Highly thermally conductive polymer nanocomposites based on boron nitride nanosheets decorated with silver nanoparticles

Abstract

The development of thermally conductive polymer composites is of critical importance to address the issue of heat aggregation in modern electronics with rapid-increasing power density. However, the thermal conductivity (K) of the polymer composites has long been limited to within 10 W m⁻¹ K⁻¹ due to the high interfacial thermal resistance. Herein, we have demonstrated a remarkable improvement in K upon the incorporation of boron nitride nanosheets doped with silver nanoparticles hybrids (BNNSs/AgNPs). The incorporation of BNNSs/AgNPs into liquid crystalline epoxy resin (LCER) yields an in-plane K of 12.55 W m⁻¹ K⁻¹ with 25.1 vol% boron nitride nanosheets (BNNSs). The interfacial thermal resistance among BNNSs has demonstrated to reduce by sintering silver nanoparticles (AgNPs) deposited on the surface of BNNSs, forming more thermally conductive networks with higher K. As a proof of concept application, the obtained composite was used as the substrate for light-emitting-diode chips and was demonstrated to be more effective in heat removal than the bulk LCER. This study establishes an efficient approach to prepare thermally conductive composites, which can be applied in the next generation of integrated circuits and three-dimensional electronics.

---

Introduction

Thermally conductive yet electrically insulating polymer materials have been considered as a choice in thermal management applications such as solar power generation substrates for commercial portable electronics.^1–3 With the continuing miniaturization of electronics and increased levels of dissipated power in portable devices, heat removal has become extremely important because the intense heat generated can cause device failure.^4–7 Current polymer materials used for conducting heat in electronics typically have a thermal conductivity (K) of 0.2 W m⁻¹ K⁻¹ at room temperature,^8–10 because the amorphous arrangement of the molecular chains will reduce the mean free path of phonons,¹¹ which makes them unfavorable for thermal conduction. Thus, there is an urgent demand for novel thermally conductive materials to provide efficient heat conduction for this challenging issue.

In recent years, lots of studies have shown that the K value of the polymers can be enhanced by introducing ceramic particles such as aluminum nitride,¹² hexagonal boron nitride (h-BN),¹³ silicon nitride¹⁴ and aluminum oxide.¹⁵ Moreover, they maintain a high electrical resistance. More specifically, boron nitride nanosheets (BNNSs), structurally analogous to few-layer graphene, have attracted increasing interest for their high thermal transport, electrically insulating properties, high chemical stability and high oxidation resistance.^16–19 Owing to these unique advantages, BNNSs are considered as excellent nanofillers for thermally conductive yet electrically insulating polymeric composites. Many different strategies have been designed and performed for the incorporation of BNNSs or their derived materials into polymers to improve the K value.^8,20–23 Xie et al. used h-BN microplatelets as thermal conductive fillers for polyvinylalcohol (PVA), whereas the in-plane K value of the h-BN/PVA composite just reached 4.41 W m⁻¹ K⁻¹ at 30 wt% loading, probably due to the poor compatibility between PVA and h-BN.¹³ The construction of covalent bonds between BNNSs and the polymer matrix to improve the compatibility is the most commonly reported strategy used towards the fabrication of composites with BNNSs. The study of Shen et al. demonstrated the surface modification of chemically inert h-BN to form a more efficient thermal pathway in a PVA matrix. However, the enhanced in-plane K value only reached 8.8 W m⁻¹ K⁻¹ with 30 vol% filler.²⁴ In specific, although a large number of examples have arisen, the in-plane K value has long been limited to within 10 W m⁻¹ K⁻¹, mainly due to the high thermal interface resistance found between the filler and the polymer matrix.

Herein, we report a new strategy used to improve the in-plane K value up to 12.5 W m⁻¹ K⁻¹ via the addition of BNNSs decorated with silver nanoparticles (BNNSs/AgNPs) hybrids into liquid crystalline epoxy resin (LCER). It is believed that the BNNSs can form thermally conductive pathways more easily by connections through the sintering of AgNPs during the curing of LCER. In addition, two physical models were applied to simulate the measured K value and the results show a lower thermal boundary resistance between the BNNSs/AgNPs and LCER at high filler loading when compared with raw BNNSs. As a proof of concept application, the obtained composite was used as a substrate for light-emitting-diode (LED) chips and further demonstrated that it was more effective in heat removal than the bulk LCER.

