Enhanced bending strength and thermal conductivity in diamond/Al composites with B4C coating

Diamond/Al composites containing B4C-coated and uncoated diamond particles were prepared by powder metallurgy. The microstructure, bending strength and thermal conductivity were characterized considering the B4C addition and diamond fraction. The influence of B4C coating and fraction of diamond on both bending strength and thermal conductivity were investigated. The bending strength increased with decreasing diamond fraction. Moreover, addition of B4C coating led to an obvious increase in bending strength. The peak value at 261.2 MPa was achieved in the composite with 30 vt.% B4C-coated diamond particles, which was about twice of that for 30 vt.% uncoated diamond/Al composite (140.1 MPa). The thermal conductivity enhanced with the increase in diamond fraction, and the highest value (352.7 W/m·K) was obtained in the composite with 50 vt.% B4C-coated diamond particles. Plating B4C on diamond gave rise to the enhancement in bending strength and thermal conductivity for diamond/Al composites, because of the improvement of the interfacial bonding between diamond and aluminum matrix.

then fabricated 7075 Al/B 4 C surface composite through friction stir processing (FSP). A 100% improvement in wear resistance was achieved compared to the parent metal. These results indicated that Al/B 4 C exhibited a strong interface. Meanwhile, B 4 C bond with diamond via B-C covalent bond 26 . Therefore, pre-plating B 4 C coating on diamond was a potential effective method for increasing both strength and TC for diamond/Al composite.
In this work, we describe the great enhancement in bending strength and TC for aluminum composites reinforced with B 4 C-coated diamond particles. The diamond/Al composites with different volume fraction of B 4 C-coated and uncoated diamond particles were prepared by powder metallurgy. The B 4 C coating on diamond is conducive to obtain a dense diamond/Al composite due to the continual bonding interface. The enhancements of both mechanical and thermal properties are dependent on the porosity and the interfacial gap width between diamond and aluminum matrix.

Results
The scanning electron microscopy (SEM) images for the B 4 C-coated diamond particles are shown in Fig. 1. For the coated diamond, a complete coverage of coating is achieved on all faces. The B 4 C layer is quite homogeneous, and the magnified image (Fig. 1b) indicates that the layer consists of submicron particles. The X-ray diffractogram (XRD) spectrum for the B 4 C-coated diamond particles is displayed in Fig. 2. As shown in Fig. 2, a high-intensity peak located at 43.9° is attributed to diamond (JCPDF#06-0675), which is partially truncated for clearly observing the other peaks for the coating on the diamond particle. Compared to the standard JCPDF#35-0798, the diffraction peaks located at 21.9°, 23.3°, 31.8°, 34.7°, 37.6° and 49.9° are ascribed to the typical B 4 C structure. The grain sizes of the B 4 C coating calculated from Williamson-Hall plot is 64 nm.
SEM micrographs of the uncoated-diamond/Al (D1-3) and coated-diamond/Al (B1-3) composites are shown in Fig. 3. It can be seen that a quite homogeneous distribution of diamond particles is achieved and most corners of diamond particles remain. For the uncoated-diamond/Al composites (D1-3, Fig. 3a,c,e), wide gap between diamond and Al matrix is observed. Meanwhile, un-wetting phenomenon between diamond particles and Al matrix is existent on the surface of uncoated-diamond/Al composites. The wide gap and un-wetting phenomenon between diamond and Al are more significant with the increase in diamond fraction. In contrast, for the coated-diamond/Al composites (B1-3, Fig. 3b,d,f), both amount and width of gap between diamond particles and Al matrix exhibit an obvious decrease. Moreover, even for the composite with 50 vt.% B 4 C-coated diamond (B3, Fig. 3f), the un-wetting phenomenon is never observed. It indicates that the B 4 C coating gives rise to a good adhesion between diamond and Al matrix, which causes the densification of diamond/Al composites.  The distribution of the composite elements (Al and C) in B 4 C-coated-diamond/Al composites is analyzed by EDS line-scanning and the results are shown in Fig. 4. Three distinct zones (Al zone, B 4 C layer and diamond zone) are presented in the figure. The thickness of the interfacial B 4 C layer is about 1.5 μm. When entering the interfacial B 4 C layer from the Al zone, a sharp decrease in intensity of the Al signal as well as an increase in intensity of C signal are observed. When entering the diamond zone from the interfacial B 4 C layer, a further increase in intensity of the C signal is observed, indicating the difference in the content of carbon between B 4 C and diamond. Figure 5 presents the XRD patterns for the diamond/Al composites with 50 vt.% uncoated and B 4 C-coated diamond particles (D3 and B3). Apart from the four typical aluminum peaks at 38.5°, 44.7°, 65.1° and 78.2° (JCPDF#65-2869), the diamond peak is observed at 43.9° (JCPDF#06-0675). For the 50 vt.% B 4 C-coated-diamond/Al composites (B3), typical B 4 C structure is confirmed by the peaks at 22.0°, 23.5°, 34.9° and 37.8° (JCPDF#35-0798).
