- Email: [email protected]
Synergistic effect of spherical Al2O3 particles and BN nanoplates on the thermal transport properties of polymer composites
Composites: Part A 98 (2017) 184–191
Contents lists available at ScienceDirect
Composites: Part A journal homepage: www.elsevier.com/locate/compositesa
Synergistic effect of spherical Al2O3 particles and BN nanoplates on the thermal transport properties of polymer composites Young-Kuk Kim a,⇑, Jae-Yong Chung a, Jung-Goo Lee a, Youn-Kyung Baek a, Pyoung-Woo Shin b a b
Korea Institute of Materials Science (KIMS), #797 Changwondaero, Changwon, Kyungnam 641-010, South Korea Changwon National University, #9 Sarimdong, Changwon, Kyungnam 641-773, South Korea
a r t i c l e
i n f o
Article history: Received 25 October 2016 Received in revised form 8 February 2017 Accepted 26 March 2017 Available online 27 March 2017 Keywords: Al2O3 Hexagonal boron nitride Nanoplatelets Surface
a b s t r a c t Efficient heat transport along through-plane direction is one of the primary requisites for thermal interface materials (TIMs) to relieve heat accumulation at the interface between chip and heat sinks. We report enhanced thermal conduction of Al2O3-based polymer composites by surface wetting and texturing of thermally conductive hexagonal boron nitride (h-BN) nanoplatelets with large anisotropy (diameter <500 nm, thickness <30 nm) in morphology and physical properties. The thermally conductive polymer composites are prepared with hybrid fillers of Al2O3 macrobeads and surface modified h-BN nanoplatelets. Here, the clustering or assembly of h-BN nanoplatelets is analyzed based on depletion interaction of colloidal particles. In addition, further addition of minimal amount of SiO2 nanoparticles shows a drastic improvement of thermal transport properties, which is attributed to the depletion interaction between nanoplatelets mediated by spherical nanoparticles. Here, the benefits of surface wetting for thermal management composite materials are illustrated. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction With development of modern microelectronic technology, power density of electronic devices is rapidly increasing due to miniaturization or integration of device elements, operation at high frequency, high power conditions [1–3]. Thermal problems are known to cause power leakage, device failure and deterioration of performances [3,4]. Thus, efficient removal of accumulated heat is essential to maintain device efficiency and to ensure long lifetime span. To meet the requirement, various thermal management materials have been devised and employed as a part of electronic packaging. Among the various thermal management materials for electronic devices, thermal interface material (TIM) which fills the air gap between chips and heat sinks suppresses the thermal bottleneck at the interface [4–6]. To fill the interfacial gaps and to extract heat generated from electronic devices, composites consisting of ceramic fillers and polymer matrix are popularly selected as TIMs owing to their moderate thermal conductivity, gap-filling capability and easy processing ability with low cost. To reinforce thermal and mechanical properties of polymers with poor thermal conducting ability but high flexibility, ceramic fillers with high thermal conductivity and excellent electrical resistivity are often ⇑ Corresponding author at: Korea Institute of Materials Science (KIMS), #797 Changwondaero, Changwon, Kyungnam 641-010, South Korea. E-mail address: [email protected] (Y.-K. Kim). http://dx.doi.org/10.1016/j.compositesa.2017.03.030 1359-835X/Ó 2017 Elsevier Ltd. All rights reserved.
employed [5–7]. Numerous inorganic particles such as diamond, alumina (Al2O3), aluminium nitride (AlN), boron nitride (BN), etc. have been employed as thermally conductive fillers and BN particles with hexagonal crystal structure (h-BN) are among the most effective fillers for thermally conducting polymer composites due to their high chemical stability and unique physical properties including high thermal conductivity and electrical resistivity. In addition, h-BN belongs to the group of crystals with 2D layered structure and manifests a significant anisotropy in physical properties including thermal conductivity, coefficient of thermal expansion [8]. As in graphene-based polymer composites, high in-plane thermal conductivity exceeding 100 W/mK was already reported for BN-containing polymer sheet [9]. Typically, h-BN is supplied as powders of crystalline platelet having high aspect ratio and h-BN fillers in polymer composites often exhibit preferential alignment along in-plane axis of the composites, resulting in large in-plane thermal conductivity [10]. However, many applications of thermally conductive composites such as thermal interface materials, substrates, etc. popularly require efficient through-plane heat transfer. So, highly conductive axis of h-BN fillers is required to be aligned along through-plane direction and various attempts have been tried to align h-BN platelets in the polymer composite by applying external electric or magnetic field [10–13]. However, only small through-plane thermal diffusivity less than 0.5 mm2/s was reported by h-BN containing
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
polymer composites assisted by external field. Recently, highly improved thermal diffusivity of 1 mm2/s along plane-normal direction of h-BN/silicone rubber composites were reported using aligned BN plates induced by directionally growing ice front [14]. Multi-component fillers with different morphology, dimension and size are often used to improve thermal transport properties of polymer composites due to their capability of forming formation of efficient thermal networks [15–17]. In this study, we demonstrate highly enhanced through-plane thermal conductivity of polymer composites by surface wetting or segregation of BN nanoplatelets with high thermal conductivity on the surface of spherical Al2O3 particles. Additionally, further improvement in the thermal diffusivity of polymer composite with multi-component fillers is achieved by adding small amount of silica nanoparticles, which can be ascribed to enhanced colloidal interaction between nanoplatelets. 2. Experimental approach Spherical Alumina (Al2O3, 45 lm, DAW-45) particles were purchased from Denka (Japan) and BN powders (1 lm, purity 98%) were purchased from Sigma-Aldrich (USA). Polydimethylsiloxane (PDMS, Sylgard 184 elastomer kit) was obtained from Dow cornings (USA) and methyl ethyl ketone (MEK) was purchased from Daejung chemicals (Korea). The epoxy resin (diglycidyl ether of bisphenol F-DGEBF, YD-170, E.E.W = 167.5 g/eq) was kindly donated by Kukdo chemicals (Korea) and 4,4-diamino diphenyl methane (DDM, TCI) was purchased from Sejin chemical (Korea). 2.1. Preparation of polymer composites Spherical alumina particles and BN powders were used as fillers for polymer composites. Spherical Alumina was used as received and BN nanoplatelets were prepared by thermal treatment at 900 °C in air to remove adsorbed organic contaminant and to introduce AOH groups to improve surface properties. After thermal treatment, BN powders were rinsed with deionized water to remove excess boron oxide. Typically, diameters of final BN powders are below 500 nm as characterized with TEM images. In polymer composites, PDMS was used as polymer matrix. To prepare uniform polymer composites, initially, mixture of silicone resin containing base resin and curing agent (containing Pt-catalyst) in weight ratio of 10:1 was simultaneously dissolved in MEK and mixed with inorganic fillers using a planetary mixer (ARM-310, Thinky, Japan) with 2000 rpm in rotation speed for 3 min. After mixing, the solvent was removed by evaporation in vacuum desiccator for 1 h. The dried mixture of polymers and fillers was casted with a Teflon mold and subsequently, it was cured at 80 °C for 1 h in drying oven. Sometimes, colloidal silica was added to improve colloidal interaction between particles and resultant thermal properties of composites. BN nanoplatelets are mixed with colloidal solution of silica nanoparticles (colloidal silica HS-40, SigmaAldrich) having average size of 12 nm in diameter. Subsequently, the mixture was dried in air for further processing to prepare polymer composites. 2.2. Characterization The bulk density (q) of the specimens was measured from the ratio of actual mass of specimen to difference in between mass of specimen and apparent mass measured with the sample fully immersed in distilled water based on Archimedes’ principle. The thermal diffusivity (DT) was measured at room temperature with disk samples using laser flash method (LFA467, Netzsch Instruments Co.). The specific heat (Cp) was measured with the same
185
instrument by scanning various temperatures from 25 °C to 50 °C using pyroceram 9606 as a standard reference material. Cp of test sample was obtained by comparing properties of test sample of under investigation to known physical properties of reference material from the following equation:
C p;means ¼
DT ref ðqref lref ÞC p;ref DT means ðqmeans lmeans Þ
ð1Þ