Experimental

Materials

h-BN micropowders (2.0 μm) were purchased from Denka (Japan). Polyvinylpyrrolidone (PVP, M_w, 58 000, k29-32) was purchased from Aladdin Chemistry Co., Ltd. Silver nitrate (AgNO₃, 99.8%) was purchased from Shanghai Lingfeng Chemical reagent Co., Ltd. N,N-Dimethylformamide (DMF, ≧99.5%) was purchased from Xilong Chemical Co., Ltd. The liquid crystalline epoxy resin, 4,4′-bis(4-hydroxybenzoyloxy)-3,3′,5,5′-tetramethyl-(1,1′-bipheyl) (LCER, 100%), was purchased from Gansu Research Institute of Chemical Industry, China. 4,4′-Diaminodiphenylsulphone (DDS) was used as the curing agent and bought from Sinopharm Chemical Reagent Co., Ltd, China. All the other materials were used as received.

Preparation of BNNSs

BNNSs were prepared by a method combining sonication-assisted liquid-phase exfoliation and centrifugation according to a previous report.²⁵ N,N-Dimethylformamide (DMF) was identified as a favorable solvent in the exfoliation of h-BN, from which the stable dispersion of BNNSs was obtained. In detail, the initial 2.0 g of h-BN powders in 200 mL of DMF was sonicated for 48 h, followed by centrifuging at 1000 rpm for 20 minutes to remove the large sediments. A homogeneous white dispersion (2.54 mg mL⁻¹) was obtained for further AgNPs decoration.

Synthesis of the BNNSs/AgNPs hybrids

To stabilize the AgNPs and control their sizes, 0.2 g of polyvinylpyrrolidone (PVP) was added to the BNNSs–DMF suspension under sonication for 30 minutes until it was dissolved completely. DMF is not only a favorable solvent, but also a remarkable reducing agent for silver ions. The mass ratio of AgNO₃ to BNNSs was maintained at 1 : 1. 0.5 g of a AgNO₃ solution was then slowly dropped into the suspension while stirring and heating at 60 °C for 1 h, followed by standing for 24 h to allow the redundant AgNPs' to deposit on the surface of the BNNSs. The BNNSs/AgNPs hybrids were obtained by vacuum filtration and sequentially washed with ethanol, distilled water and acetone followed drying in vacuum oven at 80 °C.

Fabrication of the BNNSs/AgNPs/LCER composites

Liquid crystalline epoxy resin (LCER) and the curing agent (DDS) were first subjected to a pre-curing process at 180 °C for 30 minutes with a mass ratio between LCER and DDS of 13 : 7. Then, a certain amount of BNNSs/AgNPs and pre-cured LCER were dispersed in butanone. The mixture of BNNSs/AgNPs and LCER in butanone was stirred for 2 h and then placed in an ultrasonic bath for 60 min to form a uniform suspension. The LCER with homogeneously dispersed BNNSs/AgNPs was then bar coated on copper film. The BNNSs/AgNPs/LCER composites were obtained after curing at 150 °C, 180 °C, and 220 °C for 2 h, respectively. The content of BNNSs in the BNNSs/LCER composites was varied from 0 to 25.1 vol%. For comparison, BNNSs/LCER and AgNPs/LCER were also prepared using the same procedure as mentioned above. The mass amount of AgNPs in LCER is about 8.98 wt% of BNNSs.

Characterization

Transmission electron microscopy (TEM) images were obtained on a JEOL JEM-2001F transmission electron microscope. The TEM samples were prepared by dropping the BNNSs/AgNPs solution on a copper grid with drying in an oven. Scanning electron microscopy (SEM) images of the BNNSs/AgNPs/LCER composite at 25.1 vol% BNNSs were obtained using a field SEM (Nova NanoSEM 450, FEI) with 10 kV accelerating voltage. X-ray photoelectron spectrometry (XPS) analysis was carried out on a Kratos Axis Ultra DLD with Al Kα radiation (1486.6 eV). Fourier transform infrared spectroscopy (Bruker Vertex 70) was employed to probe the functional groups on the BNNSs surface. X-Ray Diffraction (XRD) analyses of the BNNS/AgNP hybrids were recorded at a scan rate of 0.02° s⁻¹ in the 2θ range of 5–90° using a X-ray powder diffractometer with Cu-K radiation. The in-plane K value was measured using the laser flash technique (NETZSCH, LFA 467 Nanoflash). The laser flash technique has been generally known as a popular method to measure the in-plane thermal conductivity, which was calculated using eqn (1):

K = αC_pρ (1)

where ρ is the density of the LCER composites, C_p is the specific heat capacity, and α is the in-plane thermal diffusivity. The electric resistance was obtained using a Keithley 6517B Electrometer. All the measurements were carried out at room temperature.

