RETRACTED ARTICLE: Enhanced tensile strength and thermal conductivity in copper diamond composites with B4C coating

Boron carbide (B4C) coating on diamond particle is synthesized by heating diamond particles in a powder mix of H3BO3 and B in Ar atmosphere. The composition, bond state and coverage fraction of boron carbide coating on diamond particles are investigated. The boron carbide coating favors to grow on diamond (100) surface rather than on diamond (111) surface. Cu matrix composites reinforced with B4C-coated diamond particles were made by powder metallurgy. The addition of B4C coating gave rise to a dense composite. The influence of B4C coating on both tensile strength and thermal conductivity of the composite were investigated. When the B4C fully covered on diamond particles, the composite exhibited a greatly increase in tensile strength (115 MPa) which was much higher than that for uncoated-diamond/Cu (60 MPa) composites. Meanwhile, a high thermal conductivity of 687 W/mK was achieved in the B4C-coated-diamond/Cu composites.

. Furthermore, Mansourzadeh et al. 22 have found that the B 4 C particles reinforced copper composite prepared by accumulative roll-bonding in room temperature shows an excellent tensile strength. It indicates that B 4 C easily forms a strong bonding with copper at a relatively low temperature. Therefore, pre-plating B 4 C coating on diamond is a potential effective method for obtaining a diamond/Cu composite with high tensile strength and thermal conductivity.
We describe in this paper the greatly enhancement in tensile strength and thermal conductivity for copper matrix composites reinforced with B 4 C coated diamond particles. The B 4 C coating on diamond is conductive to obtain a dense diamond/Cu composite because of the continual and strong bonding interface. And the enhancements of both mechanical and thermal properties are dependent on the interfacial gap width between diamond and copper matrix.

Results
The scanning electron microscopy (SEM) images for the uncoated (D0, Fig. 1a) and coated diamond particles (D1-3, Fig. 1b-d) are shown in Fig. 1. The coating coverage is dependent on the synthesis time. For the two-hours-coated diamond (D1), only nucleation of the coating is obtained on diamond (111) surface (triangle  The X-ray diffraction (XRD) pattern for the six-hours-coated diamond particles (D3) is displayed in Fig. 2. As shown in Fig. 2, a high-intensity peak located at 43.92° 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  Raman spectroscopy was used for obtaining more information regarding the bonding conditions for the coatings on diamond particles. The Raman spectra for coated diamond particles (D1-3) in a range between 200 and 2000 cm −1 are shown in Fig. 3, wherein the high-intensity peak located at approximately 1333 cm −1 is closed to diamond 20 . In addition, a series of bonds located at 478, 530, 720, 830, 1000 and 1085 cm −1 are assigned to vibrations of the principal structural elements, icosahedra and three-atom linear chains in the B 4 C crystal [23][24][25][26][27] , which is in good agreement with the XRD analysis. Moreover, the common amorphous structure for boron carbide and diamond system, such as amorphous boron carbide (269 and 325 cm −1 ) or amorphous carbon (1350 and 1550 cm −1 ) is not obtained in the coating 20 .
