument are better than 1 and 5%, respectively [21,22].
77
2000 × 1332, and the X-ray source was applied under a voltage of 51 kV, a current of 194 μA, and at a normal pressure level. 2.3.5. Heat dissipation To measure the change in the temperature of the fabricated composites during the heating and cooling, an infrared camera (FLIR T420, Wilsonville, OR, USA) was used for the evaluation. The composite samples were put on a hot plate and heated to 100 °C. After 20 s, their heat dissipation image was recorded. In addition, the composite samples heated to 100 °C were placed on a plate at ambient temperature and then, after 20 s, their heat dissipation image was recorded [23,24]. 3. Results and discussion
2.3.4. Tomography A micro-CT device (Skyscan 1172, Bruker Co., Billerica, MA, USA) was used to evaluate the dispersion and network structure of the fillers inside the composites. Measurements were taken at a pixel size of
Fig. 2. (a) FT-IR, (b) XPS and (c) Raman spectra of EG fillers with respect to treatment time of the high-speed crusher.
As shown in Fig. 1, the raw EG becomes fragmented from the EG size of larger than 300 μm to the wEG size of 100–200 μm and smoother with respect to the treatment time in the high-speed crusher of 30 s, which is the processing time recommended by the manufacturer. In the FT-IR spectra as shown in Fig. 2(a), similar characteristic peaks of O\\H, C_C and C\\O were observed, and C/O ratio in the XPS spectra shown in Fig. 2(b) was increased as the O peak was reduced. These findings indicate that oxygen groups introduced onto the surface of the raw EG, formed by the sequential processes of the oxidation and heat treatment of the graphite based precursor materials, were removed by the high-speed crusher treatment. Raman spectroscopy is a useful tool for identifying the defect levels of carbon materials, and Fig. 2(c) shows the Raman spectra of the EG and wEG fillers. In the Raman spectra, the D and G bands were observed at ~1338 and ~1572 cm−1, respectively. The D-band is a disorder induced feature that arises from a double-resonance Raman scattering process from non-zero-center phonon modes, which are typically attributed to the presence of amorphous or disordered carbon atoms [12]. The G band results from in-plane tangential stretching of the carbon-carbon bonds in the graphene sheets [12]. Accordingly, the defect level is lowered as the intensity ratio of the D and G bands (ID/IG) is lowered, and the defect level of the raw EG was reduced during the high-speed crusher treatment. Therefore, the raw EG became smaller and smoother, and the oxidation groups were removed from the surface, leading to a reduced defect level after the high-speed crusher treatment. Fig. 3 shows the thermal conductivity of the composites filled with the EG and wEG fillers. The thermal conductivity increases as the filler content is increased. The thermal conductivity of the E20 composite was 2.34 W/m·K, an enhancement of 1014% over the thermal conductivity of the PC resin (0.21 W/m·K). The thermal conductivity of the w20
Fig. 3. Thermal conductivity of polymer composites filled with the EG and/or wEG fillers. The inset image shows the measuring instrument.
78
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
Fig. 4. (a) 2D micro-CT image, (b) micro-CT reconstruction image, (c) and (d) 3D micro-CT images of the E10w10 composite.
Fig. 5. (a) SEM and (b) micro-CT images of the E5 composite, (c) SEM and (d) micro-CT images of the E15 composite, (e) SEM and (f) micro-CT images of the w5 composite, and (g) SEM and (h) micro-CT images of the w15 composite.
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
composite was 2.17 W/m·K, which was a relatively low enhancement. The defect level of the fillers may be the critical physical factor in determining the thermal conductivity of composites because higher thermal conductivity of the filler can be expected as the defect level is reduced [25,26]. From this perspective, higher thermal conductivity of the composites filled with wEG can be expected. However, this unexpected result can be caused by interfacial thermal resistance between the EG filler and the polymer resin, and the interfacial thermal resistance negatively affects the phonon transfer between the filler and the resin [27–30]. In particular, it is considered that the wEG filler is smaller in size than the raw EG filler and, given the same content, forms more interfaces with the resin within the composites. Consequently, the thermal conductivity of the composites filled with wEG is lower. The thermal conductivity of the E10w10 composite was 2.62 W/m·K, which represents an enhancement by 12.0 and 20.7% compared to that of the E20 and w20 composites, respectively. This implies that the thermal conductivity of the composites filled with both EG and wEG of different sizes is synergistically improved over that of the composites filled with the homogeneously sized EG only. To investigate the cause of the synergistic enhancement in the thermal conductivity, an accurate analysis of the 3D network of the thermally conductive fillers is required. A morphological study based on scanning electron microscopy of the fracture surface of the composites is usually adopted to investigate the internal structures of composites. However, this method is associated with alterations of the internal structures of the samples during the sampling process, as well as image distortion owing to the two-dimensional (2D) image analysis. For a precise analysis of the internal structure of the composites, the 3D non-destructive analysis method should be utilized. A micro-CT analysis based on an X-ray source is a useful analysis tool which can meet the requirements of a precise 3D non-destructive
79