where DT, l and q are temperature change during laser flash, thickness and density, respectively. Here, the subscript of ‘‘ref” and ‘‘meas” refer to the reference material and test sample under measurement. The thermal conductivity (K) was calculated with formula: K = DT x Cp x q. The shape and the crystal structure of BN nanoplatelets were checked with transmission electron microscopy (TEM) on a JEOL 2100F. Fourier transform infrared spectroscopy (FT-IR) measurements were carried out with a Nicolet iS5 spectrometer (Thermoscientific, USA) ranging from 4000 to 400 cm1 at room temperature using ZnSe pellets. In order to characterize spatial distribution of BN nanoplatelets in the polymer composites, polymeric composites with the same composition were prepared using epoxy polymer which was prepared by mixing diglycidylether of bisphenol F (YDF-170, Kuk-do chemicals) and 2-ethyl-4-methyl imidazole (95%, Sigma-Aldrich) in weight ratio of 4:1. The specimens were cut and polished with diamond slurry to prepare flat surface for microscopy and spectral characterization. We employed a Raman spectrometer with confocal microscopy (Lab-Ram HR, Horiba-Jovin Yvon) operating with accumulation time of 30 s and aperture having 50 lm pinhole through a 100x objective lens and a diffraction grating of 1800 grooves/mm. A Raman spectrum of BN-containing polymer matrix was obtained between two Al2O3 particles using a 514 nm-laser. The intensity variation of the characteristic spectrum at 1360 cm1 for BN was collected with scanning increment of 0.2 lm and the spatial resolution was approximately 1 lm. Polarized optical microscopy was done with optical microscope (Eclipse MA200, Nikon) equipped with CCD and images were captured with software. 3. Results and discussion 3.1. Morphology and thermal properties of polymer composites containing macroscopic Al2O3 spheres as thermally conductive fillers Polymer composite containing spherical a-Al2O3 as thermally conductive fillers were prepared by simple casting. Here, polydimethylsiloxane (PDMS) is adopted as matrix for polymer composites. The through-plane thermal conductivity of polymer composites increases as filler loading increases and is analyzed with theoretical expectation to check the change in thermal conductivity. Among the various theoretical model regarding to effective thermal conductivity of polymer composites, HashinShtrikman (HS) model provides upper (HS+) and lower boundary (HS) of thermal conductivity of mixtures which can be described as [18]
KHSþ ¼
2Kf þ Kp 2/p ðKf Kp Þ Kf 2Kf þ Kp þ /p ðKf Kp Þ
ð2Þ
KHS ¼
2Kp þ Kf 2/f ðKp Kf Þ Kp 2Kp þ Kf þ /f ðKp Kf Þ
ð3Þ
where Kp, Kf denote thermal conductivity of polymer matrix and fillers, respectively. In addition, /p, /f denote volume fraction of polymer matrix and fillers. Both upper and lower boundary of thermal conductivity of Al2O3–containing composites are calculated from HS formula is shown in Fig. 1 and compared with experimental data. The measured thermal conductivity of polymer composite is
186
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
conductivity of polymer composites is hardly achieved with conventional mold casting process of polymer composites consisting of mold casting and subsequent curing at moderate temperature (80 °C). 3.2. Dispersion of BN nanoplates in the Al2O3-containing polymer composites
Fig. 1. The measured and calculated thermal conductivity of polymer composites as a function of Al2O3 filler loading (/ (Al2O3)). Calculation of thermal conductivity is based on Hashin-Shtrikman model. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
nearly consistent to HS and the overall deviation from HS is trivial, which implies near perfect isolation of highly conductive fillers and poor thermal network formation. The origin of fluctuation in experimental data can be attributed to non-uniform distribution of fillers and interfacial effect, since expectation of HS considers only uniformly distributed fillers and pays no regard to interfacial effect. In addition, the interconnectivity of fillers (Xinterconnectivity) is calculated based on deviation of measured thermal conductivity from upper and lower boundary of HS model as proposed by Schilling and Partzsch [19].
Xinterconnectivity ¼ ðKmeas: KHS Þ=ðKHSþ KHS Þ
ð4Þ
In HS model, HS+ can be interpreted as a highly interconnected 3D network, while HS is a fully separated distribution of thermally conductive fillers. The interconnectivity is a degree to estimate the formation of highly conductive thermal path. As shown in Fig. 2(a), current Al2O3-PDMS composites incorporate large amount of pores. For example, the volume fraction of pores in polymer composites with / (Al2O3) = 0.7 is estimated to be 17.2% from direct comparison of measured bulk density with estimated density of ideal mixture of fillers and polymers. So, pores inside the composite can hinder formation of thermal networks, although more chances of forming thermally conductive path are expected in the absence of pores. Then, it is shown that high thermal
In preparation of thermally conductive composites, adaptation of mixed fillers having different size, shape are proved to be useful to improve heat transport efficiency of polymer composites [20–22]. For example, thermal conductivity of phenol-formaldehyde resin composites containing 60 wt% BN is improved by substituting 30 wt% of BN for tetrapod-shaped ZnO whiskers having less thermal conductivity [22]. Here, ZnO bridges thermal conduction between BN platelets and assists to form conductive networks within the polymer matrix. That is, these low dimensional nanostructured materials are beneficial to enhance the thermal network formation in the polymer composites if provided suitable arrangement. In order to improve the thermal conductivity of polymer composites, 2 dimensional nanomaterials are incorporated as fillers for the polymer composites to form thermal bridges between spherical Al2O3 fillers. BN platelets with diameters of hundred nanometers and thickness of several nanometers show an extreme anisotropy in morphology and physical properties. The bulk thermal conductivity of BN is 300 W/mK along in-plane direction and only a few W/mK along through-plane direction [23]. By virtue of its unique thermal properties, hexagonal boron nitride can be used as thermal bridges between spherical Al2O3 particles if suitable alignment along thermally conductive direction is achieved. Then, mixed fillers of Al2O3 and BN are expected to provide an efficient route for thermal conduction in the polymer composites. However, incorporation of pristine BN into the Al2O3 containing PDMS composites results in unstable and easily decomposable composites. Since the high surface energy of nanosized BN may increase their affinity to airborne hydrocarbon molecules which enhance the hydrophobicity of BN [24], the surface of pristine BN nanoplatelets is incompatible to polydimethylsiloxane (PDMS) which have some hydrophilicity as revealed in their non-zero hydrogen bonding solubility parameter (dhydrogen bonding = 4.1 (MPa)1/2) [25]. In the present work, we use high temperature treatment in air to remove adsorbed hydrophobic molecules and to facilitate uniform dispersion by improved surface compatibility to PDMS polymer matrix. Subsequently, the partially oxidized boron nitride fillers are rinsed with hot water to remove residual
Fig. 2. (a) Measured specific heat of polymer composites and (b) a fractured microscopic image of the composite with / (Al2O3) = 0.7. Here, arrows indicate pores and scale bar denotes 20 lm; (c) Density and estimated porosity of polymer composites as a function of Al2O3 filler loading (/ (Al2O3)). Calculation of density of fully dense composite is done with volume average of fillers and polymer matrix. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
boron oxide as illustrated in Fig. 3(a) [26]. Raman spectra for both pristine and oxidized BN exhibit no identifiable difference which implies conservation of crystal structure and crystallinity of pristine BN after oxidation as shown in Fig. 3(b). Here, oxidation of BN is evidenced by presence of BAOAH vibration in FT-IR spectrum at 3250 cm1 (Fig. 3(c)). A high resolution TEM image of oxidized BN shown in Fig. 3(d) displays clear lattice fringes and hexagonal crystal structure confirmed from its Fourier transformed image. The pretreated BN nanoplatelets are mixed with spherical Al2O3 and incorporated to PDMS matrix. Distribution of Al2O3 particles are shown in the cross-sectioned image of the composite from optical microscopy (Fig. 4(a)). Between Al2O3 particles, BN nanoplatelets are expected to be distributed, which is evidenced with the characteristic Raman spectrum of BN found at 1360 cm1 shown in Fig. 4(b). The distribution of BN nanoplatelets is checked with spatial variation of the characteristic Raman spectral intensity
187
from BN nanoplatelets along the dotted line indicated in optical image shown in Fig. 4(a). The line-scanned intensity of Raman spectrum is maximized at near the interface between Al2O3 and matrix observed in optical microscopy shown in Fig. 4(c). That is, distribution of BN nanoplatelets in the PDMS matrix is inhomogeneous and they are preferentially accumulated at near surface of spherical Al2O3 particles as indicated by the microscopic image shown in Fig. 4(d). During the preparation of composites, Al2O3 and BN are dispersed in the polymer solution which is dissolved in organic solvent to make uniform mixing of inorganic particles and polymers, which is illustrated in Fig. 3(a). Addition of polymers to the stable colloidal solution of solid hard particles often induces a so-called ‘‘depletion attraction” between particles and phase separation can occur at sufficient polymer concentration. The surface of colloidal particles is impenetrable with the polymers and there exist a depletion zone at near
Fig. 3. (a) Scheme of preparing polymer composites with spherical Al2O3 and BN nanoplatelets. (b) Raman and (c) FT-IR spectra of pristine and oxidized BN. Arrow indicates BAOAH vibration. (d) TEM images of oxidized BN. Dotted line in the micrograph is an edge of the platelet. Inset images in the high resolution TEM image are Fourier transform of bright field image of BN. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
188
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
Fig. 4. Morphology and Raman spectra of composites containing 6 vol% of BN nanoplatelets (/ (Al2O3) = 0.7): (a) Optical image, (b) Raman spectra of BN, (c) line mapping of Raman spectra from BN in the composites. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
surface of particles. When two particles approach each other and the depletion zones overlap, there is difference in osmotic pressure between particle-particle gaps and outside region. The osmotic pressure difference gives rise to the depletion force between particles [27,28]. In particular, measured depletion potential is manifested with particle-particle separation at below diameter of platelet particles [29]. Examples of colloidal system showing fluid phase separation or inhomogeneity by depletion force are already reported such as clay suspensions [30] and dispersed gibbsite platelets [31]. When a substrate or wall is introduced, wall-particle interaction also occurs and results in rich phenomena including ordering of particles, wetting of colloidal fluids due to entropic effects [31–33]. Although ordering of particles is an entropydecreasing phenomenon, total entropy of the system is increased by onset of further free volume available for each particle.
