Optimized thermal properties in diamond particles reinforced copper-titanium matrix composites produced by gas pressure infiltration

Composites: Part A 91 (2016) 189–194

Contents lists available at ScienceDirect

Composites: Part A journal homepage: www.elsevier.com/locate/compositesa

Optimized thermal properties in diamond particles reinforced coppertitanium matrix composites produced by gas pressure infiltration Jianwei Li a, Hailong Zhang a, Luhua Wang a, Zifan Che a, Yang Zhang a,1, Jinguo Wang b, Moon J. Kim b, Xitao Wang a,⇑ a b

State Key Laboratory for Advanced Metals and Materials, University of Science and Technology Beijing, Beijing 100083, China Department of Materials Science and Engineering, The University of Texas at Dallas, Richardson, TX 75080, USA

a r t i c l e

i n f o

Article history: Received 9 May 2016 Received in revised form 12 August 2016 Accepted 1 October 2016 Available online 5 October 2016 Keywords: A. Metal-matrix composites (MMCs) B. Interface/interphase B. Thermal properties E. Liquid metal infiltration

a b s t r a c t Interface modification is crucial to exploit high thermal conductive potential of diamond in the metal matrix composites reinforced with diamond particles (Cu/diamond composites). With an attempt to modify the Cu/diamond interface, we add a carbide-forming element of Ti to the Cu matrix and use a liquid-phase processing technique to attain sound interfacial bonding. The Cu-xTi/diamond composites were characterized by using scanning electron microscopy, transmission electron microscopy, and Xray diffraction. The interface layer is confirmed as TiC, the amount of which increases with increasing Ti concentration in the Cu-xTi alloy matrix. As the Ti concentration increases, the thermal conductivity of the Cu-xTi/diamond composites first increases and then decreases, giving an optimized thermal conductivity of 752 W/m K and a coefficient of thermal expansion of 6.50
106/K at x = 0.5 wt.%. The results show that an appropriate amount of Ti addition in Cu matrix can enhance the thermal conductivity of Cu/diamond composites. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction The miniaturization of electronics is urgently demanding packaging materials with superior thermal conductivity. Effective thermal management is fundamental to consistent long-term performance and reliability of electronic devices [1,2]. As a promising candidate, diamond particles dispersed copper matrix (Cu/diamond) composites are attracting more and more interests [3–17]. In preparing Cu/diamond composites, the major restriction originates from the mismatch in thermal expansion coefficient between diamond and copper, which easily induces large thermal stress. In addition, the non-carbide forming nature of copper and the large contact angle between diamond and Cu mean an absence of strong chemical bonds at the interface. Consequently, the diamond particles are inclined to delaminate from the pure copper matrix after cooling down. Poor interfacial bonding becomes a main obstacle to obtaining high thermal conductivity in the Cu/diamond composites. The thermal conductivity of the Cu/diamond composites is reduced even with increasing the amount of diamonds [4]. ⇑ Corresponding author. E-mail address: [email protected] (X. Wang). Present address: Advanced Manufacture Technology Center, China Academy of Machinery Science & Technology, Beijing 100083, China. 1

http://dx.doi.org/10.1016/j.compositesa.2016.10.005 1359-835X/Ó 2016 Elsevier Ltd. All rights reserved.

At present, it is popular to improve the interfacial bonding by introducing a third component between diamond and copper, usually through surface metallization of diamond particles [3,7,8,11,18] or through alloying of copper matrix [6,19,20]. Among broad researches, TiC is suggested to connect the diamond reinforcements with the metal matrix tightly. TiC has a Gibbs free energy of 170.5 kJ/mol [21], which is favorable to its formation at the interface. Through Ti coating on diamond particles, the derived Cu-Ti/diamond composites display a thermal conductivity as high as 630 W/m K, owing to TiC formation at the interface [22]. Nevertheless, the coating on diamond is expensive and demanding. Moreover, the coating is oxidized during high-temperature processing [7]. The presence of oxygen on TiC surface strongly inhibits the interaction between the carbide and the molten Cu matrix, causing non-wetting conditions [23]. By means of alloying Ti to Cu matrix, the diamond particles and the Cu-Ti alloys can react directly, and the formation of oxide layer on the diamond particles is thus avoided. From the above, the route of Cu matrix alloying is proved to be more feasible in the fabrication of Cu/diamond composites. Here we add Ti to Cu matrix to demonstrate the enhancement of thermal conductivity in the Cu-Ti/diamond composites. Intuitively, the amount of Ti addition should be limited in order to maintain high thermal conductivity in the Cu matrix. Literature