analysis. The 2D micro-CT images were measured sequentially in the direction of the sample thickness, as shown in Fig. 4(a), with the results then reconstructed as 3D images, as shown in Fig. 4(b). In this way, it was possible to measure the internal structure within the composites in a non-destructive manner, as shown in Fig. 4(c) and (d), and determine how the EG fillers are distributed and oriented. However, it is difficult to identify the internal structure of the composites according to the contents and sizes of the EG fillers merely by studying the fracture surface, as shown in the SEM images in Figs. 5 and 6. Therefore, in this study, it is sought to identify the dispersion and 3D network of the EG fillers in the composites in both the SEM and 3D micro-CT analyses. In the micro-CT images shown in Fig. 4(c) and (d), the dispersion and orientation of the EG and wEG fillers can be clearly observed. The EG and wEG with the interlayer spacings are denoted here by the blue and red circles, respectively. The micro-CT image of the composites in Fig. 5 shows a 3D thermal conduction network that was formed as the EG filler content increased inside the composites. A better 3D thermal conduction network of the composites filled with both the EG and wEG fillers was formed than that of the composites filled with EG filler alone when considering the empty spaces inside the composites without fillers, as represented by the red marks in Fig. 6(b) and (d). Therefore, it can be inferred that the synergistically improved thermal conductivity of the composites containing both the EG and wEG fillers is caused by the efficient structure of the 3D thermally conductive pathways [31]. Fig. 7 shows the heat dissipation properties of the composites. This indicates that the composites filled with both the EG and wEG fillers were most sensitive to the temperature changes caused by heating and cooling. This results show that the composites with high thermal conductivity values were excellent for heat dissipation and that the heat dissipation trend was consistent with the thermal conductivity trend. Accordingly,
Fig. 6. (a) SEM and (b) micro-CT images of the E20 composite, (c) SEM and (d) micro-CT images of the E10w10 composite, and (e) SEM and (f) micro-CT images of the w20 composite.
80
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
Fig. 7. (a) Infrared camera images of composites filled with the EG and/or wEG fillers during heating and cooling for transient temperature response and heat transport, and temperaturetime profiles of the composites during (b) heating and (c) cooling.
it was found that optimizing the filler and the efficient construction of a 3D thermally conductive filler network are the most critical physical factors for optimizing the thermal conductivity of composites and for maximally tuning their heat dissipation capabilities. 4. Conclusions EG is considered as an effective filler for improving the thermal conductivity of composites owing to its unique structure of graphite sheets with interlayer spacings. It can efficiently fill polymer matrix and form thermal transfer pathways. In this study, in order to optimize the thermal conductivity of composites filled with EG fillers, wEG fillers which had been cut into small pieces using a high-speed crusher were prepared and investigated, as the high-speed crusher treatment was considered efficient in terms of both the process time and the cost. The thermal conductivity of the composite filled with both 10 wt% EG and wEG fillers was improved compared to those of the composites filled with only one type of filler at 20 wt%. The filler network structure was precisely observed in 3D non-destructive micro-CT analysis, which confirmed that there was a significant effect on the thermal conductivity of the hybrid composites filled with two types of fillers. Higher thermal conductivity of the composites was observed as a better 3D network of the fillers was formed, thus providing more efficient pathways to transfer phonons. These thermal conductivity results were in good agreement with the heat dissipation results during the heating and cooling process. Acknowledgements This study was supported by Korea Institute of Science and Technology (KIST) Institutional Program and the Technological innovation R&D program of SMBA [S2177379].
References [1] T. Evgin, H.D. Koca, N. Horny, A. Turgut, I.H. Tavman, M. Chirtoc, M. Omastová, I. Novak, Effect of aspect ratio on thermal conductivity of high density polyethylene/ multi-walled carbon nanotubes nanocomposites, Compos. A: Appl. Sci. Manuf. 82 (2016) 208. [2] Z. Wang, R. Qi, J. Wang, S. Qi, Thermal conductivity improvement of epoxy composite filled with expanded graphite, Ceram. Int. 41 (2015) 13541. [3] H.S. Kim, J.-u. Jang, J. Yu, S.Y. Kim, Thermal conductivity of polymer composites based on the length of multi-walled carbon nanotubes, Compos. Part B 79 (2015) 505. [4] J.G. Cevallos, A.E. Bergles, A. Bar-cohen, P. Rodgers, S.K. Gupta, Polymer heat exchangers-history, opportunities, and challenges, Heat Transfer Eng. 33 (2012) 1075. [5] Y.J. Noh, S.Y. Kim, Synergistic improvement of thermal conductivity in polymer composites filled with pitch based carbon fiber and graphene nanoplatelets, Polym. Test. 45 (2015) 132. [6] Z. Han, A. Fina, Thermal conductivity of carbon nanotubes and their polymer nanocomposites: a review, Prog. Polym. Sci. 36 (2011) 914. [7] C. Yu, L. Shi, Z. Yao, D. Li, A. Majumdar, Thermal conductance and thermopower of an individual single-wall carbon nanotube, Nano Lett. 5 (2005) 1842. [8] P. Kim, L. Shi, A. Majumdar, P.L. McEuen, Thermal transport measurements of individual multiwalled nanotubes, Phys. Rev. Lett. 87 (2001) 215502. [9] K.M.F. Shahil, A.A. Balandin, Graphene-multilayer graphene nanocomposites as highly efficient thermal interface materials, Nano Lett. 12 (2012) 861. [10] A.A. Balandin, Thermal properties of graphene and nanostructured carbon materials, Nat. Mater. 10 (2011) 569–581. [11] A.A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, C.N. Lau, Superior thermal conductivity of sigle-layer garphene, Nano Lett. 8 (2008) 902. [12] S.Y. Pak, H.M. Kim, S.Y. Kim, J.R. Youn, Synergistic improvement of thermal conductivity of thermoplastic composites with mixed boron nitride and multi-walled carbon nanotube fillers, Carbon 50 (2012) 4830. [13] S.Y. Kim, Y.J. Noh, J. Yu, Improved thermal conductivity of polymeric composites fabricated by solvent-free processing for the enhanced dispersion of nanofillers and a theoretical approach for composites containing multiple heterogeneities and geometrized nanofillers, Compos. Sci. Technol. 101 (2014) 79. [14] S.Y. Kim, Y.J. Noh, J. Yu, Thermal conductivity of graphene nanoplatelets filled composites fabricated by solvent-free processing for the excellent filler dispersion and a theoretical approach for the composites containing the geometrized fillers, Compos. A: Appl. Sci. Manuf. 69 (2015) 219. [15] Y.J. Noh, H.S. Kim, B.-C. Ku, M.-S. Khil, S.Y. Kim, Thermal conductivity of polymer composites with geometric characteristics of carbon allotropes, Adv. Eng. Mater. 18 (2016) 1127.