Theoretical expectation based on density functional theory and numerical simulation also supports entropic wetting of disk-like colloidal particles at near the hard wall which is impenetrable with colloidal particles [34]. In addition, thickness of wetting layers increases with increasing number density of disk-like particles in the system. Fig. 5 shows the polarized optical microscopy (POM) of the polymer composites. In the absence of BN nanoplatelets, contrast between spherical Al2O3 particles and polymer matrix is trivial irrespective of whether polarizers are crossed or not. In contrast to this, significant contrast between spherical Al2O3 particles and matrix is observed from BN-containing composites in the crossed polarizer configuration in POM. This implies formation of textures by alignment of disk-like particles in the matrix of the polymer composites. Here, the surface of the Al2O3 particles (Davg 45 lm) which is much bigger than BN nanoplatelets
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
189
Fig. 5. Polarized optical microscopy (POM) of cross-sectioned surface of Al2O3-PDMS composites with and without BN nanoplatelets under parallel (//) or crossed (\) polarization setup. Scale bar denotes 25 mm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
(Davg < 0.5 lm) can be regarded as an impenetrable wall in colloidal systems and textured arrangement of disk-like BN nanoplatelets can be attributed to particle-wall interaction. Accumulation at near Al2O3 surface and texturing of BN nanoplatelets having high in-plane thermal conductivity is expected to be beneficial for thermal transport properties of the polymer composites. Fig. 6 shows thermal properties and density of polymer composites with various amount of BN nanoplatelets (/ (Al2O3) = 0.7). Various amounts of BN nanoplatelets are incorporated to Al2O3-containing PDMS composites. The porosity of the polymer composites remains constant up to / (BN) = 0.12 and further addition of BN nanoplatelets significantly reduces the bulk density of polymer composites inducing porosity of the composites. The thermal diffusivity (DT) of the polymer composites increases as the amount of incorporated BN nanoplatelets (diameter 0.5 lm) increases up to 12 vol%. Further BN addition to the composites lowers DT as a consequence of limited thermal connectivity owing to incorporation of pores. DT of 1.9 mm2/s is achieved with Al2O3BN-PDMS composites, which is much higher than that of silicone composites assisted with external field-assisted alignment of BN particles [14]. Here, BN nanoplatelets forms thermal networks
between Al2O3 macroparticles by surface wetting and textured arrangement in polymer composites. In addition, textured arrangement of BN nanoplatelets can provide thermally conductive networks enable to enhance thermal transport of composites. Thermal conductivity (K) of the polymer composites are calculated as a product of thermal diffusivity (DT), density (q), specific heat (Cp). Since specific heat of composites is linearly dependent on the volume fraction of BN according to the Neumann–Kopp additive rule [35], the variation of thermal conductivity is dominated by change in DT and q. Then, the thermal conductivity of the composites is peaked at / (BN) = 0.12 which is consistent to change of DT in the same composites. However, further addition of BN nanoplatelets reduces thermal conductivity, which can be attributed to increased porosity. The level of estimated porosity of composites is maintained at about 17% for BN loading less than / (BN) = 0.12 and further addition of BN increases porosity of composites above 25%. The maximum value of thermal conductivity of K = 3.6 W/mK is achieved with / (BN) = 0.12 by mold casting method without external pressure. Instead of simple casting, composites was formed by pressing at the pressure of 50 MPa and thermal conductivity of the composites is enhanced up to 4.5 W/mK with the same
Fig. 6. (a) Porosity, Thermal diffusivity (DT), (b) thermal conductivity (K) of silicone composites containing different amount of BN (Amount of Al2O3 in composites is fixed to be / (Al2O3) = 0.7). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
190
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
Fig. 7. (a) Thermal diffusivity of Al2O3-BN polymer composites with various amounts of silica nanoparticles. Inset is a TEM image of BN-SiO2 mixtures (without polymer molecules) (b) Thermal conductivity (K) and density of silicone composites containing 1 wt% of silica as a function of BN content (Amount of Al2O3 in composites is fixed to be / (Al2O3) = 0.7). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
amount of loaded BN nanoplatelets. The porosity of composites formed by pressing is reduced to be 13% which is nearly 40% smaller than the porosity of casted one. Here, it is demonstrated that additional control of porosity in the composite can modify thermal transport properties of polymer composites, where the same explanation can be applied to diminished thermal conductivity of composites with extremely higher loading of BN nanoplatelets (/ (BN) > 0.12).
3.3. Effect of nanosized silica addition on the thermal diffusivity of polymer composites It is known that addition of secondary particles of different size or shape affects the properties of colloidal systems. Resultant disparity in size and shape leads to effective attraction between similar particles due to depletion interaction [36]. Here, silica nanoparticles are added to the colloidal system to enhance interaction between colloidal particles. In colloidal system containing platelets, addition of small spherical particles can play a role of ‘‘depletant” which mediate depletion attraction between platelets [37]. As mentioned in the previous section, depletion interaction between colloidal particles or particle-wall assists to enhance thermal transport properties of the polymer composites. Fig. 7 shows effect of silica nanoparticles addition on thermal diffusivity of Al2O3 – BN polymer composites. As shown in the TEM image of mixture of BN nanoplatelets and silica nanoparticles, silica nanoparticles are much smaller than BN nanoplatelets and inhomogeneously dispersed. Phase stability of colloidal system is varied with the ratio of radius of plate to radius of sphere. Demixing or two phase coexistence in large plate-small sphere system is theoretically expected from density functional theory [37]. In particular, the phase behavior of this colloidal system is viewed as being driven by depletion attraction between platelets induced by small spherical particles. The phase separation or inhomogeneity in this polymer composite system is thought to be profitable to their thermal transport properties. Experimental measured thermal diffusivity of Al2O3-BN-PDMS composite (/ (Al2O3) = 0.7, / (BN) = 0.06) is enhanced by 40% after addition of very small amount of silica nanoparticles (1 wt% compared with weight fraction of BN nanoplatelets) as shown in Fig. 7(a). The silica-doped polymer composite has ca. 40% higher thermal conductivity than that of undoped one (Fig. 7(b)). Hence, it is shown that addition of small amount of depletant can improve thermal transport properties of polymer composites.
4. Conclusions In conclusion, we demonstrated synergistic effect of spherical Al2O3 and BN nanoplatelets for efficient thermal transport in polymer composites. Polymer composites with single fillers of spherical Al2O3 have poor network formation due to isolation of conductive Al2O3 fillers by thermally insulating polymer matrix. Thermally treated BN nanoplatelets in oxidative atmosphere are easily incorporated into Al2O3-containing polymer composites formed by simple casting. It is shown that the BN nanoplatelets introduced in the matrix of the polymer composites are accumulated at near surface of spherical Al2O3 fillers. From polarized optical microscopy, textured arrangement of BN nanoplatelets is also exhibited by bright contrast of BN-containing polymer matrix. Both texturing and surface wetting of BN nanoplatelets at near surface are explained with depletion interaction in the colloidal systems and expected to contribute thermal network formation between spherical Al2O3 fillers. Addition of BN nanoplatelets boosts thermal conductivity of Al2O3containing polymer composites up to K = 3.6 W/mK which is more than double the thermal conductivity of pristine polymer composites without BN nanoplatelets. Further addition of less-conductive silica nanoparticles by 1 wt% is shown to be beneficial to augment thermal conductivity of polymer composites.