190

J. Li et al. / Composites: Part A 91 (2016) 189–194

has not reported detailed information about the effect of alloying element content on the microstructure and thermal properties of the composites. Several groups [20,24] prepare the Cu-Ti/ diamond composites using a solid-state sintering route, but obtain relatively low thermal conductivities. A liquid-state processing route of gas pressure infiltration has been successfully utilized to achieve high thermal conductivity in Cu/diamond composites [7,16]. The gas pressure infiltration is capable of realizing dense microstructure and promoting interfacial reaction in the diamond composites. So far the promising route has not been used to prepare the Cu-Ti/diamond composites. Another difficulty in Cu/diamond composites community comes from delicate characterization of the Cu/diamond interface. Due to huge difference in hardness between diamond and metal, it is extremely tough to mill an eligible specimen for transmission electron microscopy (TEM) observations. For this reason, actual interface structure of the Cu/diamond composites is rarely reported [10,15], which is certainly critical to understand the thermal conducting mechanism involved. Taking advantage of focused ion beam (FIB) technique, we have developed thin foils containing the Cu/diamond interface and obtained the state-of-the-art images depicting carbides at the interface. In this article, we produce the Cu-Ti/diamond composites by the gas pressure infiltration and investigate the effect of Ti addition on the microstructure and thermal properties with respect to diamond surface state and interface structure of the composites. The effect of interfacial layer thickness on thermal conductivity of the Cu/diamond composites is elucidated. The thermal conductivity and coefficient of thermal expansion (CTE) of the Cu-Ti/diamond composites are correlated with various modeling schemes.

Fig. 1. Schematic drawing of the gas pressure infiltration device. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

according to the equation K = aqc, where q is sample density and c is specific heat. The density was measured by the Archimedes method using alcohol as the immersion medium, and the measurement was repeated three times for each sample. The heat capacities were calculated from the rule of mixture based on mass fraction of each component. The CTEs of the composites were measured by a dilatometer (DIL 402C, NETZSCH, Germany) with a heating rate of 5 K/min in the temperature range between 323 K and 573 K, using a cylinder 5 mm in diameter and 25 mm in length. 3. Results and discussion

2. Experimental The Cu-xTi/diamond composites were produced by a gas pressure infiltration method. Fig. 1 shows the schematic drawing of the gas pressure infiltration device. The starting materials were HHD90 synthetic single-crystalline diamond powders with an average particle size of 230 lm (Henan Huanghe Whirlwind Co., China) and Cu-xTi alloys (x = 0.3, 0.5, 2.0 wt.%). The Cu-xTi alloys were melted by a vacuum induction route using 99.999 wt.% Cu and 99.99 wt.% Ti bulks (TaiYu Materials Science & Technology Co., China). The infiltration was conducted at 1423 K for 30 min, under an Ar gas pressure of 1.0 MPa. The details of the infiltration are referred to elsewhere [7]. X-ray diffraction (XRD, Rigaku DMAX-RB, Japan) was used to characterize the phase structure of the composites. The polished surface and fracture surface were observed by field emission scanning electron microscope (SEM, ZEISS SUPRA 55, Germany). A dual beam FIB system (FEI Nova 200 FIB, USA) was used to mill the CuTi/diamond samples to thin foils. The FIB-milled samples were characterized using a scanning transmission electron microscope (STEM, JEOL, ARM200, Japan). Besides, the metal matrix was electrochemically etched and the diamond particles were collected to directly characterize the interface structure [25]. The composites were electrochemically etched in a 10 vol.% HNO3 and 90 vol.% H2O solution for 5 min. The sample acts as the anode and a steel plate as the cathode. The current density was controlled by a direct current power supply and the current was fixed at 1 A. After electrochemical etching, the samples were cleaned ultrasonically in an acetone bath for 40 s. The thermal diffusivity (a) was measured by a laser flash method (LFA427, NETZSCH, Germany) with an international standard using a disc 10 mm in diameter and 3 mm in thickness. Three times were repeated for one sample to get an average value of the thermal diffusivity. The thermal conductivity (K) can be calculated