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81 [16] S.Y. Kim, Y.J. Noh, J. Yu, Prediction and experimental validation of electrical percolation by applying a modified micromechanics model considering multiple heterogeneous inclusions, Compos. Sci. Technol. 106 (2015) 156. [17] Y.J. Noh, S.Y. Pak, S.H. Hwang, J.Y. Hwang, S.Y. Kim, J.R. Youn, Enhanced dispersion for electrical percolation behavior of multi-walled carbon nanotubes in polymer nanocomposites using simple powder mixing and in situ polymerization with surface treatment of the fillers, Compos. Sci. Technol. 89 (2013) 29. [18] M. Chirtoc, N. Horny, I. Tavman, A. Turgut, I. Kökey, M. Omastová, Preparation and photothermal characterization of nanocomposites based on high density polyethylene filled with expanded and unexpanded graphite: particle size and shape effects, Int. J. Therm. Sci. 62 (2012) 50. [19] J. Kratochvíla, A. Boudenne, I. Krupa, Effect of filler size on thermophysical and electrical behavior of nanocomposites based on expanded graphite nanoparticles filled in low-density polyethylene matrix, Polym. Compos. 34 (2013) 149. [20] K.C. Yung, H. Liem, Enhanced thermal conductivity of boron nitride epoxy-matrix composite through multi-modal particle size mixing, J. Appl. Polym. Sci. 160 (2007) 3587. [21] M. Gustavsson, E. Karawacki, S.E. Gustafsson, Thermal conductivity, thermal diffusivity, and specific heat of thin samples from transient measurements with hot disk sensors, Rev. Sci. Instrum. 65 (1994) 3856. [22] V. Bohac, M.K. Gustavsson, L. Kubicar, S.E. Gustafsson, Parameter estimations for measurements of thermal transport properties with the hot disk thermal constants analyzer, Rev. Sci. Instrum. 71 (2000) 2452.
81
[23] H.S. Kim, H.S. Bae, J. Yu, S.Y. Kim, Thermal conductivity of polymer composites with the geometrical characteristics of graphene nanoplatelets, Sci Rep. 6 (2016) 26825. [24] E.-C. Cho, J.-H. Huang, C.-P. Li, C.-W. Chang-Jian, K.-C. Lee, Y.-S. Hsiao, J.-H. Huang, Graphene-based thermoplastic composites and their application for LED thermal management, Carbon 102 (2016) 66. [25] N.-J. Song, C.-M. Chen, C. Lu, Z. Liu, Q.-Q. Kong, R. Cai, Thermally reduced graphene oxide films as flexible lateral heat spreaders, J. Mater. Chem. A 39 (2014) 16563. [26] G. Xin, H. Sun, S.P. Scott, T. Yao, F. Lu, D. Shao, T. Hu, G. Wang, G. Ran, J. Lian, Advanced phase change composite by thermally annealed defect-free graphene for thermal energy storage, ACS Appl. Mater. Interfaces 6 (2014) 15262. [27] S.T. Huxtable, D.G. Cahill, S. Shenogin, L. Xue, R. Ozisik, P. Barone, M. Usrey, M.S. Strano, G. Siddons, M. Shim, P. Keblinski, Interfacial heat flow in carbon nanotube suspensions, Nat. Mater. 2 (2003) 731. [28] C.-W. Nan, Z. Shi, Y. Lin, A simple model for thermal conductivity of carbon nanotube-based composites, Chem. Phys. Lett. 375 (2003) 666. [29] S. Shenogin, L. Xue, R. Ozisik, P. Keblinski, D.G. Cahill, Role of thermal boundary resistance on the heat flow in carbon-nanotube composites, J. Appl. Phys. 95 (2004) 8136. [30] N. Shenogina, S. Shenogin, L. Xue, P. Keblinski, On the lack of thermal percolation in carbon nanotube composites, Appl. Phys. Lett. 87 (2005) 133106. [31] U. Szeluga, B. Kumanek, B. Trzebicka, Synergy in hybrid polymer/nanocarbon composites. A review, Compos. A: Appl. Sci. Manuf. 73 (2015) 20.