Acknowledgment This work was partially supported by the basic research program in Korea Institute of Materials Science (KIMS) and partially supported by the Center for Advanced Meta-Materials (CAMM) funded by the Ministry of Science, ICT & Future Planning as Global Frontier Research Project (Code No. NRF-2014M3A6B3063704) in Republic of Korea.
References [1] Tong XC. Advanced materials for thermal management of electronic packaging, Springer Science+Business Media; 2011. [2] Vadivelu MA, Kumar CR, Joshi GM. Polymer composites for thermal management: a review. Compos Interfaces 2016;23(9):847–72. [3] Rotkin SV, Perebeinos V, Petrov AG, Avouris P. An essential mechanism of heat dissipation in carbon nanotube electronics nano lett 2009;9(5):1850–5. [4] Gwinn JP, Webb RL. Performance and testing of thermal interface materials. Microelectron J 2003;34(3):215–22. [5] Hsieh CY, Chung SL. High thermal conductivity epoxy molding compound filled with a combustion synthesized AlN powder. J Appl Polym Sci 2006;102 (5):4734–40.
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191 [6] Wong CP, Bollampally RS. Thermal conductivity, elastic modulus, and coefficient of thermal expansion of polymer composites filled with ceramic particles for electronic packaging. J Appl Polym Sci 1999;74:3396–403. [7] Choi S, Kim J. Thermal conductivity of epoxy composites with a binary-particle system of aluminum oxide and aluminum nitride fillers. Compos Part B 2013;51(16):140–7. [8] Duclaux L, Nysten B, Issi JP, Moore AW. Structure and low-temperature thermal conductivity of pyrolytic boron nitride. Phys Rev B 1992;46 (6):3362–7. [9] Zhu H, Li Y, Fang Z, Xu J, Cao F, Wan J, et al. Highly thermally conductive papers with percolative layered boron nitride nanosheets. ACS Nano 2014;8 (4):3606–13. [10] Lin Z, Liu Y, Raghavan S, Moon KS, Sitaraman SK, Wong CP. Magnetic alignment of hexagonal boron nitride platelets in polymer matrix: toward high performance anisotropic polymer composites for electronic encapsulation. ACS Appl Mater Interfaces 2013;5(15):7633–40. [11] Yuan C, Li L, Duan B, Xie B, Zhu Y, Luo X. Locally reinforced polymer-based composites for efficient heat dissipation of local heat source. Inter J Therm Sci 2016;102:202. [12] Yuan C, Xie B, Huang M, Wu R, Luo X. Thermal conductivity enhancement of platelets aligned composites with volume fraction from 10% to 20%. Inter J Heat Mass Transf 2016;94:20–8. [13] Kim K, Ju H, Kim J. Filler orientation of boron nitride composite via external electric field for thermal conductivity enhancement. Ceram Inter 2016;42 (7):8657–63. [14] Bouville F, Maire E, Deville S. Self-assembly of faceted particles triggered by a moving ice front. Langmuir 2014;30(29):8656–63. [15] Shao L, Shi L, Li X, Song N, Ding P. Synergistic effect of BN and graphene nanosheets in 3D framework on the enhancement of thermal conductive properties of polymeric composites. Composites Sci Tech 2016;135(27):83–91. [16] Xiao Y-J, Wang W-Y, Lin T, Chen X-J, Zhang Y-T, Yang J-H, et al. Largely enhanced thermal conductivity and high dielectric constant of poly(vinylidene fluoride)/boron nitride composites achieved by adding a few carbon nanotubes. J Phys Chem C 2016;120(12):6344. [17] Cui X, Ding P, Zhuang N, Shi L, Song N, Tang S. Thermal conductive and mechanical properties of polymeric composites based on solution-exfoliated boron nitride and graphene nanosheets: a morphology-promoted synergistic effect. ACS Appl Mater Interfaces 2015;7(34):19068–75. [18] Hashin Z, Shtrikman SA. Variational approach to the theory of the elastic behaviour of multiphase materials. J Mech Phys Solids 1963;11(2):127. [19] Schilling FR, Partzsch GM. Quantifying partial melt portion in the crust beneath the central andes and the tibetan plateau. Phys Chem Earth(A) 2001;26(4-5):239–46. [20] Zakaria MdH, Akil MR, Kundus MHA Md, Saleh SS. Enhancement of tensile and thermal properties of epoxy nanocomposites through chemical hybridization of carbon nanotubes and alumina. Composites A 2014;66:109–16. [21] Safdari M, Al-Haik MS. Synergistic electrical and thermal transport properties of hybrid polymeric nanocomposites based on carbon nanotubes and graphite nanoplatelets. Carbon 2013;64:111–21.
191
[22] Yuan FY, Zhang HB, Li X, Yu ZZ. Synergistic effect of boron nitride flakes and tetrapod-shaped ZnO whiskers on the thermal conductivity of electrically insulating phenol formaldehyde composites. Compos A 2013;53:137–44. [23] Sichel EK, Miller RE, Abrahams MS, Buiocchi CJ. Heat capacity and thermal conductivity of hexagonal pyrolytic boron nitride. Phys Rev B 1976;13 (10):4607–11. [24] Boinovich LB, Emelyanenko AM, Pashinin AS, Lee CH, Drelich J, Yap YK. Origins of thermodynamically stable superhydrophobicity of boron nitride nanotubes coatings. Langmuir 2012;28(2):1206–16. [25] Alizadeh M, Abbasi F, Farahi M, Jalili K. Silicone-based hydrogels prepared by interpenetrating polymer network synthesis: swelling properties and confinements effects on the formation kinetics. J Appl Polymer Sci 2011;124 (2):985–92. [26] Yu B, Xing W, Guo W, Qiu S, Wang X, Lo S, et al. Thermal exfoliation of hexagonal boron nitride for effective enhancements on thermal stability, flame retardancy and smoke suppression of epoxy resin nanocomposites via sol–gel process. J Mater Chem A 2016;4:7330–40. [27] Lekkerkerker HNW, Tuinier R. Colloids and the depletion interaction. LLC: Springer Science+Business Media; 2011. [28] Yang S, Yan D, Tan H, Shi A-C. Depletion interaction between two colloidal particles in a nonadsorbing polymer solution. Phys Rev E 2006;74:041808. [29] July C, Kleshchanok D, Lang PR. Depletion interactions caused by polydisperse, hard platelets. Soft Matter 2011;7:6444–50. [30] Dijkstra M, Hansen JP. Madden PA statistical model for the structure and gelation of smectite clay suspensions. Phys Rev E 1997;55(3):3044–53. [31] van der Beek D, Petukhov AV, Oversteegen SM, Vroege GJ, Lekkerkerker HNW. Evidence of the hexagonal columnar liquid-crystal phase of hard colloidal platelets by high-resolution SAXS. Eur Phys J E 2005;16(3):253–8. [32] Dijkstra M, van Roij R. Entropic wetting in colloidal suspensions. J Phys: Condens Matter 2005;17(45):S3507–14. [33] Reich H, Schmidt M. Capillary nematization of hard colloidal platelets confined between two parallel hard walls. J Phys: Condens Matter 2007;19(32):326103. [34] Reich H, Dijkstra M, van Roij R, Schmidt M. Entropic wetting and the free isotropicnematic interface of hard colloidal platelets. J Phys Chem B 2007;111(27):7825–35. [35] Tjahjono M, Schreyer MK, Guo L, Garland M. Determination of the individual specific heat capacities of solids from multi-component powder mixtures and polymorphic mixtures-A combined analysis of quantitative PXRD and calorimetry measurements. J Therm Anal Calorim 2012;108(1):361. [36] Doshi N, Cinacchi G, van Duijneveldt JS, Cosgrove T, Prescott SW, Grillo I, et al. Structure of colloidal sphere–plate mixtures. J Phys: Condens Matter 2011;23 (19):194109. [37] de las Heras D, Schmidt M. Bulk fluid phase behaviour of colloidal platelet– sphere and platelet–polymer mixtures. Phil Trans R Soc A 2013;371:20120259.