Results and discussion

The fabrication process of the BNNSs/AgNPs involves the exfoliation of BNNSs and reduction of Ag⁺. During the reaction, the solution was initially yellow, then brown and finally slightly metal-like (ESI, Fig. S1†). Observation of the TEM images of the BNNSs/AgNPs shows that the AgNPs were indeed attached onto the surface of the BNNSs without aggregation (Fig. 1a and b). The size distribution of the AgNPs was between 5 and 10 nm. Especially, the AgNPs were preferentially attached on the edges of the BNNSs, as shown in Fig. 1c. High-resolution TEM analysis confirms the AgNPs' multi-crystalline structure and the interplanar spacing of d = 0.23 nm (Fig. 1d).

Fig. 1 TEM images of the BNNSs/AgNPs hybrids. (a)–(c) Low-resolution TEM images and (d) high-resolution TEM images of the multi-crystalline structural Ag nanoparticles.

To explore the mechanism of the AgNPs decoration on the surfaces of the BNNSs, an investigation on the interaction between the Ag atoms and B, N atoms was performed using XPS, as presented in Fig. 2. The elemental scan of the Ag 3d5/2 and Ag 3d3/2 core level peaks in the BNNSs/AgNPs hybrids were centered at 367.6 eV and 373.6 eV, respectively (Fig. 2a). This can be attributed to the valence state of the Ag⁰ species, suggesting that the Ag⁺ ions have been reduced by DMF. When compared with the values for the AgNPs (368.5 eV and 374.4 eV), both peaks shift to lower binding energies by 0.9 eV and 0.8 eV, suggesting that the AgNPs were attached mainly on the surface of the BNNSs by either physical adsorption or electrostatic adsorption, which is consistent with a previous report.²⁶ The main peaks of the N 1s and B 1s spectra (Fig. 2b and c) of the BNNSs/AgNPs are at 397.6 eV and 190.2 eV, a little lower than the values of N 1s and B 1s in BNNSs (Fig. 2e and f). The atomic percentage of the different elements in BNNSs/AgNPs was detected by XPS (ESI, Table S1†). The Ag element can be detected and the mass amount of AgNPs in the BNNSs/AgNPs hybrids was about 8.98 wt%. Fig. 2g shows the XRD patterns of the BNNSs and BNNSs/AgNPs. The BNNSs reveal reflections corresponding to the (002), (100), (004) and (110) crystalline planes.^27,28 When compared with the BNNSs, the peaks at the (111), (220), and (311) positions can be observed in the patterns of the BNNSs/AgNPs materials, which are in agreement with the previous reports on AgNPs.^27,29 The mechanism for successful AgNPs decoration on the BNNSs surfaces is schematically shown in Fig. 2i. Hexagonal BN is a layered material similar to graphite. Boron and nitrogen atoms are bound by strong covalent bonds and the layers are held together by weak van der Waals forces.¹⁷ During vigorous sonication of h-BN in DMF, defect sites, such as vacancy defects, topological defects, and exposed edges, are introduced on the surfaces of the BNNSs.³⁰ The reactivity of the defects on h-BN has been theoretically discussed in previous reports.^31–33 Similar to the functionalization of boron nitride nanotubes,^34,35 the electron-deficient B defects are more prone to be attacked by oxygen atoms and nitrogen atoms due to Lewis acid–base complexation reactions.³⁶ Thus, the exposed edges and vacancy defects have functional groups such as hydroxyl groups (–OH) and amino groups (–NH₂).^25,30,37 The FTIR spectrum (ESI, Fig. S3†) of the BNNSs shows two different peaks at 815 cm⁻¹ and 1402 cm⁻¹, which can be assigned to the B–N bending mode and B–N stretching mode of BNNSs.³⁸ The typical B–O signal (e.g., B–O stretching at 1350 cm⁻¹)³⁹ and the strong B–N band overlapped. The presence of amino groups was confirmed by the symmetric NH₂ stretching vibrations in the NH₂ functional groups at 3417 cm⁻¹ in the spectrum.^37,40 The absorption features around 3700 cm⁻¹ come from the free OH stretching vibrations of the B–OH functional groups.^41,42 In the system containing silver salt (AgNO₃), Ag⁺ can attach to the functional groups of the BNNSs through coordination interactions.⁴³ In a solution of DMF, the attached Ag⁺ ions are reduced to Ag atoms with a gradual increase in the Ag atom concentration. Then, nucleation is triggered and subsequent nanocrystal growth into the final AgNPs.⁴⁴ Similar metal nucleation has also occurred in other reports.⁴⁵

Fig. 2 (a–c) The typical XPS spectra of the Ag 3d5/2 and Ag 3d3/2 (a), N 1s (b), B 1s (c) core levels in the BNNSs/AgNPs hybrids. The spectra curves were deconvoluted by Gaussian fitting in (a) showing the main multi-bonding information. (d–f) The typical XPS spectra of the Ag 3d5/2 and Ag 3d3/2 (d), N 1s (e), B 1s (f) core levels in the AgNPs and BNNSs. (g) The X-ray diffraction pattern of BNNSs and the BNNSs/AgNPs. (h) Schematic of synthesis of the BNNSs/AgNPs nanohybrids via liquid-phase exfoliation of h-BN, followed by the reduction of Ag⁺.