The theoretical density of the composte is determined by.
where ρ D (3.52 g/cm 3 27 ) and ρ m (2.70 g/cm 3 27 ) is theoretical density of diamond and aluminum matrix, respectively. V D and V M are the volume fraction of diamond particles and aluminum matrix respectively. Thereby, the theoretical density for composite with 30, 40 and 50 vt.% diamond is 2.95, 3.03 and 3.11 g/cm 3 , respectively. With increasing diamond from 30 to 50 vt.%, the density of uncoated-diamond/Al composites is 2.86, 2.92 and 2.96 g/ cm 3 (D1-3), which is quite lower than the corresponding theoretical density. In contrast, plating B 4 C coating on diamond gives rise to dense composites. The composite with 30 vt.% B 4 C-coated diamond (B1) exhibits a density of 2.92 g/cm 3 , which is quite close to the theoretical density. The composite with 50 vt.% B 4 C-coated diamond particles (B3) still maintains a high density (3.05 g/cm 3 ), which is denser than all uncoated-diamond/Al composites (D1-3). The relative density ρ relative is determined by.
where ρ measured is the measured density. The relative densities for Al/diamond composites with different diamond fractions of uncoated and B 4 C-coated diamond particles are shown in Fig. 6. As shown in Fig. 6, the relative densities for composites with both uncoated and B 4 C coated diamond particles decrease with the increase in diamond fraction, whereas the relative densities for composites with B 4 C coated diamond particles (>97.9%) are higher than those for composites with uncoated diamonds (up to 97.2%). Furthermore, by comparing relative density for the B 4 C-coated-diamond/Al with that for the corresponding uncoated-diamond/Al composite, the improvement in relative density increases from 1.8 to 2.6% with increasing diamonds fraction from 30 to 50 vt.%, meaning that plating B 4 C coating on diamond is benefit to the densification of diamond/Al composite.   SEM micrographs for bending fracture surfaces of the diamond/Al composites with uncoated (D1-3) and B 4 C-coated (B1-3) diamond particles are given in Fig. 8. For the uncoated-diamond/Al composites (D1-3, Fig. 8a,c,e), large amounts of wide gap around diamond particles are observed, which is marked by white arrows. Meanwhile, un-wetting phenomenon between aluminum matrix and diamond particles is existent, which is highlighted by brown arrows. For the B 4 C-coated-diamond/Al composites (B1-3, Fig. 8b,d,f), no obvious gap around diamond particles is found. EDS mapping was used to evaluate the bonding condition between diamond particles and aluminum matrix. Figure 9 presents the elements distribution maps of diamond/Al composites containing 50 vt.% uncoated and B 4 C-coated diamond particles. As shown in Fig. 9a,b, obvious gap at the interface boundary is observed for the uncoated-diamond/Al composite (D3). Meanwhile, most surfaces of the diamond particle remain naked without aluminum matrix surrounded, indicating a poor interfacial bonding caused by the un-wetting between diamond and aluminum matrix. It implies the fracture of composite D3 occurred from interface between diamond and Al matrix, because of the poor interfacial bonding. According to the elements distribution map of sample B3 (Fig. 9c,d), the adhesion of the aluminum matrix to the diamond particle is remarkably improved. Most diamond surfaces are found to be embedded in the aluminum matrix, while the exposed part is covered by large amounts of aluminum dimples. It indicates that ductile fracture of Al matrix has replaced interfacial fracture to be the dominant fracture mode in composite B3, since better wettability and interfacial bonding between diamond particles and aluminum achieved by plating B 4 C. Figure 10 shows the TC of diamond/Al composites with uncoated (D1-3) and B 4 C-coated (B1-3) diamond particles. As shown in Fig. 10, uncoated-diamond/Al composites (D1-3) exhibit lower TC, and the TC for sample D1 (208.4 W/m·K) is even lower than that for pure aluminum (237 W/m·K 28 ). With increasing diamond fraction from 30 to 50 vt.%, the TC for uncoated-diamond/Al composite increases from 208.4 to 283.8 W/m·K, meanwhile that for B 4 C-coated-diamond/Al composite increases from 311.4 to 352.7 W/m·K. Moreover, it should be noted that B 4 C coating gives rise to an obvious increase in thermal conductivity, because the TC for B 4 C-coated diamond/Al composites are much higher than those for uncoated diamond/Al composites. The TC for composite containing 30 vt.% B 4 C-coated diamond particles (B1) even performs better than that for the composite with 50 vt.% uncoated diamond particles (D1).