The XPS C1s spectra for boron carbide coated diamond are displayed in Fig. 4a, wherein two peaks are identified. The peak located at 285.2 eV and 282.4 eV is assigned to C-C and C-B bonds 20, 28 , respectively. The high fraction of C-C in C1s spectrum for two-hours-coated diamond (D1) is attributed to the bare plane of diamond particles. The increase in plating time gives rise to a decrease in the fraction of C-C in C1s spectrum since the growth of boron carbide on diamond (111) plane. When plating time reaches 6 hours, the peak for C-C bond is hardly apparent, because diamond particles are completely covered by coating, which agrees well with SEM results. The XPS B1s spectra are shown in Fig. 4b. As shown in Fig. 4b, there is only one peak located at 187.5 eV, which is assigned to B-C bonds 20,29 . In addition, the B:C atomic ratio for six-hours-coated-diamond (D3) is close to 4:1, confirming the coating exhibit B 4 C phase. Figure 5 shows the density for diamond/Cu composite as a function of plating time. The theoretical density of composite ρ theoretical is determined by where ρ D (3.52 g/cm 3,30 ) and ρ M (8.96 g/cm 3,30 ) is theoretical density of diamond and copper matrix, respectively. In addition, V D (50 vt.%) and V M (50 vt.%) is the volume fraction of diamond particles and copper matrix, respectively. Therefore, the theoretical density of composite ρ theoretical is 6.24 g/cm 3 , which is also shown in Fig. 5. However, the density for uncoated-diamond/Cu composite (C0) exhibits a low value (5.56 g/cm 3 ). By plating boron carbide coating on diamond reinforcement particles, the diamond/Cu composite exhibits a significant densification. For the two-hours-coated diamond/Cu composite (C1), the density obviously increases to 5.86 g/ cm 3 . With further increasing plating time, the density of diamond/Cu composite continually increases. When the plating time increases to 6 hours, the density for sample D3 reaches 6.21 g/cm 3 , which is close to the theoretical density of composite. The relative density ρ relative is determined by where ρ measured is the measured density. Thereby, the relative density for uncoated-diamonds/Cu composite (C0) is 89.05%. Plating boron carbide coating on diamond reinforcement gives rise to a densification of composite. With increasing plating time of boron carbide coating to 6 hours, the composite shows a nearly full relative density (99.52%).
The tensile strength as a function of plating time is displayed in Fig. 6. As shown in Fig. 6, plating boron carbide coating on diamonds is benefited to the improvement of tensile strength of diamond/Cu composites. Furthermore, the tensile strength of diamond/Cu composite increases from 60 to 115 MPa with increasing plating time of boron carbide coating from 0 to 6 hours.
SEM micrographs of tensile fracture surfaces of the uncoated-diamond/Cu (C0) and coated-diamond/Cu (C1-3) composites are shown in Fig. 7. For the uncoated-diamond/Cu composite (C0, Fig. 7a), large amounts of wide gap around diamond particles is observed. Meanwhile, un-wetting phenomenon between copper matrix and diamond particles is existence, which is marked in Fig. 7a. As shown in Fig. 7b-d, plating boron carbide coating on diamonds gives rise to a narrow gap, indicating that plating boron carbide coating on diamonds contributed to the densification of diamond/Cu composite. Figure 8 shows the thermal conductivity of uncoated-diamond/Cu (C0) and coated-diamond/Cu (C1-3) composites. As shown in Fig. 8, the uncoated-diamond/Cu composite (C0) exhibits a low thermal conductivity of 210 W/mK, which is even lower than the thermal conductivity of pure copper (385 W/mK 7 ). By plating boron carbide on diamonds, the thermal conductivity of composite is significantly improved. The thermal conductivity of composite increases with the increase in plating time of coating, and reaches 687 W/mK when six-hours-coated diamonds are used.

Discussion
The relative density for uncoated-diamond/Cu composite (C0) is only 89.05%, indicating large amount of hole exists in the sample C0. Meanwhile, as shown in the cross section image (Fig. 6a), wide interfacial gap is observed between uncoated diamond and copper matrix. The separation between diamond and copper is formed during the cooling process because of large different expansion coefficients between copper (17.0 × 10 −6 °C −1 31 ) and carbon materials (1.0 × 10 −6 °C −1 32 ). Therefore, the low density for uncoated-diamond/Cu composite (C0) is where a is the radius of diamond particle, V D is the volume fraction of diamond particle, and V p is porosity. As displayed in Fig. 9, plating boron carbide coating on diamond reinforcement gives rise to a decrease in average gap width from 8.40 to 0.36 μm, which is agreement well with SEM results (Fig. 7). Because of the modest expansion coefficients of boron carbide (5.65 × 10 −6 °C −1 33 ), B 4 C interlayer is benefited to relieve the interfacial thermal stress between copper and diamond during cooling process. Weber and Tavangar 10 reports that 2.5 at.% boron alloyed in copper matrix effectively prevented the copper separated from diamond. Hu and Kong 8 also found that diamond/Cu composite exhibited a continual interface via forming a B 4 C interface between diamond and copper. In addition, Ahn et al. 34 and Mansourzadeh et al. 22 observed continual interfaces between copper and boron carbide in Cu/B 4 C composites. Therefore, we suggest that the formation of boron carbide is benefited to the densification of diamond/Cu composite. The relationship between tensile strength and average gap width is summarized in Fig. 10, wherein the tensile strength is strongly dependent on the average gap width. Together gap width data (Fig. 9) and SEM images (Fig. 7a), for uncoated-diamond/Cu composite the interface between copper matrix and diamond is very weak, which is not benefited to the stress transfer, resulting in a low tensile strength. According to the analysis of composite densification, plating boron carbide is benefited to the improvement of interface structure between diamond and copper matrix. Obviously, the gap around diamond contributed to the extension of the crack during tensile test. Therefore, plating boron carbide is benefited to the increase in tensile strength of diamond/Cu composite, since the decrease in average gap width between diamond and copper matrix. For six-hours-coated diamond/Cu composite (C3), there is no obvious gap between diamond and copper matrix. Thus the tensile strength reaches the maximum value, because the continual interface is conducive to distribution of stress.