Contents lists available at ScienceDirect
Composites: Part A journal homepage: www.elsevier.com/locate/compositesa
Synergistic effect of spherical Al2O3 particles and BN nanoplates on the thermal transport properties of polymer composites Young-Kuk Kim a,⇑, Jae-Yong Chung a, Jung-Goo Lee a, Youn-Kyung Baek a, Pyoung-Woo Shin b a b
Korea Institute of Materials Science (KIMS), #797 Changwondaero, Changwon, Kyungnam 641-010, South Korea Changwon National University, #9 Sarimdong, Changwon, Kyungnam 641-773, South Korea
a r t i c l e
i n f o
Article history: Received 25 October 2016 Received in revised form 8 February 2017 Accepted 26 March 2017 Available online 27 March 2017 Keywords: Al2O3 Hexagonal boron nitride Nanoplatelets Surface
a b s t r a c t Efficient heat transport along through-plane direction is one of the primary requisites for thermal interface materials (TIMs) to relieve heat accumulation at the interface between chip and heat sinks. We report enhanced thermal conduction of Al2O3-based polymer composites by surface wetting and texturing of thermally conductive hexagonal boron nitride (h-BN) nanoplatelets with large anisotropy (diameter <500 nm, thickness <30 nm) in morphology and physical properties. The thermally conductive polymer composites are prepared with hybrid fillers of Al2O3 macrobeads and surface modified h-BN nanoplatelets. Here, the clustering or assembly of h-BN nanoplatelets is analyzed based on depletion interaction of colloidal particles. In addition, further addition of minimal amount of SiO2 nanoparticles shows a drastic improvement of thermal transport properties, which is attributed to the depletion interaction between nanoplatelets mediated by spherical nanoparticles. Here, the benefits of surface wetting for thermal management composite materials are illustrated. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction With development of modern microelectronic technology, power density of electronic devices is rapidly increasing due to miniaturization or integration of device elements, operation at high frequency, high power conditions [1–3]. Thermal problems are known to cause power leakage, device failure and deterioration of performances [3,4]. Thus, efficient removal of accumulated heat is essential to maintain device efficiency and to ensure long lifetime span. To meet the requirement, various thermal management materials have been devised and employed as a part of electronic packaging. Among the various thermal management materials for electronic devices, thermal interface material (TIM) which fills the air gap between chips and heat sinks suppresses the thermal bottleneck at the interface [4–6]. To fill the interfacial gaps and to extract heat generated from electronic devices, composites consisting of ceramic fillers and polymer matrix are popularly selected as TIMs owing to their moderate thermal conductivity, gap-filling capability and easy processing ability with low cost. To reinforce thermal and mechanical properties of polymers with poor thermal conducting ability but high flexibility, ceramic fillers with high thermal conductivity and excellent electrical resistivity are often ⇑ Corresponding author at: Korea Institute of Materials Science (KIMS), #797 Changwondaero, Changwon, Kyungnam 641-010, South Korea. E-mail address: [email protected] (Y.-K. Kim). http://dx.doi.org/10.1016/j.compositesa.2017.03.030 1359-835X/Ó 2017 Elsevier Ltd. All rights reserved.
employed [5–7]. Numerous inorganic particles such as diamond, alumina (Al2O3), aluminium nitride (AlN), boron nitride (BN), etc. have been employed as thermally conductive fillers and BN particles with hexagonal crystal structure (h-BN) are among the most effective fillers for thermally conducting polymer composites due to their high chemical stability and unique physical properties including high thermal conductivity and electrical resistivity. In addition, h-BN belongs to the group of crystals with 2D layered structure and manifests a significant anisotropy in physical properties including thermal conductivity, coefficient of thermal expansion [8]. As in graphene-based polymer composites, high in-plane thermal conductivity exceeding 100 W/mK was already reported for BN-containing polymer sheet [9]. Typically, h-BN is supplied as powders of crystalline platelet having high aspect ratio and h-BN fillers in polymer composites often exhibit preferential alignment along in-plane axis of the composites, resulting in large in-plane thermal conductivity [10]. However, many applications of thermally conductive composites such as thermal interface materials, substrates, etc. popularly require efficient through-plane heat transfer. So, highly conductive axis of h-BN fillers is required to be aligned along through-plane direction and various attempts have been tried to align h-BN platelets in the polymer composite by applying external electric or magnetic field [10–13]. However, only small through-plane thermal diffusivity less than 0.5 mm2/s was reported by h-BN containing
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
polymer composites assisted by external field. Recently, highly improved thermal diffusivity of 1 mm2/s along plane-normal direction of h-BN/silicone rubber composites were reported using aligned BN plates induced by directionally growing ice front [14]. Multi-component fillers with different morphology, dimension and size are often used to improve thermal transport properties of polymer composites due to their capability of forming formation of efficient thermal networks [15–17]. In this study, we demonstrate highly enhanced through-plane thermal conductivity of polymer composites by surface wetting or segregation of BN nanoplatelets with high thermal conductivity on the surface of spherical Al2O3 particles. Additionally, further improvement in the thermal diffusivity of polymer composite with multi-component fillers is achieved by adding small amount of silica nanoparticles, which can be ascribed to enhanced colloidal interaction between nanoplatelets. 2. Experimental approach Spherical Alumina (Al2O3, 45 lm, DAW-45) particles were purchased from Denka (Japan) and BN powders (1 lm, purity 98%) were purchased from Sigma-Aldrich (USA). Polydimethylsiloxane (PDMS, Sylgard 184 elastomer kit) was obtained from Dow cornings (USA) and methyl ethyl ketone (MEK) was purchased from Daejung chemicals (Korea). The epoxy resin (diglycidyl ether of bisphenol F-DGEBF, YD-170, E.E.W = 167.5 g/eq) was kindly donated by Kukdo chemicals (Korea) and 4,4-diamino diphenyl methane (DDM, TCI) was purchased from Sejin chemical (Korea). 2.1. Preparation of polymer composites Spherical alumina particles and BN powders were used as fillers for polymer composites. Spherical Alumina was used as received and BN nanoplatelets were prepared by thermal treatment at 900 °C in air to remove adsorbed organic contaminant and to introduce AOH groups to improve surface properties. After thermal treatment, BN powders were rinsed with deionized water to remove excess boron oxide. Typically, diameters of final BN powders are below 500 nm as characterized with TEM images. In polymer composites, PDMS was used as polymer matrix. To prepare uniform polymer composites, initially, mixture of silicone resin containing base resin and curing agent (containing Pt-catalyst) in weight ratio of 10:1 was simultaneously dissolved in MEK and mixed with inorganic fillers using a planetary mixer (ARM-310, Thinky, Japan) with 2000 rpm in rotation speed for 3 min. After mixing, the solvent was removed by evaporation in vacuum desiccator for 1 h. The dried mixture of polymers and fillers was casted with a Teflon mold and subsequently, it was cured at 80 °C for 1 h in drying oven. Sometimes, colloidal silica was added to improve colloidal interaction between particles and resultant thermal properties of composites. BN nanoplatelets are mixed with colloidal solution of silica nanoparticles (colloidal silica HS-40, SigmaAldrich) having average size of 12 nm in diameter. Subsequently, the mixture was dried in air for further processing to prepare polymer composites. 2.2. Characterization The bulk density (q) of the specimens was measured from the ratio of actual mass of specimen to difference in between mass of specimen and apparent mass measured with the sample fully immersed in distilled water based on Archimedes’ principle. The thermal diffusivity (DT) was measured at room temperature with disk samples using laser flash method (LFA467, Netzsch Instruments Co.). The specific heat (Cp) was measured with the same
185
instrument by scanning various temperatures from 25 °C to 50 °C using pyroceram 9606 as a standard reference material. Cp of test sample was obtained by comparing properties of test sample of under investigation to known physical properties of reference material from the following equation:
C p;means ¼
DT ref ðqref lref ÞC p;ref DT means ðqmeans lmeans Þ
ð1Þ
where DT, l and q are temperature change during laser flash, thickness and density, respectively. Here, the subscript of ‘‘ref” and ‘‘meas” refer to the reference material and test sample under measurement. The thermal conductivity (K) was calculated with formula: K = DT x Cp x q. The shape and the crystal structure of BN nanoplatelets were checked with transmission electron microscopy (TEM) on a JEOL 2100F. Fourier transform infrared spectroscopy (FT-IR) measurements were carried out with a Nicolet iS5 spectrometer (Thermoscientific, USA) ranging from 4000 to 400 cm1 at room temperature using ZnSe pellets. In order to characterize spatial distribution of BN nanoplatelets in the polymer composites, polymeric composites with the same composition were prepared using epoxy polymer which was prepared by mixing diglycidylether of bisphenol F (YDF-170, Kuk-do chemicals) and 2-ethyl-4-methyl imidazole (95%, Sigma-Aldrich) in weight ratio of 4:1. The specimens were cut and polished with diamond slurry to prepare flat surface for microscopy and spectral characterization. We employed a Raman spectrometer with confocal microscopy (Lab-Ram HR, Horiba-Jovin Yvon) operating with accumulation time of 30 s and aperture having 50 lm pinhole through a 100x objective lens and a diffraction grating of 1800 grooves/mm. A Raman spectrum of BN-containing polymer matrix was obtained between two Al2O3 particles using a 514 nm-laser. The intensity variation of the characteristic spectrum at 1360 cm1 for BN was collected with scanning increment of 0.2 lm and the spatial resolution was approximately 1 lm. Polarized optical microscopy was done with optical microscope (Eclipse MA200, Nikon) equipped with CCD and images were captured with software. 3. Results and discussion 3.1. Morphology and thermal properties of polymer composites containing macroscopic Al2O3 spheres as thermally conductive fillers Polymer composite containing spherical a-Al2O3 as thermally conductive fillers were prepared by simple casting. Here, polydimethylsiloxane (PDMS) is adopted as matrix for polymer composites. The through-plane thermal conductivity of polymer composites increases as filler loading increases and is analyzed with theoretical expectation to check the change in thermal conductivity. Among the various theoretical model regarding to effective thermal conductivity of polymer composites, HashinShtrikman (HS) model provides upper (HS+) and lower boundary (HS) of thermal conductivity of mixtures which can be described as [18]