Contents lists available at ScienceDirect
Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/locate/jnoncrysol
Synergistic enhancement of thermal conductivity in polymer composites filled with self-hybrid expanded graphite fillers Hyun Su Kim, Jung Hyun Na, Yong Chae Jung, Seong Yun Kim ⁎ Multifunctional Structural Composite Research Center, Institute of Advanced Composite Materials, Korea Institute of Science and Technology (KIST), 92 Chudong-ro, Bongdong-eup, Wanju-gun, Jeonbuk 55324, Republic of Korea
a r t i c l e
i n f o
Article history: Received 4 May 2016 Received in revised form 27 July 2016 Accepted 28 July 2016 Available online 4 August 2016 Keywords: Polymer composite Thermal conductivity X-ray micro computed tomography Expanded graphite
a b s t r a c t The size of expanded graphite (EG) was controlled using a high-speed crusher because a hybrid of differently sized fillers can induce a synergistic enhancement of the thermal conductivity in polymer composites. We found that the thermal conductivity of a polymer composite filled with both 10 wt% EG and 10 wt% highspeed crusher treated EG (wEG) was synergistically improved by 12.0 and 20.7% compared to that of polymer composites filled with 20 wt% EG and 20 wt% wEG alone, respectively. A three-dimensional (3D) non-destructive analysis using X-ray micro-computed tomography (micro-CT) was applied to explain the synergistic enhancement and to identify the dispersion and 3D network of EG fillers in the composites accurately. According to the non-destructive analysis results, the synergistic enhancement was caused by the formation of efficient thermally conductive pathways due to the hybrid of the differently sized EG and wEG fillers. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Electronic devices have become increasingly integrated into ever smaller spaces. It is therefore becoming more critical to dissipate the heat generated efficiently from the insides of devices, as failure adequately to discharge residual heat to the outside environment can cause heat damage or even a fire [1–3]. Particularly, when applying heat-dissipating structures to light-emitting diodes and electronic device housing units, it is necessary to develop polymer composites with high thermal conductivity; and which also have relatively simple processing conditions and a good degree of product flexibility [4]. These advanced thermally conductive polymer composites are typically fabricated by mixing thermally conductive fillers such as metals, ceramics and carbons into a polymer resin [5]. Recently, nano-carbon fillers such as graphene and carbon nanotubes are receiving more attention due to their excellent thermal conductivity between 1950 and 7000 W/m·K [6–11]. However, when a polymer matrix is mixed with nano-carbon fillers, incomplete contact can be produced at the interface between the nano-filler and the polymer matrix, or between the nano-fillers themselves [6,12]. As a result, it has been reported that this incomplete contact can produce interfacial thermal resistance and thermal contact resistance, resulting in phonon scattering and low thermal conductivity near the lower bound of the rule of mixture [6,12–14]. In addition, the high price of nano-carbon fillers compared to other types of thermal conductive fillers is an ⁎ Corresponding author. E-mail address: [email protected] (S.Y. Kim).
http://dx.doi.org/10.1016/j.jnoncrysol.2016.07.038 0022-3093/© 2016 Elsevier B.V. All rights reserved.
obstacle to the commercial use of thermally conductive polymer composites containing nano-carbon fillers. Expanded graphite (EG) is a carbon material with the unique structure of graphite sheets with interlayer spacings. When filled within polymers, it can become distributed inside the composite in an efficient manner, forming a thermally conductive network. Noh and Kim [15] evaluated the thermal conductivity of thermally conductive polymer composites filled with various carbon fillers such as EG, pitch-based carbon fiber, graphite, graphene nanoplatelet, multi-walled carbon nanotube and carbon black. They reported that the polymer composite filled with EG fillers showed the best performance in terms of thermal conductivity [15]. In addition, given that EG is one of the most cost-effective carbon fillers, it is highly advantageous to use from a commercialization perspective. It is known that the electrical conductivity of polymer composites rapidly increases at a specific range of filler content, according to percolation theory [16,17]. However, the thermal conductivity of composites filled with EG fillers show a monotonous increase with respect to the filler loading amount [18,19]. Yung and Liem [20] reported that the introduction of hybrid fillers consisting of different particle sizes contributed to the synergistic enhancement of the thermal conductivity of thermally conductive polymer composites and that this synergistic effect was related to the internal structure of the composites. Accordingly, to induce the synergistic enhancement of the thermal conductivity of a composite filled with EG fillers, it can be an effective approach to mix differently sized EG fillers into a polymer matrix. In this study, inexpensive commercial EG was treated by a highspeed crusher, which is an advantageous method in terms of the
76
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
Table 1 Sample composition Sample
EG (wt%)
wEG (wt%)
PC (wt%)
E5 E10 E15 E20 w5 w10 w15 w20 E10w10
5 10 15 20 0 0 0 0 10
0 0 0 0 5 10 15 20 10
95 90 85 80 95 90 85 80 80