Discussion
As shown in Fig. 6, the relative density for uncoated diamond/Al composites (D1-3) is lower than that for B 4 C-coated diamond/Al composites (B1-3) at each diamond fraction, indicating that a larger number of pores exist in samples D1-3. Meanwhile, SEM images for bending fracture surface show that wider interfacial gap is observed between uncoated diamond and aluminum matrix. The separation between diamond and aluminum is caused during the cooling process by the large difference in expansion coefficients between aluminum (23.0 × 10 −6 K −1 29 ) and carbon materials (1.0 × 10 −6 K −1 27 ). Therefore, the low density for uncoated diamond/ Al composites is attributed to large amounts of wide gap around diamond particles. Moreover, the decrease in relative density for uncoated diamond/Al composite with the increase in diamond fraction is caused by the larger amount of wide gap provided by longer interface. Assuming that the diamond used in this study is isotropic spherical particle, the average gap width (x) can be evaluated by where a is the radius of diamond particle, V D is the volume fraction of diamond particle, and V P is porosity (Fig. 11a). As displayed in Fig. 11b, plating B 4 C coating on diamond reinforcement gives rise to a decrease in average gap width from 1.54 to 0.58 μm, which agrees well with SEM results (Fig. 8). B 4 C interlayer is benefited to relieve the interfacial thermal stress between aluminum and diamond during the cooling process due to the modest expansion coefficients of B 4 C (5.65 × 10 −6 K −1 30 ). In addition, the improvement rate attributed to B 4 C interface is more significant for the composite with larger diamond fraction. Therefore, we suggest that the formation of B 4 C is benefited to the densification of diamond/Al composite. Figure 12 summarizes the relationship between bending strength and porosity. It can be seen that the bending strength is strongly dependent on the porosity of diamond/Al composites. Together with the average gap width (Fig. 11) and SEM images (Fig. 8), uncoated-diamond/Al composites (D1-3) are fabricated with high porosity along with a large number of gap and pores, which contributes to the extension of the crack under stress. Therefore, the significant decrease in bending strength for uncoated diamond/Al composites with the increase in diamond fraction (as shown in Fig. 7) is due to that the larger diamond fraction results in a longer weak interface and larger number of gap and pores. For the B 4 C-coated diamond/Al composites (B1-3), plating B 4 C on diamond contributes to improve the wettability between diamond and aluminum matrix and optimize the interface structure. B 4 C coating is benefited to the increase in bending strength due to the decrease in average gap width and porosity. For the 30 vt.% B 4 C-coated-diamond/Al composite (B1), no obvious gap between diamond and aluminum matrix is observed and the highest bending strength is achieved, because the continual interface is  conducive to distribution of stress. Thereby the enhancement is more evident for the composite with larger diamond fraction, since longer weak interface between naked diamond and Al matrix is enhanced by B 4 C interface. Furthermore, the 50 vt.% B 4 C-coated-diamond/Al composite (B3) still maintain a high bending strength which is even larger than the bending strength for the composite with 30 vt.% uncoated diamond particles (D1), because the number of gap and pores for B3 is lower than that for D1 via adding a B 4 C interlayer between diamond and Al matrix.