where K m and K r are thermal conductivity of matrix and reinforcement particles respectively (K Cu =385 W/mK 7 , K diamond = 1350 W/mK 8 ), a is the radius of reinforcement particle, and V r is the volume fraction of reinforcement particle. The interfacial thermal conductance hc is identified as where ρ m and ρ r is theoretical density of matrix and reinforcement particle respectively. And v m and v r is phonon velocity in matrix and reinforcement particle respectively (v Cu = 2881 m/s, v diamond = 13924 m/s 8 ). c m is the specific heat of matrix (c Cu =385 J/kgK 8 ). It is worth noting that the additional thermal resistance caused by the interphase should be considered, because Eq. (6) is on the assumption of perfect bonding between matrix and reinforcements. However, in this work, the interface between copper matrix and diamond reinforcement particles is constituted by Cu/B 4 C interface, B 4 C interlayer and B 4 C/diamond interface. Therefore, based on the concept of an electrical resistance analogy the interfacial thermal conductance hc can be estimated by where h Cu-B4C , h B4C and h B4C-diamond is thermal conductance of Cu/B 4 C interface, B 4 C interlayer and B 4 C/diamond interface, respectively. K B4C (67 W/mK 8 ) is thermal conductivity of B 4 C, and t is thickness of B 4 C layer. The calculated results are compared with experimental data in Fig. 11a, wherein the results show a clear difference. The increase in plating time gives rise to an increase in B 4 C layer thickness, resulting in a decrease in theoretically estimated thermal conductivity. However, experimental data shows an increase with the increase in plating time.
Only for the six-hours-coated diamond/Cu composite (C3), the theoretically estimated data (707 W/mK) is close to experimental data (687 W/mK). Cross-section SEM image in Fig. 6 shows large amount of gap around diamond. However, the theoretically estimation is only in the consideration of perfect contact between matrix and reinforcement particles. It is well known that air (or vacuum) is an excellent heat insulation layer. Therefore, the existence of gap in composite contributed to the deterioration in thermal conductivity of composite. As shown in Fig. 11b, a narrower gap width is benefited to the improvement of thermal conductivity of composite, and the experimental data is close to the theoretically data when the gap is unapparent (sample C3). To summarize, the B 4 C coating on the diamond particles was synthesized by heating the diamond particles in a powder mixture consisting of H 3 BO 3 and B under Ar atmosphere. Coverage fraction of the B 4 C coating was strongly dependent on the synthesized time. Diamond/Cu composites were made by powder metallurgy. The addition of B 4 C coating gave rise to a dense composite. In addition, both tensile strength and thermal conductivity of composite were dependent on the interfacial gap width between diamond and copper matrix. When the B 4 C fully covered on diamond particles, the composite exhibited a nearly full relative density (99.52%) and narrow interfacial gap width (0.36 μm), leading to a greatly increase in tensile strength (115 MPa) which was much higher than that for uncoated-diamond/Cu (60 MPa) and diamond/CuB composites 3 . Meanwhile, the six-hours-coated-diamond/Cu composite exhibited a thermal conductivity of 687 W/mK, which was close to the theoretically estimated data (707 W/mK) and much higher than that of uncoated-diamond/Cu composite (210 W/mK).