processing time and cost, to control the size of the EG. The resulting selfhybrid composites filled with different sizes of the same fillers (raw EG and wEG) were fabricated using a typical and cost-effective melt mixing method. Their thermal conductivity values were measured in an effort to investigate the possible synergistic enhancement of the thermal conductivity by the self-hybrid. In addition, a non-destructive X-ray microCT method was used to observe the internal structure of the composite to analyze the synergistic effect and to optimize the thermal conductivity of the composites. 2. Experimental 2.1. Materials EG powder was provided by Hana Chemtech (ES 250 B5D, Hana Chemtech, Seoul, Korea). The EG has a unique structure consisting of graphite sheets with interlayer spacings which effectively enhances the thermal conductivity of composites filled with EG filler. The EG had a carbon content of 95% and an expansion coefficient of 250 cm3/g. Average size of the EG was larger than 300 μm. To control the size of the EG, a high-speed crusher (WB-1, Osaka Chemical Co. Ltd., Osaka, Japan) was used to treat it at a speed of 25,000 rpm for 30 s at room temperature. The crusher crushes samples instantly from hard to soft, such as ceramics, cereals, chips of wood, clothes, crude drugs, medicinal herbs and minerals, based on its super-high-speed rotation and strong motor power of 700 W. A linear polycarbonate (PC) resin (LUPOY PC 1300-03, LG Chemistry Co., Gyeonggi-do, Korea) which was designed for extrusion or injection, was used as the matrix.
The Vicat softening point was 151 °, as measured under conditions of 50 °/hr and a 50 N load according to ASTM D 696. The density of the resin was 1200 kg/m3, as measured according to ASTM D 792. The melt flow rate was 3 g/10 min, as measured according to ASTM D 1238. 2.2. Composite fabrication The prepared EG and/or wEG and PC resin samples were weighed to the target contents as shown in Table 1. They were put into a HAAKE Rheomix internal Mixer (HAAKE™ Rheomix OS Lab Mixers, Thermo Scientific Inc., Marietta, GA, USA) heated to 260 °C and mixed at a constant screw speed of 60 rpm for 30 min, and then pelletized. The fabricated pellets were put into a mold (2.5 cm2 × 2 mm) for the thermal conductivity tests. The pellets were pressed using a heating press (D3P-30J, Daheung Science, Incheon, Korea) at a pressure of 15 MPa under a temperature of 260 °C for 15 min and were then quenched to 30 °C using cold water.
2.3. Characterization 2.3.1. EG filler For an analysis of the surface functional groups of the EG fillers, Fourier transform infrared spectroscopy (FT-IR, Nicolet 6700, Thermo Scientific, MA, USA) and X-ray photoelectron spectroscopy (XPS, KAlpha, Thermo Scientific, Massachusetts, USA) were used. The FT-IR spectra were measured in the range of 1000–4000 cm−1 at a resolution of 16 cm−1. The XPS spectra were obtained with an Al X-ray source at a pressure of 1 × 10−8 Pa. To evaluate the defect level of the EG fillers, Raman spectroscopy (LabRAM HR 800, HORIBA Jobin Yvon, Japan) was used along with a 514 nm Ar ion laser. 2.3.2. Morphology The fabricated composites were fractured using liquid nitrogen to obtain the samples. The EG, wEG and composite samples were coated with platinum for 120 s under a vacuum using a sputter-coating machine (Ion Sputter E-1030, Hitachi High Technologies, Tokyo, Japan). In this case, 10 kV of electricity was applied to the coated samples under a nitrogen vacuum using a field emission scanning microscope (FE-SEM, Nova NanoSEM 450, FEI Corp., OR, USA).
Fig. 1. SEM images of EG fillers with respect to treatment time of the high-speed crusher. (a) Raw EG, (b) 10 s, (c) 20 s, and (d) 30 s.
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
2.3.3. Thermal conductivity The thermal conductivity of the composites was measured at ambient temperature and normal pressure using a thermal conductivity measurement instrument (TPS 2500S, Hot Disk ab, Gothenburg, Sweden) according to the ISO 22007-2 standard. The sensor used was constructed with dual spirals made of thin nickel wires and was operated by a sequential plane heat source. In order to measure the change in the temperature of the sensor itself, using the change in the sensor's resistance, the sensor was supplied with a certain level of electric power (P), which induced a temperature rise (ΔT). The thermal conductivity of the samples was obtained from the Fourier thermal conductivity formula based on the supplied power and induced temperature changes. The reproducibility and accuracy of the instrument are better than 1 and 5%, respectively [21,22].
77
2000 × 1332, and the X-ray source was applied under a voltage of 51 kV, a current of 194 μA, and at a normal pressure level. 2.3.5. Heat dissipation To measure the change in the temperature of the fabricated composites during the heating and cooling, an infrared camera (FLIR T420, Wilsonville, OR, USA) was used for the evaluation. The composite samples were put on a hot plate and heated to 100 °C. After 20 s, their heat dissipation image was recorded. In addition, the composite samples heated to 100 °C were placed on a plate at ambient temperature and then, after 20 s, their heat dissipation image was recorded [23,24]. 3. Results and discussion
2.3.4. Tomography A micro-CT device (Skyscan 1172, Bruker Co., Billerica, MA, USA) was used to evaluate the dispersion and network structure of the fillers inside the composites. Measurements were taken at a pixel size of
Fig. 2. (a) FT-IR, (b) XPS and (c) Raman spectra of EG fillers with respect to treatment time of the high-speed crusher.