The in-plane K values of the LCER composites filled with three different fillers (AgNPs, BNNSs and BNNSs/AgNPs) are shown in Fig. 3a. Table S2 (ESI†) shows the measured ρ, C_p and α values. The K value of pure LCER is about 0.34 W m⁻¹ K⁻¹ at room temperature. After the addition of the inorganic fillers (BNNSs and BNNSs/AgNPs), both composites show an improvement of the K value when compared to pure LCER. In the absence of the AgNPs, the highest K value for the LCER composites reaches 7.81 W m⁻¹ K⁻¹ with 25.1 vol% of BNNSs, which is in agreement with Zeng et al.'s report.⁴⁶ After AgNPs decoration on the surface of the BNNSs, the highest K value for the BNNSs/AgNPs/LCER composites can reach up to 12.55 W m⁻¹ K⁻¹ at the same loading, which was 60.7% higher than that of the LCER composite filled with BNNSs. For comparison, pure AgNPs were mixed with LCER followed by curing and heating in vacuum with a 8.98 wt% mass ratio of Ag to BNNSs. There is almost no thermal conductivity enhancement in the AgNPs/LCER composites due to the small loading fractions of AgNPs. It should be noted that the enhancement of the K value in the LCER composites was non-linear: at a higher loading of BNNSs, a more effective enhancement was observed. This indicates that efficient thermal transport pathways begin to form at a high loading of BNNSs in the BNNSs/AgNPs/LCER composites due to the sintering of AgNP–AgNP contacts during the high-temperature curing process of LCER. To shed light on the BNNSs/AgNPs arrangement in the LCER matrix, the SEM cross-section image of the BNNSs/AgNPs/LCER with 25.1 vol% BNNSs is shown in Fig. 3b. The image shows that BNNSs are stacked into a network structure with AgNPs bridges. The thermally conductive network can be formed during the curing process at relatively high temperature (220 °C). To confirm this hypothesis, the BNNSs/AgNPs hybrids were heated individually at 220 °C for 2 h. The temperature and time of heating was similar to those used to cure the composites. The obtained hybrids were characterized by TEM. As shown in Fig. 3c and d, the size of the AgNPs after heating was larger than that of the AgNPs before heating. This was due to the fact that the dispersed AgNPs in the close vicinity are sintered together to decrease the surface energy. As a result, the adjacent BNNSs are connected by linking bridges of AgNPs, as illustrated in Fig. 3e.

Fig. 3 (a) The thermal conductivity of the AgNPs/LCER, BNNSs/LCER and BNNSs/AgNPs/LCER composites as a function of AgNPs loading fraction on the accurate ratio of Ag to BNNSs and BNNSs volume loading fraction, (b) the SEM fracture morphology of the BNNSs/AgNPs/LCER composites with 25.1 vol% BNNSs content, (c) and (d) TEM images of the BNNSs/AgNPs heat-treated at 220 °C for 2 h and (e) a schematic of how heat is transmitted between the BNNSs via the bridge of AgNPs.

In general, when compared with the theoretical predictions, the major limitation to the efficiency of BNNSs and their under performance during thermal conduction, originates from the high thermal boundary resistance (R_B) between the BNNSs and matrix material.^47,48 The physics that control heat flow at the interface between the fillers and matrix are critical for the thermal transport in the next-generation nanocomposite materials and devices. A lot of theoretical and computational research has been developed in the past few years to calculate the R_B value between the filler and polymer matrix.^49–52 However, a complete theoretical analysis of the thermal transport behavior of these BNNSs media is still missing. Herein, an effective medium approximation (EMA) for calculating the K value in which the interfacial effect and the filler size are considered was used to analyze our experimental data.⁵³ The validity of the approximation has been known to be reasonable for composites with low filler loading.⁵² In our study, the effect of these parameters, including the size of the fillers, aspect ratios α and R_B value of the composites would be emphasized. To see how the R_B value may affect the effective conductivities of the composites reinforced with BNNSs/AgNPs, we investigated the K value as a function of BNNSs volume loading fraction V_f with different assumed R_B values from theoretical point of view. According to the EMA, the K value of the composites was described by the following function:⁵²