KHSþ ¼
2Kf þ Kp 2/p ðKf Kp Þ Kf 2Kf þ Kp þ /p ðKf Kp Þ
ð2Þ
KHS ¼
2Kp þ Kf 2/f ðKp Kf Þ Kp 2Kp þ Kf þ /f ðKp Kf Þ
ð3Þ
where Kp, Kf denote thermal conductivity of polymer matrix and fillers, respectively. In addition, /p, /f denote volume fraction of polymer matrix and fillers. Both upper and lower boundary of thermal conductivity of Al2O3–containing composites are calculated from HS formula is shown in Fig. 1 and compared with experimental data. The measured thermal conductivity of polymer composite is
186
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
conductivity of polymer composites is hardly achieved with conventional mold casting process of polymer composites consisting of mold casting and subsequent curing at moderate temperature (80 °C). 3.2. Dispersion of BN nanoplates in the Al2O3-containing polymer composites
Fig. 1. The measured and calculated thermal conductivity of polymer composites as a function of Al2O3 filler loading (/ (Al2O3)). Calculation of thermal conductivity is based on Hashin-Shtrikman model. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
nearly consistent to HS and the overall deviation from HS is trivial, which implies near perfect isolation of highly conductive fillers and poor thermal network formation. The origin of fluctuation in experimental data can be attributed to non-uniform distribution of fillers and interfacial effect, since expectation of HS considers only uniformly distributed fillers and pays no regard to interfacial effect. In addition, the interconnectivity of fillers (Xinterconnectivity) is calculated based on deviation of measured thermal conductivity from upper and lower boundary of HS model as proposed by Schilling and Partzsch [19].
Xinterconnectivity ¼ ðKmeas: KHS Þ=ðKHSþ KHS Þ
ð4Þ
In HS model, HS+ can be interpreted as a highly interconnected 3D network, while HS is a fully separated distribution of thermally conductive fillers. The interconnectivity is a degree to estimate the formation of highly conductive thermal path. As shown in Fig. 2(a), current Al2O3-PDMS composites incorporate large amount of pores. For example, the volume fraction of pores in polymer composites with / (Al2O3) = 0.7 is estimated to be 17.2% from direct comparison of measured bulk density with estimated density of ideal mixture of fillers and polymers. So, pores inside the composite can hinder formation of thermal networks, although more chances of forming thermally conductive path are expected in the absence of pores. Then, it is shown that high thermal
In preparation of thermally conductive composites, adaptation of mixed fillers having different size, shape are proved to be useful to improve heat transport efficiency of polymer composites [20–22]. For example, thermal conductivity of phenol-formaldehyde resin composites containing 60 wt% BN is improved by substituting 30 wt% of BN for tetrapod-shaped ZnO whiskers having less thermal conductivity [22]. Here, ZnO bridges thermal conduction between BN platelets and assists to form conductive networks within the polymer matrix. That is, these low dimensional nanostructured materials are beneficial to enhance the thermal network formation in the polymer composites if provided suitable arrangement. In order to improve the thermal conductivity of polymer composites, 2 dimensional nanomaterials are incorporated as fillers for the polymer composites to form thermal bridges between spherical Al2O3 fillers. BN platelets with diameters of hundred nanometers and thickness of several nanometers show an extreme anisotropy in morphology and physical properties. The bulk thermal conductivity of BN is 300 W/mK along in-plane direction and only a few W/mK along through-plane direction [23]. By virtue of its unique thermal properties, hexagonal boron nitride can be used as thermal bridges between spherical Al2O3 particles if suitable alignment along thermally conductive direction is achieved. Then, mixed fillers of Al2O3 and BN are expected to provide an efficient route for thermal conduction in the polymer composites. However, incorporation of pristine BN into the Al2O3 containing PDMS composites results in unstable and easily decomposable composites. Since the high surface energy of nanosized BN may increase their affinity to airborne hydrocarbon molecules which enhance the hydrophobicity of BN [24], the surface of pristine BN nanoplatelets is incompatible to polydimethylsiloxane (PDMS) which have some hydrophilicity as revealed in their non-zero hydrogen bonding solubility parameter (dhydrogen bonding = 4.1 (MPa)1/2) [25]. In the present work, we use high temperature treatment in air to remove adsorbed hydrophobic molecules and to facilitate uniform dispersion by improved surface compatibility to PDMS polymer matrix. Subsequently, the partially oxidized boron nitride fillers are rinsed with hot water to remove residual
Fig. 2. (a) Measured specific heat of polymer composites and (b) a fractured microscopic image of the composite with / (Al2O3) = 0.7. Here, arrows indicate pores and scale bar denotes 20 lm; (c) Density and estimated porosity of polymer composites as a function of Al2O3 filler loading (/ (Al2O3)). Calculation of density of fully dense composite is done with volume average of fillers and polymer matrix. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
boron oxide as illustrated in Fig. 3(a) [26]. Raman spectra for both pristine and oxidized BN exhibit no identifiable difference which implies conservation of crystal structure and crystallinity of pristine BN after oxidation as shown in Fig. 3(b). Here, oxidation of BN is evidenced by presence of BAOAH vibration in FT-IR spectrum at 3250 cm1 (Fig. 3(c)). A high resolution TEM image of oxidized BN shown in Fig. 3(d) displays clear lattice fringes and hexagonal crystal structure confirmed from its Fourier transformed image. The pretreated BN nanoplatelets are mixed with spherical Al2O3 and incorporated to PDMS matrix. Distribution of Al2O3 particles are shown in the cross-sectioned image of the composite from optical microscopy (Fig. 4(a)). Between Al2O3 particles, BN nanoplatelets are expected to be distributed, which is evidenced with the characteristic Raman spectrum of BN found at 1360 cm1 shown in Fig. 4(b). The distribution of BN nanoplatelets is checked with spatial variation of the characteristic Raman spectral intensity
187
from BN nanoplatelets along the dotted line indicated in optical image shown in Fig. 4(a). The line-scanned intensity of Raman spectrum is maximized at near the interface between Al2O3 and matrix observed in optical microscopy shown in Fig. 4(c). That is, distribution of BN nanoplatelets in the PDMS matrix is inhomogeneous and they are preferentially accumulated at near surface of spherical Al2O3 particles as indicated by the microscopic image shown in Fig. 4(d). During the preparation of composites, Al2O3 and BN are dispersed in the polymer solution which is dissolved in organic solvent to make uniform mixing of inorganic particles and polymers, which is illustrated in Fig. 3(a). Addition of polymers to the stable colloidal solution of solid hard particles often induces a so-called ‘‘depletion attraction” between particles and phase separation can occur at sufficient polymer concentration. The surface of colloidal particles is impenetrable with the polymers and there exist a depletion zone at near
Fig. 3. (a) Scheme of preparing polymer composites with spherical Al2O3 and BN nanoplatelets. (b) Raman and (c) FT-IR spectra of pristine and oxidized BN. Arrow indicates BAOAH vibration. (d) TEM images of oxidized BN. Dotted line in the micrograph is an edge of the platelet. Inset images in the high resolution TEM image are Fourier transform of bright field image of BN. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
188
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
Fig. 4. Morphology and Raman spectra of composites containing 6 vol% of BN nanoplatelets (/ (Al2O3) = 0.7): (a) Optical image, (b) Raman spectra of BN, (c) line mapping of Raman spectra from BN in the composites. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
surface of particles. When two particles approach each other and the depletion zones overlap, there is difference in osmotic pressure between particle-particle gaps and outside region. The osmotic pressure difference gives rise to the depletion force between particles [27,28]. In particular, measured depletion potential is manifested with particle-particle separation at below diameter of platelet particles [29]. Examples of colloidal system showing fluid phase separation or inhomogeneity by depletion force are already reported such as clay suspensions [30] and dispersed gibbsite platelets [31]. When a substrate or wall is introduced, wall-particle interaction also occurs and results in rich phenomena including ordering of particles, wetting of colloidal fluids due to entropic effects [31–33]. Although ordering of particles is an entropydecreasing phenomenon, total entropy of the system is increased by onset of further free volume available for each particle.