As shown in Fig. 1, the raw EG becomes fragmented from the EG size of larger than 300 μm to the wEG size of 100–200 μm and smoother with respect to the treatment time in the high-speed crusher of 30 s, which is the processing time recommended by the manufacturer. In the FT-IR spectra as shown in Fig. 2(a), similar characteristic peaks of O\\H, C_C and C\\O were observed, and C/O ratio in the XPS spectra shown in Fig. 2(b) was increased as the O peak was reduced. These findings indicate that oxygen groups introduced onto the surface of the raw EG, formed by the sequential processes of the oxidation and heat treatment of the graphite based precursor materials, were removed by the high-speed crusher treatment. Raman spectroscopy is a useful tool for identifying the defect levels of carbon materials, and Fig. 2(c) shows the Raman spectra of the EG and wEG fillers. In the Raman spectra, the D and G bands were observed at ~1338 and ~1572 cm−1, respectively. The D-band is a disorder induced feature that arises from a double-resonance Raman scattering process from non-zero-center phonon modes, which are typically attributed to the presence of amorphous or disordered carbon atoms [12]. The G band results from in-plane tangential stretching of the carbon-carbon bonds in the graphene sheets [12]. Accordingly, the defect level is lowered as the intensity ratio of the D and G bands (ID/IG) is lowered, and the defect level of the raw EG was reduced during the high-speed crusher treatment. Therefore, the raw EG became smaller and smoother, and the oxidation groups were removed from the surface, leading to a reduced defect level after the high-speed crusher treatment. Fig. 3 shows the thermal conductivity of the composites filled with the EG and wEG fillers. The thermal conductivity increases as the filler content is increased. The thermal conductivity of the E20 composite was 2.34 W/m·K, an enhancement of 1014% over the thermal conductivity of the PC resin (0.21 W/m·K). The thermal conductivity of the w20
Fig. 3. Thermal conductivity of polymer composites filled with the EG and/or wEG fillers. The inset image shows the measuring instrument.
78
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
Fig. 4. (a) 2D micro-CT image, (b) micro-CT reconstruction image, (c) and (d) 3D micro-CT images of the E10w10 composite.
Fig. 5. (a) SEM and (b) micro-CT images of the E5 composite, (c) SEM and (d) micro-CT images of the E15 composite, (e) SEM and (f) micro-CT images of the w5 composite, and (g) SEM and (h) micro-CT images of the w15 composite.
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
composite was 2.17 W/m·K, which was a relatively low enhancement. The defect level of the fillers may be the critical physical factor in determining the thermal conductivity of composites because higher thermal conductivity of the filler can be expected as the defect level is reduced [25,26]. From this perspective, higher thermal conductivity of the composites filled with wEG can be expected. However, this unexpected result can be caused by interfacial thermal resistance between the EG filler and the polymer resin, and the interfacial thermal resistance negatively affects the phonon transfer between the filler and the resin [27–30]. In particular, it is considered that the wEG filler is smaller in size than the raw EG filler and, given the same content, forms more interfaces with the resin within the composites. Consequently, the thermal conductivity of the composites filled with wEG is lower. The thermal conductivity of the E10w10 composite was 2.62 W/m·K, which represents an enhancement by 12.0 and 20.7% compared to that of the E20 and w20 composites, respectively. This implies that the thermal conductivity of the composites filled with both EG and wEG of different sizes is synergistically improved over that of the composites filled with the homogeneously sized EG only. To investigate the cause of the synergistic enhancement in the thermal conductivity, an accurate analysis of the 3D network of the thermally conductive fillers is required. A morphological study based on scanning electron microscopy of the fracture surface of the composites is usually adopted to investigate the internal structures of composites. However, this method is associated with alterations of the internal structures of the samples during the sampling process, as well as image distortion owing to the two-dimensional (2D) image analysis. For a precise analysis of the internal structure of the composites, the 3D non-destructive analysis method should be utilized. A micro-CT analysis based on an X-ray source is a useful analysis tool which can meet the requirements of a precise 3D non-destructive
79
analysis. The 2D micro-CT images were measured sequentially in the direction of the sample thickness, as shown in Fig. 4(a), with the results then reconstructed as 3D images, as shown in Fig. 4(b). In this way, it was possible to measure the internal structure within the composites in a non-destructive manner, as shown in Fig. 4(c) and (d), and determine how the EG fillers are distributed and oriented. However, it is difficult to identify the internal structure of the composites according to the contents and sizes of the EG fillers merely by studying the fracture surface, as shown in the SEM images in Figs. 5 and 6. Therefore, in this study, it is sought to identify the dispersion and 3D network of the EG fillers in the composites in both the SEM and 3D micro-CT analyses. In the micro-CT images shown in Fig. 4(c) and (d), the dispersion and orientation of the EG and wEG fillers can be clearly observed. The EG and wEG with the interlayer spacings are denoted here by the blue and red circles, respectively. The micro-CT image of the composites in Fig. 5 shows a 3D thermal conduction network that was formed as the EG filler content increased inside the composites. A better 3D thermal conduction network of the composites filled with both the EG and wEG fillers was formed than that of the composites filled with EG filler alone when considering the empty spaces inside the composites without fillers, as represented by the red marks in Fig. 6(b) and (d). Therefore, it can be inferred that the synergistically improved thermal conductivity of the composites containing both the EG and wEG fillers is caused by the efficient structure of the 3D thermally conductive pathways [31]. Fig. 7 shows the heat dissipation properties of the composites. This indicates that the composites filled with both the EG and wEG fillers were most sensitive to the temperature changes caused by heating and cooling. This results show that the composites with high thermal conductivity values were excellent for heat dissipation and that the heat dissipation trend was consistent with the thermal conductivity trend. Accordingly,
Fig. 6. (a) SEM and (b) micro-CT images of the E20 composite, (c) SEM and (d) micro-CT images of the E10w10 composite, and (e) SEM and (f) micro-CT images of the w20 composite.