Furthermore, the EDS mapping results reveal the conversion of fracture behavior of diamond/Al composites after plating B 4 C on diamond particles. The dominant position of interfacial fracture (Fig. 9b) suggests a weak interfacial bonding between diamond and Al matrix for uncoated-diamond/Al composites, which is not benefited to the stress transfer and leads to a lower bending strength and brittle fracture. In contrast, for the composite with B 4 C-coated diamond (e.g. B3), large number of dimples are observed on the surface of diamond, which is confirmed to be Al by EDS analysis. It demonstrates that the fracture mode of B 4 C-coated-diamond/Al composite is converted to ductile fracture, and a strong interfacial bonding between diamond particles and aluminum are achieved by plating B 4 C.
As shown in Fig. 10, plating B 4 C on diamond particles contributes to a significant increase in TC for diamond/ Al composites. The TC for composites depends on many factors, such as component, reinforcement fraction and size, distribution, and interfacial bonding between the matrix and reinforcement 18 . To better understand the thermal conductivity behavior of diamond/Al composites, it is necessary to compare experimental results with theoretical predictions. Among those models developed by researchers, the Hasselman-Johnson (H-J) model 31 was used to estimate effective thermal conductivity of composites K c by taking interfacial thermal barrier into consideration.  where K m and K r are TC of matrix and reinforcement particles respectively (K Al = 237 W/m·K 32 , K diamond = 1350 W/m·K 7 ), a is the radius of reinforcement particle. The interfacial thermal conductance h c is identified as where ρ m and ρ r are theoretical density of matrix and reinforcement particle respectively, v m and v r are phonon velocity in matrix and reinforcement particle respectively (v Al = 3040 m/s, v diamond = 13924 m/s 26 ), and c m is the specific heat of matrix (C Al = 880 J/kg·K 7 ). The experimental data and theoretical prediction of the TC of diamond/Al composites with uncoated and B 4 C-coated diamond particles are displayed in Fig. 13. As shown in Fig. 13, measured TCs of B 4 C-coated-diamond/ Al composites are close to theoretical values, whereas for uncoated-diamond/Al composites measured TCs exhibit a clear difference with theoretical results. Near 100% of the theoretical value (316.1 W/m·K) is reached for the 30 vt.% B 4 C-coated-diamond/Al composite (B1, 311.4 W/m·K), while only around 80% of the theoretical value is reached for D1 with uncoated diamond particles. In addition, it is worth noted in Fig. 13 that with the diamond fraction increasing, the deviation between the measured TC and theoretical value becomes larger, which is hard to be explained through the H-J model.
Although many factors are taken in the H-J model to evaluate the TC of composites, the porosity in theses specimens is not taken into consideration. It is well known that air (or vacuum) is an excellent heat insulation layer. Furthermore, the influence of porosity on TC was investigated in diamond/Cu composites 26 . Therefore, the existence of gap in composite contributed to the deterioration in TC of composite, leading to a larger deviation between measured and theoretical results. To further understand the effect of porosity on the TC of diamond/ Al composites, The relationship between K measured /K theoretical and porosity for all composites is displayed in Fig. 14. It reveals that K measured /K theoretical is highly dependent on the porosity of diamond/Al composites. The measured TC of diamond/Al composite is closer to the theoretical prediction by the H-J model with a higher density. The relative density of diamond/Al composites (Fig. 6) suggests that the relative density decreases with increasing diamond fraction. Thereby, lower fraction of diamond in composite gives rise to an increase in K measured /K theoretical   Fig. 14) because of the decrease in porosity. In addition, combined with densification and TC analyze, the addition of B 4 C interlayer is contributed to the densification of composite, and benefit to achieve a continual interface between diamond reinforcement and Al matrix, which gives rise to an obvious enhancement in thermal conductivity. Therefore, the composites with a B 4 C interlayer display higher TCs (Fig. 10) and larger K measured /K theoretical (Fig. 14).