where K_p and K_m are thermal conductivity of the filler and matrix materials, respectively. h is the thickness of the BNNSs, assumed as 3.5 nm according to the TEM characterization (ESI, Fig. S2†). For simplicity, we assume the in-plane thermal conductivity of BN (K_p1) was 200 W m⁻¹ K⁻¹. We take the thermal conductivity of the BNNSs/AgNPs (K_p2) as 241 W m⁻¹ K⁻¹, supported by the theoretical estimate (see details in the ESI†). Fig. 4a shows a comparison between the calculated K values and experimental data versus V_f for the BNNSs/AgNPs/LCER and BNNSs/LCER composites. The good fit of this model to the experimental data at lower loadings (V_f < 11.2 vol%) obtains two R_B values of 3.84 × 10⁻⁷ m² K W⁻¹ for BNNSs/AgNPs/LCER and 3.50 × 10⁻⁷ m² K W⁻¹ for BNNSs/LCER, which are in agreement with other reports and theoretical results.^52,54,55 The slight difference between the two R_B values indicates the negligible effect of the AgNPs on the heat conduction and thermal coupling to the LCER at low filler loading. Furthermore, it was observed that with the assumed R_B values, the K values do not increase with R_B linearly but start to saturate.

Fig. 4 Fitting of the theoretical curves to the experimental data used to extract the actual values of the thermal boundary resistance between the BNNSs and LCER. (a) The measured (spherical dots) and predicted (colored lines) in-plane K of the LCER composites at low filler content. (b) The measured (spherical dots) and predicted in-plane K from Foygel's (colored lines) of the LCER composites at high filler content.

The simple EMA theory for BNNSs and BNNSs/AgNPs is valid for the composites in which fillers are surrounded by the matrix. For our type of composites at higher V_f values, the EMA would underestimate the effective K values of the BNNSs/LCER and BNNSs/AgNPs/LCER composites. Therefore, an additional model presented by Foygel et al. was applied to our K data of the resulting composites at high V_f values.⁵⁶ The Foygel model is based on the morphology of the filler entanglement, the heat dissipation among fillers and developing free parameters (the K value of the individual fillers networks and thermal conductivity exponent of the percolation model) to compute the contact resistance (R₀) among fillers and R_B value between the fillers and the matrix. The K values of the LCER composites near the percolation threshold are represented by eqn (4).

K = K₀[V_f − V_c(β)]^t(β) (4)

In eqn (4), the K value is the thermal conductivity of the LCER composites, β is the aspect ratio of BNNSs (β = L/H, H and L are the thickness and size of the BNNSs), K₀ is a pre-exponential factor that depends on the conductivity of the individual filler networks, V_c is the critical volume fraction of fillers, and t(β) is a thermal conductivity exponent that is dependent on the aspect ratio of the fillers. Eqn (4) can be used to find the K₀ and t(β) values based on the best match with the experimental data. According to eqn (5) and the TEM images of the BNNSs (ESI, Fig. S2†), we estimated the critical fractional volume V_c = 0.01 with H = 3.5 nm and L = 200 nm. Then, we used eqn (4) to fit the K data of the BNNSs/AgNPs/LCER composites with one free parameter K₀ = 88 W m⁻¹ K⁻¹ (using t(β) = 1.3 as an additional free parameter). For BNNSs/LCER, the K₀ and t(β) values can be estimated as K₀ = 67 W m⁻¹ K⁻¹ and t(β) = 1.4 on the basis of the best fit with the experimental data, as shown in Fig. 4b. The R₀ value can be defined as

R₀ = (K₀LV_c^t(β))⁻¹ (6)

using K₀ and t(β), the obtained values for R₀ are 2.26 × 10⁷ K W⁻¹ for the BNNSs/AgNPs/LCER and 4.71 × 10⁷ K W⁻¹ for the BNNSs/LCER. The R_B values obtained from this data analysis can be extracted by assuming that 1/10 surface of each BNNS was involved in heat dissipation of the filler network. Then, the surface of BNNS could be estimated as 4.0 × 10⁻¹⁶ m² (L = 200 nm), from which we could obtain the R_B values of 9.04 × 10⁻⁸ m² K W⁻¹ for BNNSs/AgNPs/LCER and 1.88 × 10⁻⁷ m² K W⁻¹ for BNNSs/LCER. It was noted that in Foygel's model, the best matched R_B value with the experimental data for the BNNSs/AgNPs/LCER composites, was smaller than that found for the BNNSs/LCER with high BNNSs content. This data indicates that an effective interfacial thermal transfer was indeed achieved in the BNNSs/AgNPs/LCER. The interfacial surface conditions of BNNSs can be tuned by the decoration of AgNPs at the vacancy defects or edges of the BNNSs. The decrease in the R_B value after the attachment of the AgNPs may be associated with the increased affinity between the BNNSs/AgNPs and LCER. It was possible that the AgNPs are acting as an interstitial material used to bridge the different phonon spectra between the BNNSs and LCER. Our experimental K values suggest that the rate of thermal transport across the BNNSs/AgNPs–LCER interface would be increased substantially by the AgNPs bridges. The phonon scattering of BNNSs/LCER may be higher than that of BNNSs/AgNPs/LCER, especially on the filler–matrix interface.