80
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81
Fig. 7. (a) Infrared camera images of composites filled with the EG and/or wEG fillers during heating and cooling for transient temperature response and heat transport, and temperaturetime profiles of the composites during (b) heating and (c) cooling.
it was found that optimizing the filler and the efficient construction of a 3D thermally conductive filler network are the most critical physical factors for optimizing the thermal conductivity of composites and for maximally tuning their heat dissipation capabilities. 4. Conclusions EG is considered as an effective filler for improving the thermal conductivity of composites owing to its unique structure of graphite sheets with interlayer spacings. It can efficiently fill polymer matrix and form thermal transfer pathways. In this study, in order to optimize the thermal conductivity of composites filled with EG fillers, wEG fillers which had been cut into small pieces using a high-speed crusher were prepared and investigated, as the high-speed crusher treatment was considered efficient in terms of both the process time and the cost. The thermal conductivity of the composite filled with both 10 wt% EG and wEG fillers was improved compared to those of the composites filled with only one type of filler at 20 wt%. The filler network structure was precisely observed in 3D non-destructive micro-CT analysis, which confirmed that there was a significant effect on the thermal conductivity of the hybrid composites filled with two types of fillers. Higher thermal conductivity of the composites was observed as a better 3D network of the fillers was formed, thus providing more efficient pathways to transfer phonons. These thermal conductivity results were in good agreement with the heat dissipation results during the heating and cooling process. Acknowledgements This study was supported by Korea Institute of Science and Technology (KIST) Institutional Program and the Technological innovation R&D program of SMBA [S2177379].
References [1] T. Evgin, H.D. Koca, N. Horny, A. Turgut, I.H. Tavman, M. Chirtoc, M. Omastová, I. Novak, Effect of aspect ratio on thermal conductivity of high density polyethylene/ multi-walled carbon nanotubes nanocomposites, Compos. A: Appl. Sci. Manuf. 82 (2016) 208. [2] Z. Wang, R. Qi, J. Wang, S. Qi, Thermal conductivity improvement of epoxy composite filled with expanded graphite, Ceram. Int. 41 (2015) 13541. [3] H.S. Kim, J.-u. Jang, J. Yu, S.Y. Kim, Thermal conductivity of polymer composites based on the length of multi-walled carbon nanotubes, Compos. Part B 79 (2015) 505. [4] J.G. Cevallos, A.E. Bergles, A. Bar-cohen, P. Rodgers, S.K. Gupta, Polymer heat exchangers-history, opportunities, and challenges, Heat Transfer Eng. 33 (2012) 1075. [5] Y.J. Noh, S.Y. Kim, Synergistic improvement of thermal conductivity in polymer composites filled with pitch based carbon fiber and graphene nanoplatelets, Polym. Test. 45 (2015) 132. [6] Z. Han, A. Fina, Thermal conductivity of carbon nanotubes and their polymer nanocomposites: a review, Prog. Polym. Sci. 36 (2011) 914. [7] C. Yu, L. Shi, Z. Yao, D. Li, A. Majumdar, Thermal conductance and thermopower of an individual single-wall carbon nanotube, Nano Lett. 5 (2005) 1842. [8] P. Kim, L. Shi, A. Majumdar, P.L. McEuen, Thermal transport measurements of individual multiwalled nanotubes, Phys. Rev. Lett. 87 (2001) 215502. [9] K.M.F. Shahil, A.A. Balandin, Graphene-multilayer graphene nanocomposites as highly efficient thermal interface materials, Nano Lett. 12 (2012) 861. [10] A.A. Balandin, Thermal properties of graphene and nanostructured carbon materials, Nat. Mater. 10 (2011) 569–581. [11] A.A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, C.N. Lau, Superior thermal conductivity of sigle-layer garphene, Nano Lett. 8 (2008) 902. [12] S.Y. Pak, H.M. Kim, S.Y. Kim, J.R. Youn, Synergistic improvement of thermal conductivity of thermoplastic composites with mixed boron nitride and multi-walled carbon nanotube fillers, Carbon 50 (2012) 4830. [13] S.Y. Kim, Y.J. Noh, J. Yu, Improved thermal conductivity of polymeric composites fabricated by solvent-free processing for the enhanced dispersion of nanofillers and a theoretical approach for composites containing multiple heterogeneities and geometrized nanofillers, Compos. Sci. Technol. 101 (2014) 79. [14] S.Y. Kim, Y.J. Noh, J. Yu, Thermal conductivity of graphene nanoplatelets filled composites fabricated by solvent-free processing for the excellent filler dispersion and a theoretical approach for the composites containing the geometrized fillers, Compos. A: Appl. Sci. Manuf. 69 (2015) 219. [15] Y.J. Noh, H.S. Kim, B.-C. Ku, M.-S. Khil, S.Y. Kim, Thermal conductivity of polymer composites with geometric characteristics of carbon allotropes, Adv. Eng. Mater. 18 (2016) 1127.