Theoretical expectation based on density functional theory and numerical simulation also supports entropic wetting of disk-like colloidal particles at near the hard wall which is impenetrable with colloidal particles [34]. In addition, thickness of wetting layers increases with increasing number density of disk-like particles in the system. Fig. 5 shows the polarized optical microscopy (POM) of the polymer composites. In the absence of BN nanoplatelets, contrast between spherical Al2O3 particles and polymer matrix is trivial irrespective of whether polarizers are crossed or not. In contrast to this, significant contrast between spherical Al2O3 particles and matrix is observed from BN-containing composites in the crossed polarizer configuration in POM. This implies formation of textures by alignment of disk-like particles in the matrix of the polymer composites. Here, the surface of the Al2O3 particles (Davg 45 lm) which is much bigger than BN nanoplatelets
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
189
Fig. 5. Polarized optical microscopy (POM) of cross-sectioned surface of Al2O3-PDMS composites with and without BN nanoplatelets under parallel (//) or crossed (\) polarization setup. Scale bar denotes 25 mm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
(Davg < 0.5 lm) can be regarded as an impenetrable wall in colloidal systems and textured arrangement of disk-like BN nanoplatelets can be attributed to particle-wall interaction. Accumulation at near Al2O3 surface and texturing of BN nanoplatelets having high in-plane thermal conductivity is expected to be beneficial for thermal transport properties of the polymer composites. Fig. 6 shows thermal properties and density of polymer composites with various amount of BN nanoplatelets (/ (Al2O3) = 0.7). Various amounts of BN nanoplatelets are incorporated to Al2O3-containing PDMS composites. The porosity of the polymer composites remains constant up to / (BN) = 0.12 and further addition of BN nanoplatelets significantly reduces the bulk density of polymer composites inducing porosity of the composites. The thermal diffusivity (DT) of the polymer composites increases as the amount of incorporated BN nanoplatelets (diameter 0.5 lm) increases up to 12 vol%. Further BN addition to the composites lowers DT as a consequence of limited thermal connectivity owing to incorporation of pores. DT of 1.9 mm2/s is achieved with Al2O3BN-PDMS composites, which is much higher than that of silicone composites assisted with external field-assisted alignment of BN particles [14]. Here, BN nanoplatelets forms thermal networks
between Al2O3 macroparticles by surface wetting and textured arrangement in polymer composites. In addition, textured arrangement of BN nanoplatelets can provide thermally conductive networks enable to enhance thermal transport of composites. Thermal conductivity (K) of the polymer composites are calculated as a product of thermal diffusivity (DT), density (q), specific heat (Cp). Since specific heat of composites is linearly dependent on the volume fraction of BN according to the Neumann–Kopp additive rule [35], the variation of thermal conductivity is dominated by change in DT and q. Then, the thermal conductivity of the composites is peaked at / (BN) = 0.12 which is consistent to change of DT in the same composites. However, further addition of BN nanoplatelets reduces thermal conductivity, which can be attributed to increased porosity. The level of estimated porosity of composites is maintained at about 17% for BN loading less than / (BN) = 0.12 and further addition of BN increases porosity of composites above 25%. The maximum value of thermal conductivity of K = 3.6 W/mK is achieved with / (BN) = 0.12 by mold casting method without external pressure. Instead of simple casting, composites was formed by pressing at the pressure of 50 MPa and thermal conductivity of the composites is enhanced up to 4.5 W/mK with the same
Fig. 6. (a) Porosity, Thermal diffusivity (DT), (b) thermal conductivity (K) of silicone composites containing different amount of BN (Amount of Al2O3 in composites is fixed to be / (Al2O3) = 0.7). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
190
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191
Fig. 7. (a) Thermal diffusivity of Al2O3-BN polymer composites with various amounts of silica nanoparticles. Inset is a TEM image of BN-SiO2 mixtures (without polymer molecules) (b) Thermal conductivity (K) and density of silicone composites containing 1 wt% of silica as a function of BN content (Amount of Al2O3 in composites is fixed to be / (Al2O3) = 0.7). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
amount of loaded BN nanoplatelets. The porosity of composites formed by pressing is reduced to be 13% which is nearly 40% smaller than the porosity of casted one. Here, it is demonstrated that additional control of porosity in the composite can modify thermal transport properties of polymer composites, where the same explanation can be applied to diminished thermal conductivity of composites with extremely higher loading of BN nanoplatelets (/ (BN) > 0.12).
3.3. Effect of nanosized silica addition on the thermal diffusivity of polymer composites It is known that addition of secondary particles of different size or shape affects the properties of colloidal systems. Resultant disparity in size and shape leads to effective attraction between similar particles due to depletion interaction [36]. Here, silica nanoparticles are added to the colloidal system to enhance interaction between colloidal particles. In colloidal system containing platelets, addition of small spherical particles can play a role of ‘‘depletant” which mediate depletion attraction between platelets [37]. As mentioned in the previous section, depletion interaction between colloidal particles or particle-wall assists to enhance thermal transport properties of the polymer composites. Fig. 7 shows effect of silica nanoparticles addition on thermal diffusivity of Al2O3 – BN polymer composites. As shown in the TEM image of mixture of BN nanoplatelets and silica nanoparticles, silica nanoparticles are much smaller than BN nanoplatelets and inhomogeneously dispersed. Phase stability of colloidal system is varied with the ratio of radius of plate to radius of sphere. Demixing or two phase coexistence in large plate-small sphere system is theoretically expected from density functional theory [37]. In particular, the phase behavior of this colloidal system is viewed as being driven by depletion attraction between platelets induced by small spherical particles. The phase separation or inhomogeneity in this polymer composite system is thought to be profitable to their thermal transport properties. Experimental measured thermal diffusivity of Al2O3-BN-PDMS composite (/ (Al2O3) = 0.7, / (BN) = 0.06) is enhanced by 40% after addition of very small amount of silica nanoparticles (1 wt% compared with weight fraction of BN nanoplatelets) as shown in Fig. 7(a). The silica-doped polymer composite has ca. 40% higher thermal conductivity than that of undoped one (Fig. 7(b)). Hence, it is shown that addition of small amount of depletant can improve thermal transport properties of polymer composites.
4. Conclusions In conclusion, we demonstrated synergistic effect of spherical Al2O3 and BN nanoplatelets for efficient thermal transport in polymer composites. Polymer composites with single fillers of spherical Al2O3 have poor network formation due to isolation of conductive Al2O3 fillers by thermally insulating polymer matrix. Thermally treated BN nanoplatelets in oxidative atmosphere are easily incorporated into Al2O3-containing polymer composites formed by simple casting. It is shown that the BN nanoplatelets introduced in the matrix of the polymer composites are accumulated at near surface of spherical Al2O3 fillers. From polarized optical microscopy, textured arrangement of BN nanoplatelets is also exhibited by bright contrast of BN-containing polymer matrix. Both texturing and surface wetting of BN nanoplatelets at near surface are explained with depletion interaction in the colloidal systems and expected to contribute thermal network formation between spherical Al2O3 fillers. Addition of BN nanoplatelets boosts thermal conductivity of Al2O3containing polymer composites up to K = 3.6 W/mK which is more than double the thermal conductivity of pristine polymer composites without BN nanoplatelets. Further addition of less-conductive silica nanoparticles by 1 wt% is shown to be beneficial to augment thermal conductivity of polymer composites.