Thus, the results suggest that the thermal boundary resistances extracted by the two approaches can be almost the same order of magnitude (0.9–4 × 10⁻⁷ m² K W⁻¹), which is almost in agreement with that of randomly dispersed CNTs in epoxy resin (10⁻⁸ to 10⁻⁷ m² K W⁻¹). The EMA model shows the best fit with the experimental data at lower loadings (V_f < 11.2%) and Foygel's model was best applied to higher loadings (V_f < 25.1% in this study). The theoretical predictions and experimental data discussed above have demonstrated that the BNNSs/AgNPs are an effective thermally conductive filler.

BNNSs are electrically insulating with a wide band gap of 5.6 eV.⁵⁷ We expect that the intrinsic electrical insulating properties of the BNNSs could be retained in the BNNSs/AgNPs/LCER. A two-point probe method, one of the most common ways to measure a material's surface resistivity (δ), was used in a square pattern in contact with the surface of the test material.⁵⁸ According to Ohm's law, the δ of a sample with rectangular shape can be defined as

where R is the electrical resistance of the material, d and l are the width and length of the sample, respectively.⁵⁹ Eqn (7) allows us to calculate the δ value to be of the order of 10¹² Ω, meaning that the decoration of AgNPs on the surface of the BNNSs was not accompanied with decreasing δ. As shown in Fig. 5, incorporating BNNSs/AgNPs into the LCER matrix slightly increases the δ value of neat LCER (δ₀ = 2.1 TΩ). In the LCER composite, the conduction of electrons among the AgNPs was completely blocked by the BNNSs and LCER. Therefore, the electrically insulating properties of the composites was controlled well, being important for the integration of these materials in a wide range of packing industrial applications.

Fig. 5 Electrical resistivity measurements of pure LCER, BNNSs/LCER and BNNSs/AgNPs/LCER composites at different BNNSs content.

To show the thermal management application of our BNNSs/AgNPs/LCER composites, we use the prepared composite film with 25.1 vol% BNNSs content as a substrate to fix two LED chips (Fig. 6a). As a control, the same procedure was applied to the pure LCER film (Fig. 6c). The two films were tested using a thermal infrared camera. After steady-state (5 minutes) of the same heat flow caused by the equal standard LED chips, the heat distribution was uniform for the BNNSs/AgNPs/LCER film with a low center spot temperature (32 °C, Fig. 6b), which can positively influence the lifetime of devices. However, the heat flow generated from the LED chips was highly concentrated (Fig. 6c) with high temperature gradients from the center to the edge of the film (Fig. 6d). The highest center spot temperature for the pure LCER film increased to 47 °C, which will finally degrade the performance, reliability and lifetime of the LED chips.

Fig. 6 Application of the BNNS based LCER composite film in substrates. Operation of a LED chip on the BNNSs/AgNPs/LCER film with 25.1 vol% BNNSs (a) and pure LCER film (c). The corresponding thermal images of BNNSs/AgNPs/LCER film (b) and pure LCER film (d) recorded using an infrared camera at the same steady-state.

Conclusions

We have successfully fabricated high-performance LCER composites based on BNNSs/AgNPs hybrid fillers with a maximum in-plane thermal conductivity value reaching 12.55 W m⁻¹ K⁻¹ at 25.1 vol% BNNSs content. We have demonstrated that sintered AgNPs among the BNNSs can form efficient heat transferring bridges, which are the main contributor of the significant thermal enhancement. Fitting the thermal conductivity data of the LCER composites with EMA and Foygel models indicates that the thermal boundary resistance (R_B) at the fillers–matrix interfaces decreased with AgNPs decoration at a high BNNS content. In addition, the prepared LCER composites maintain high electrical insulating properties. The resulting composite film shows its potential applications in portable devices as a substrate to support light-emitting-diode (LED) chips and demonstrated that it was more effective in heat removal. This study offers a new insight into the strategies used towards the enhancement of thermal conductivity in polymer-based composites for the widely identified and pursued potential applications in modern electronics.

Acknowledgements

This study was supported by Guangdong and Shenzhen Innovative Research Team Program (No. 2011D052 and KYPT20121228160843692).