H.S. Kim et al. / Journal of Non-Crystalline Solids 450 (2016) 75–81 [16] S.Y. Kim, Y.J. Noh, J. Yu, Prediction and experimental validation of electrical percolation by applying a modified micromechanics model considering multiple heterogeneous inclusions, Compos. Sci. Technol. 106 (2015) 156. [17] Y.J. Noh, S.Y. Pak, S.H. Hwang, J.Y. Hwang, S.Y. Kim, J.R. Youn, Enhanced dispersion for electrical percolation behavior of multi-walled carbon nanotubes in polymer nanocomposites using simple powder mixing and in situ polymerization with surface treatment of the fillers, Compos. Sci. Technol. 89 (2013) 29. [18] M. Chirtoc, N. Horny, I. Tavman, A. Turgut, I. Kökey, M. Omastová, Preparation and photothermal characterization of nanocomposites based on high density polyethylene filled with expanded and unexpanded graphite: particle size and shape effects, Int. J. Therm. Sci. 62 (2012) 50. [19] J. Kratochvíla, A. Boudenne, I. Krupa, Effect of filler size on thermophysical and electrical behavior of nanocomposites based on expanded graphite nanoparticles filled in low-density polyethylene matrix, Polym. Compos. 34 (2013) 149. [20] K.C. Yung, H. Liem, Enhanced thermal conductivity of boron nitride epoxy-matrix composite through multi-modal particle size mixing, J. Appl. Polym. Sci. 160 (2007) 3587. [21] M. Gustavsson, E. Karawacki, S.E. Gustafsson, Thermal conductivity, thermal diffusivity, and specific heat of thin samples from transient measurements with hot disk sensors, Rev. Sci. Instrum. 65 (1994) 3856. [22] V. Bohac, M.K. Gustavsson, L. Kubicar, S.E. Gustafsson, Parameter estimations for measurements of thermal transport properties with the hot disk thermal constants analyzer, Rev. Sci. Instrum. 71 (2000) 2452.
81
[23] H.S. Kim, H.S. Bae, J. Yu, S.Y. Kim, Thermal conductivity of polymer composites with the geometrical characteristics of graphene nanoplatelets, Sci Rep. 6 (2016) 26825. [24] E.-C. Cho, J.-H. Huang, C.-P. Li, C.-W. Chang-Jian, K.-C. Lee, Y.-S. Hsiao, J.-H. Huang, Graphene-based thermoplastic composites and their application for LED thermal management, Carbon 102 (2016) 66. [25] N.-J. Song, C.-M. Chen, C. Lu, Z. Liu, Q.-Q. Kong, R. Cai, Thermally reduced graphene oxide films as flexible lateral heat spreaders, J. Mater. Chem. A 39 (2014) 16563. [26] G. Xin, H. Sun, S.P. Scott, T. Yao, F. Lu, D. Shao, T. Hu, G. Wang, G. Ran, J. Lian, Advanced phase change composite by thermally annealed defect-free graphene for thermal energy storage, ACS Appl. Mater. Interfaces 6 (2014) 15262. [27] S.T. Huxtable, D.G. Cahill, S. Shenogin, L. Xue, R. Ozisik, P. Barone, M. Usrey, M.S. Strano, G. Siddons, M. Shim, P. Keblinski, Interfacial heat flow in carbon nanotube suspensions, Nat. Mater. 2 (2003) 731. [28] C.-W. Nan, Z. Shi, Y. Lin, A simple model for thermal conductivity of carbon nanotube-based composites, Chem. Phys. Lett. 375 (2003) 666. [29] S. Shenogin, L. Xue, R. Ozisik, P. Keblinski, D.G. Cahill, Role of thermal boundary resistance on the heat flow in carbon-nanotube composites, J. Appl. Phys. 95 (2004) 8136. [30] N. Shenogina, S. Shenogin, L. Xue, P. Keblinski, On the lack of thermal percolation in carbon nanotube composites, Appl. Phys. Lett. 87 (2005) 133106. [31] U. Szeluga, B. Kumanek, B. Trzebicka, Synergy in hybrid polymer/nanocarbon composites. A review, Compos. A: Appl. Sci. Manuf. 73 (2015) 20.