To summarize, diamond/Al composites with different fraction of uncoated and B 4 C-coated diamond particles were prepared by powder metallurgy. Interfacial bonding and porosity were found to be the key factors in determining the properties of diamond/Al composites. The addition of B 4 C coating was benefit to the decrease in interfacial gap between diamond and Al, which gave rise to a dense composite. In addition, both bending strength and thermal conductivity of composites were dependent on the interfacial gap width between diamond and Al matrix. The bending strength for composites with B 4 C coated diamond was about twice of that for composites with un-coated diamond. Meanwhile, the specimens with B 4 C-coated diamond particles exhibited high TC (up to 352.7 W/m·K), even the sample with 30 vt.% B 4 C-coated diamond exhibited a TC as 311.4 W/m·K, which is quite larger than that for composite with un-coated diamond (up to 283.8 W/m·K) and pure Al (237 W/m·K 28 ).

Method
As the composite matrix, the pure aluminum powder with an average size of 74 μm was provided from Shanghai Chaowei Nanotechnology Co. Ltd., China. Synthetic HPHT diamond particles (HSD90, particle size 140/170 mesh (89~104 μm), Henan huanghe whirlwind international Co., Ltd., China) with cubic-octahedral monocrystalline were used as the reinforcement. For forming boron carbide (B 4 C) coating on diamond particles, the mixture of boron (B), boric acid (H 3 BO 3 ) and diamond particles was heated in a tube furnace at 1200 °C for 6 h in Ar atmosphere. The method was described in detail in the previous work 26 . Different fraction of diamonds (30,40 and 50 vt.%) were mechanically mixed with aluminum powder at room temperature. For comparison, the uncoated and B 4 C-coated diamond particles were both applied. The powder mixtures were then subjected to vacuum hot pressing in a graphite die. The furnace was heated up to 600 °C at a heating speed of 10 °C/min, then held for 30 min under a uniaxial pressure at 30 MPa in order to ensure the density. The temperature during hot pressing was monitored through a thermocouple inserted into the graphite die. After sintering, the specimens were cooled in furnace to the room temperature. The vacuum was needed before the specimens being cooled to room temperature. The composites with different fractions of B 4 C-coated and un-coated diamond reinforcements were referred as B1-3 and D1-3 respectively, as shown in Table 1.
To investigated the microstructure of coating and the interfacial product, X-ray diffraction (XRD) was performed on a Bruker D8 with a Cu K α source in the step mode from 20° to 80° at a scanning speed of 0.08 °/s. A Hitachi S-4800 scanning electron microscope (SEM) was used to characterize the distribution of diamond particles in composites. The density of composites was measured by a high precision ceramic porosity volume density tester (Dahometer, DE-120M) using Archimedes method 29 . Three-point bending strength was measured on the specimens with a dimension of 5 × 8 × 50 mm. The bending test was carried out with an initial speed of . 0 50 mm/ min using an electronic universal test machine (DDL 100, CIMACH, Changchun, China). The morphologies of fracture surfaces were also obtained by SEM. X-ray energy dispersive spectrometer (EDS) attached to the SEM equipment was applied to analyze the elements distribution on the fracture surface of composites. Thermal diffusivity of the composites at room temperature was measured by a laser flash method by a NETZSCH LFA427/3/G thermal physical testing instrument. Specific heat of the composites was derived from the theoretical value calculated according to the rule of mixture (ROM). Finally, the thermal conductivity was calculated by the product of density, thermal diffusivity and specific heat according to the following equation 28 : where α was thermal diffusivity, ρ measured was the measured density of composites and C was specific heat. Detailed results were shown in Table 2.