Notes and references

1. E.-S. Lee, S.-M. Lee, D. J. Shanefield and W. R. Cannon, J. Am. Ceram. Soc., 2008, 91, 1169–1174.
2. C. Y. Zhi, Y. Bando, C. C. Tang, Q. Huang and D. Golberg, J. Mater. Chem., 2008, 18, 3900–3908.
3. Y. Xu and D. D. L. Chung, Compos. Interfaces, 2000, 7, 243–256.
4. A. S. Luyt, J. A. Molefi and H. Krump, Polym. Degrad. Stab., 2006, 91, 1629–1636.
5. J. Yang, Y. Yang, S. W. Waltermire, X. Wu, H. Zhang, T. Gutu, Y. Jiang, Y. Chen, A. A. Zinn, R. Prasher, T. T. Xu and D. Li, Nat. Nanotechnol., 2012, 7, 91–95.
6. M. D. Losego, M. E. Grady, N. R. Sottos, D. G. Cahill and P. V. Braun, Nat. Mater., 2012, 11, 502–506.
7. Q. Liang, X. Yao, W. Wang, Y. Liu and C. P. Wong, ACS Nano, 2011, 5, 2392–2401.
8. M. Donnay, S. Tzavalas and E. Logakis, Compos. Sci. Technol., 2015, 110, 152–158.
9. C. Zhi, Y. Bando, T. Terao, C. Tang, H. Kuwahara and D. Golberg, Adv. Funct. Mater., 2009, 19, 1857–1862.
10. K. Kim, M. Kim and J. Kim, Compos. Sci. Technol., 2014, 103, 72–77.
11. V. Singh, T. L. Bougher, A. Weathers, Y. Cai, K. Bi, M. T. Pettes, S. A. McMenamin, W. Lv, D. P. Resler, T. R. Gattuso, D. H. Altman, K. H. Sandhage, L. Shi, A. Henry and B. A. Cola, Nat. Nanotechnol., 2014, 9, 384–390.
12. Y. Xu, D. D. L. Chung and C. Mroz, J. Electron. Packag., 2001, 32, 1749–1757.
13. B.-H. Xie, X. Huang and G.-J. Zhang, Compos. Sci. Technol., 2013, 85, 98–103.
14. H. He, R. Fu, Y. Han, Y. Shen and D. Wang, J. Electron. Packag., 2007, 129, 469–472.
15. Y. Yao, X. Zeng, K. Guo, R. Sun and J.-b. Xu, Composites, Part A, 2015, 69, 49–55.
16. I. Jo, M. T. Pettes, J. Kim, K. Watanabe, T. Taniguchi, Z. Yao and L. Shi, Nano Lett., 2013, 13, 550–554.
17. L. Song, L. Ci, H. Lu, P. B. Sorokin, C. Jin, J. Ni, A. G. Kvashnin, D. G. Kvashnin, J. Lou, B. I. Yakobson and P. M. Ajayan, Nano Lett., 2010, 10, 3209–3215.
18. M. Xu, D. Fujita, H. Chen and N. Hanagata, Nanoscale, 2011, 3, 2854–2858.
19. L. H. Li, J. Cervenka, K. Watanabe, T. Taniguchi and Y. Chen, ACS Nano, 2014, 8, 1457–1462.
20. Z. Kuang, Y. Chen, Y. Lu, L. Liu, S. Hu, S. Wen, Y. Mao and L. Zhang, Small, 2015, 11, 1655–1659.
21. D. Lee, S. H. Song, J. Hwang, S. H. Jin, K. H. Park, B. H. Kim, S. H. Hong and S. Jeon, Small, 2013, 9, 2602–2610.
22. K. Wattanakul, H. Manuspiya and N. Yanumet, Colloids Surf., A, 2010, 369, 203–210.
23. X. Huang, S. Wang, M. Zhu, K. Yang, P. Jiang, Y. Bando, D. Golberg and C. Zhi, Nanotechnology, 2015, 26, 015705.
24. H. Shen, J. Guo, H. Wang, N. Zhao and J. Xu, ACS Appl. Mater. Interfaces, 2015, 7, 5701–5708.
25. C. Zhi, Y. Bando, C. Tang, H. Kuwahara and D. Golberg, Adv. Mater. Processes, 2009, 21, 2889–2893.
26. E. Foroozmehr, H. Alemohammad, X. Wang, E. Toyserkani and S. Esmaeili, J. Phys. D: Appl. Phys., 2011, 44, 495305.
27. H. Shen, C. Duan, J. Guo, N. Zhao and J. Xu, J. Mater. Chem. A, 2015, 3, 16663–16669.
28. L. Huang, P. Zhu, G. Li, F. Zhou, D. Lu, R. Sun and C. Wong, J. Mater. Sci.: Mater. Electron., 2015, 26, 3564–3572.