Acknowledgment This work was partially supported by the basic research program in Korea Institute of Materials Science (KIMS) and partially supported by the Center for Advanced Meta-Materials (CAMM) funded by the Ministry of Science, ICT & Future Planning as Global Frontier Research Project (Code No. NRF-2014M3A6B3063704) in Republic of Korea.
References [1] Tong XC. Advanced materials for thermal management of electronic packaging, Springer Science+Business Media; 2011. [2] Vadivelu MA, Kumar CR, Joshi GM. Polymer composites for thermal management: a review. Compos Interfaces 2016;23(9):847–72. [3] Rotkin SV, Perebeinos V, Petrov AG, Avouris P. An essential mechanism of heat dissipation in carbon nanotube electronics nano lett 2009;9(5):1850–5. [4] Gwinn JP, Webb RL. Performance and testing of thermal interface materials. Microelectron J 2003;34(3):215–22. [5] Hsieh CY, Chung SL. High thermal conductivity epoxy molding compound filled with a combustion synthesized AlN powder. J Appl Polym Sci 2006;102 (5):4734–40.
Y.-K. Kim et al. / Composites: Part A 98 (2017) 184–191 [6] Wong CP, Bollampally RS. Thermal conductivity, elastic modulus, and coefficient of thermal expansion of polymer composites filled with ceramic particles for electronic packaging. J Appl Polym Sci 1999;74:3396–403. [7] Choi S, Kim J. Thermal conductivity of epoxy composites with a binary-particle system of aluminum oxide and aluminum nitride fillers. Compos Part B 2013;51(16):140–7. [8] Duclaux L, Nysten B, Issi JP, Moore AW. Structure and low-temperature thermal conductivity of pyrolytic boron nitride. Phys Rev B 1992;46 (6):3362–7. [9] Zhu H, Li Y, Fang Z, Xu J, Cao F, Wan J, et al. Highly thermally conductive papers with percolative layered boron nitride nanosheets. ACS Nano 2014;8 (4):3606–13. [10] Lin Z, Liu Y, Raghavan S, Moon KS, Sitaraman SK, Wong CP. Magnetic alignment of hexagonal boron nitride platelets in polymer matrix: toward high performance anisotropic polymer composites for electronic encapsulation. ACS Appl Mater Interfaces 2013;5(15):7633–40. [11] Yuan C, Li L, Duan B, Xie B, Zhu Y, Luo X. Locally reinforced polymer-based composites for efficient heat dissipation of local heat source. Inter J Therm Sci 2016;102:202. [12] Yuan C, Xie B, Huang M, Wu R, Luo X. Thermal conductivity enhancement of platelets aligned composites with volume fraction from 10% to 20%. Inter J Heat Mass Transf 2016;94:20–8. [13] Kim K, Ju H, Kim J. Filler orientation of boron nitride composite via external electric field for thermal conductivity enhancement. Ceram Inter 2016;42 (7):8657–63. [14] Bouville F, Maire E, Deville S. Self-assembly of faceted particles triggered by a moving ice front. Langmuir 2014;30(29):8656–63. [15] Shao L, Shi L, Li X, Song N, Ding P. Synergistic effect of BN and graphene nanosheets in 3D framework on the enhancement of thermal conductive properties of polymeric composites. Composites Sci Tech 2016;135(27):83–91. [16] Xiao Y-J, Wang W-Y, Lin T, Chen X-J, Zhang Y-T, Yang J-H, et al. Largely enhanced thermal conductivity and high dielectric constant of poly(vinylidene fluoride)/boron nitride composites achieved by adding a few carbon nanotubes. J Phys Chem C 2016;120(12):6344. [17] Cui X, Ding P, Zhuang N, Shi L, Song N, Tang S. Thermal conductive and mechanical properties of polymeric composites based on solution-exfoliated boron nitride and graphene nanosheets: a morphology-promoted synergistic effect. ACS Appl Mater Interfaces 2015;7(34):19068–75. [18] Hashin Z, Shtrikman SA. Variational approach to the theory of the elastic behaviour of multiphase materials. J Mech Phys Solids 1963;11(2):127. [19] Schilling FR, Partzsch GM. Quantifying partial melt portion in the crust beneath the central andes and the tibetan plateau. Phys Chem Earth(A) 2001;26(4-5):239–46. [20] Zakaria MdH, Akil MR, Kundus MHA Md, Saleh SS. Enhancement of tensile and thermal properties of epoxy nanocomposites through chemical hybridization of carbon nanotubes and alumina. Composites A 2014;66:109–16. [21] Safdari M, Al-Haik MS. Synergistic electrical and thermal transport properties of hybrid polymeric nanocomposites based on carbon nanotubes and graphite nanoplatelets. Carbon 2013;64:111–21.
191
[22] Yuan FY, Zhang HB, Li X, Yu ZZ. Synergistic effect of boron nitride flakes and tetrapod-shaped ZnO whiskers on the thermal conductivity of electrically insulating phenol formaldehyde composites. Compos A 2013;53:137–44. [23] Sichel EK, Miller RE, Abrahams MS, Buiocchi CJ. Heat capacity and thermal conductivity of hexagonal pyrolytic boron nitride. Phys Rev B 1976;13 (10):4607–11. [24] Boinovich LB, Emelyanenko AM, Pashinin AS, Lee CH, Drelich J, Yap YK. Origins of thermodynamically stable superhydrophobicity of boron nitride nanotubes coatings. Langmuir 2012;28(2):1206–16. [25] Alizadeh M, Abbasi F, Farahi M, Jalili K. Silicone-based hydrogels prepared by interpenetrating polymer network synthesis: swelling properties and confinements effects on the formation kinetics. J Appl Polymer Sci 2011;124 (2):985–92. [26] Yu B, Xing W, Guo W, Qiu S, Wang X, Lo S, et al. Thermal exfoliation of hexagonal boron nitride for effective enhancements on thermal stability, flame retardancy and smoke suppression of epoxy resin nanocomposites via sol–gel process. J Mater Chem A 2016;4:7330–40. [27] Lekkerkerker HNW, Tuinier R. Colloids and the depletion interaction. LLC: Springer Science+Business Media; 2011. [28] Yang S, Yan D, Tan H, Shi A-C. Depletion interaction between two colloidal particles in a nonadsorbing polymer solution. Phys Rev E 2006;74:041808. [29] July C, Kleshchanok D, Lang PR. Depletion interactions caused by polydisperse, hard platelets. Soft Matter 2011;7:6444–50. [30] Dijkstra M, Hansen JP. Madden PA statistical model for the structure and gelation of smectite clay suspensions. Phys Rev E 1997;55(3):3044–53. [31] van der Beek D, Petukhov AV, Oversteegen SM, Vroege GJ, Lekkerkerker HNW. Evidence of the hexagonal columnar liquid-crystal phase of hard colloidal platelets by high-resolution SAXS. Eur Phys J E 2005;16(3):253–8. [32] Dijkstra M, van Roij R. Entropic wetting in colloidal suspensions. J Phys: Condens Matter 2005;17(45):S3507–14. [33] Reich H, Schmidt M. Capillary nematization of hard colloidal platelets confined between two parallel hard walls. J Phys: Condens Matter 2007;19(32):326103. [34] Reich H, Dijkstra M, van Roij R, Schmidt M. Entropic wetting and the free isotropicnematic interface of hard colloidal platelets. J Phys Chem B 2007;111(27):7825–35. [35] Tjahjono M, Schreyer MK, Guo L, Garland M. Determination of the individual specific heat capacities of solids from multi-component powder mixtures and polymorphic mixtures-A combined analysis of quantitative PXRD and calorimetry measurements. J Therm Anal Calorim 2012;108(1):361. [36] Doshi N, Cinacchi G, van Duijneveldt JS, Cosgrove T, Prescott SW, Grillo I, et al. Structure of colloidal sphere–plate mixtures. J Phys: Condens Matter 2011;23 (19):194109. [37] de las Heras D, Schmidt M. Bulk fluid phase behaviour of colloidal platelet– sphere and platelet–polymer mixtures. Phil Trans R Soc A 2013;371:20120259.