Method
Diamond/Cu composite materials were made by powder metallurgy methods of vacuum hot-pressing sintering, which the process included mixing, compacting and sintering. Copper powder (99.9% purity) was used as matrix material, and synthetic HPHT diamond particles (HSD90, particle size 70/80 mesh (180 ~ 212 μm), Henan huanghe whirlwind international Co., Ltd., China) with boron carbide coating were used as reinforcements.
For forming boron carbide coating on diamond particles, 25 g of diamond particles were immersed in a mixture consisting of 33 g of boron (B), and 23 g of boric acid (H 3 BO 3 ) powders. The diamond-powder mixture was mixed using vigorously mechanical stirring at room temperature. This mixture was placed on an alumina boat and placed into a tube furnace. The diamond-powder mixture was heated to 1200 °C for 2, 4 and 6 hours in Ar atmosphere to synthesize the boron carbide coating on the diamond particles. The reaction equation was 3 3 1200 C 4 2 + +  →  + +°Th e Gibbs energy changes of the reactions calculated through thermodynamic calculation was −366.6 kJ/mol. After cooling of the furnace, the samples were treated with dilute nitric acid to remove B 2 O 3 from the product. The coated diamonds were separated from the excess B powders using a sieve. The coated diamond particles synthesized for 2, 4 and 6 is referred to as D1, D2 and D3, respectively. In addition, the uncoated diamond particles are referred to as D0.
Copper powders were mixed with diamond particles with boron carbide coating to prepare composite with 50 vol.% reinforcement. The furnace chamber was evacuated using a pump prior to sample heating. The composite powders were heated at 950 °C for 20 min and pressed by a pressure of about 60 MPa before being inside the furnace to room temperature. The vacuum was only broken when the furnace temperature reached room temperature. The composite with D0-3 as reinforcement is referred to as C0-3, respectively.
The microstructure of coating on diamond particles was characterized by X-ray diffraction (XRD) using a Bruker D8 with a Cu Kα source. To obtain more information regarding the bonding conditions, the coatings on diamond particles were also analysed by Raman spectroscopy using a Renishaw Micro Raman system 2000 with an excitation wavelength of 514 nm. The bonding states of coatings were further characterized with X-ray photoelectron spectroscopy (XPS) using a Physical Systems Quantum2000 spectrometer with monochromatic Al Kα radiation. The XPS analysis area was set to a diameter of 1 mm. The samples in the XPS analysis were cleaned by Ar + ion sputter etching with energy of 1 keV to remove the surface oxides. The surface morphologies and topographies of the films were characterized using a Hitachi S-4800 scanning electron microscope (SEM). The average thickness of coating was measured by SEM cross-section images, at least five measurements were carried out on each plane. The average thickness of coating is 0.2, 0.7 and 1.0 μm for two-hours-coated diamond (D1), four-hours-coated diamond (D2) and six-hours-coated diamond (D3), respectively.
The density of composite was measured by high precision ceramic porosity volume density tester (Dahometer, DE-120M) via Archimedes method. The tensile strength and fracture elongation were tested at room temperature by an electronic universal test machine (DDL 100, CIMACH, Changchun, China) at the speed of 0.18 mm/min. The morphologies of cross section were also obtained by SEM. Thermal diffusivity was measured by a laser flash method (ASTM E1461-13). The uncertainty in the thermal measurements is ±2%. Differential scanning calorimetry (DSC) analyses were performed to estimate the specific heat. Finally, the thermal conductivity was calculated by the product of density, thermal diffusivity and specific heat according to the following equation 36,37 : where α was thermal Diffusivity, ρ measured was the measured density and C was specific heat. Detail results were shown in Table 1.

Sample
Density g/cm 3 Specific heat J/g K Thermal diffusivity mm 2 /s Thermal conductivity W/(m·K)