29. R. Zhang, K.-s. Moon, W. Lin and C. P. Wong, J. Mater. Chem., 2010, 20, 2018–2023.
30. Y. Lin, T. V. Williams, W. Cao, H. E. Elsayed-Ali and J. W. Connell, J. Phys. Chem. C, 2010, 114, 17434–17439.
31. Y.-H. Zhang, K.-G. Zhou, X.-C. Gou, K.-F. Xie, H.-L. Zhang and Y. Peng, Chem. Phys. Lett., 2010, 484, 266–270.
32. J. Ma, D. Alfè, A. Michaelides and E. Wang, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80, 033407.
33. W. Chen, Y. Li, G. Yu, Z. Zhou and Z. Chen, J. Chem. Theory Comput., 2009, 5, 3088–3095.
34. S. Y. Xie, W. Wang, K. A. Fernando, X. Wang, Y. Lin and Y. P. Sun, Chem. Commun., 2005, 3670–3672.
35. Z. Cui, A. J. Oyer, A. J. Glover, H. C. Schniepp and D. H. Adamson, Small, 2014, 10, 2352–2355.
36. S. Pal, S. R. C. Vivekchand, A. Govindaraj and C. N. R. Rao, J. Mater. Chem., 2007, 17, 450–452.
37. K. Sato, H. Horibe, T. Shirai, Y. Hotta, H. Nakano, H. Nagai, K. Mitsuishi and K. Watari, J. Mater. Chem., 2010, 20, 2749–2752.
38. Y. Lin, T. V. Williams, T.-B. Xu, W. Cao, H. E. Elsayed-Ali and J. W. Connell, J. Phys. Chem. C, 2011, 115, 2679–2685.
39. Y. Liu, J. Gao and R. Zhang, Polym. Degrad. Stab., 2002, 77, 495–501.
40. R. T. Paine and C. K. Narula, Chem. Rev., 1990, 90, 73–91.
41. J. D. Ferguson, A. W. Weimer and S. M. George, Thin Solid Films, 2000, 371, 95–104.
42. M. T. Huang and H. Ishida, Surf. Interface Anal., 2005, 37, 621–627.
43. P. C. Ma, B. Z. Tang and J.-K. Kim, Carbon, 2008, 46, 1497–1505.
44. L. He and S. C. Tjong, RSC Adv., 2015, 5, 15070–15076.
45. K. Jasuja and V. Berry, ACS Nano, 2009, 3, 2358–2366.
46. X. Zeng, L. Ye, S. Yu, H. Li, R. Sun, J. Xu and C. P. Wong, Nanoscale, 2015, 7, 6774–6781.
47. M. B. Bryning, D. E. Milkie, M. F. Islam, J. M. Kikkawa and A. G. Yodh, Appl. Phys. Lett., 2005, 87, 161909.
48. J. Su, Y. Xiao and M. Ren, Phys. Status Solidi A, 2013, 210, 2699–2705.
49. C.-W. Nan, G. Liu, Y. Lin and M. Li, Appl. Phys. Lett., 2004, 85, 3549–3551.
50. S. H. Xie, Y. Y. Liu and J. Y. Li, Appl. Phys. Lett., 2008, 92, 243121.
51. P. Bonnet, D. Sireude, B. Garnier and O. Chauvet, Appl. Phys. Lett., 2007, 91, 201910.
52. K. M. Shahil and A. A. Balandin, Nano Lett., 2012, 12, 861–867.
53. C.-W. Nan, R. Birringer, D. R. Clarke and H. Gleiter, J. Appl. Phys., 1997, 81, 6692.
54. Z. Lin, A. McNamara, Y. Liu, K.-s. Moon and C.-P. Wong, Compos. Sci. Technol., 2014, 90, 123–128.
55. R. J. Warzoha and A. S. Fleischer, Nano Energy, 2014, 6, 137–158.
56. M. Foygel, R. D. Morris, D. Anez, S. French and V. L. Sobolev, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 104201.
57. W. Meng, Y. Huang, Y. Fu, Z. Wang and C. Zhi, J. Mater. Chem. C, 2014, 2, 10049–10061.
58. T. C. Li, C. S. Chang, W. L. Liang, W. F. Tsai, C. F. Ai and J. F. Lin, J. Instrum., 2012, 7, P07013.
59. P. P. Tsai and R. R. Bresee, J. Appl. Polym. Sci., 2001, 82, 2856–2862.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra00358c

---