Fig. 2 shows the representative SEM images of the polished surface and fracture surface of the Cu-0.5 wt.%Ti/diamond composite. The diamond particles are found to be uniformly distributed in the Cu matrix (Fig. 2a) and no noticeable defects such as cracks or flaws are observed at the interface. The pull-out of diamond particles is rarely seen in the polished surface, which indicates strong interface bonding between the diamond particles and Cu matrix. The diamond particles maintain their original shape without any degradation. The observations demonstrate excellent quality of the composites produced by the gas pressure infiltration. As seen from the fractured surface, the Cu matrix closely adheres to the surfaces of the diamond particles (red circles in Fig. 2b). Some diamond particles are found to fracture transgranularly. This fracture only occurs when the interfacial bonding strength is higher than the fracture strength of the diamond particles. The synthetic diamond has some flaws inside. It can be concluded that Ti addition has significantly improved the interfacial bonding between the Cu matrix and diamond particles. In order to study the effect of Ti addition on the phase composition, the XRD patterns of the Cu-Ti/diamond composites were characterized, as shown in Fig. 3a. The results indicate the coexistence of Cu, diamond and TiC phases in the composites. Fig. 3b shows the XRD patterns of the diamond particles extracted from the Cu-Ti/diamond composites. Only diamond and TiC phases are detected. The morphology of the diamond particles is clearly seen in the inset to Fig. 3b. The diamond particle surface is covered uniformly by the carbides. Owing to the interfacial reaction between the Cu-Ti matrix and the diamond, TiC is formed during the infiltrating process by the reaction between Ti and C atoms (Ti + C ? TiC). Thermodynamic calculation shows that titanium carbide will be produced by the reaction of carbon with copper-tin-titanium alloys that have a titanium activity larger than 0.1 at 1400 K and close to 1.0 at 1423 K [26], which is the infiltration temperature

J. Li et al. / Composites: Part A 91 (2016) 189–194

191

Fig. 2. SEM observations of (a) the polished surface and (b) the fractured surface of the Cu-0.5 wt.%Ti/diamond composite. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

in this study. Accordingly, the Ti atoms dissolved in the Cu matrix will react with the C atoms on diamond surface. The peak intensity grows stronger with increasing Ti content, indicating more amount of TiC formed. The interfacial bonding of the Cu-Ti/diamond composites is therefore enhanced by the interfacial phase TiC. To further study the effect of Ti addition on interface microstructure, the FIB-milled thin foils were used for SEM characterizations. As shown in Fig. 4a and b, it is clear that the composites consist of three distinct zones: the matrix, the interphase and the diamond. The diamond is found to be bound tightly to the matrix. The carbides act as a binder between the diamond and the matrix. The average size of the carbides is found to increase from 100 nm to 200–300 nm when Ti content increases from 0.3 wt.% to 0.5 wt.%. The carbides are not continuous and they are separated by the Cu alloy matrix. The result is explained by the heterogeneous nucleation of carbides at interface and the following preferential growth along certain crystal orientations from the diamond into the Cu alloy matrix. This is similar to carbide characteristics in the Al/diamond composites [27]. When Ti content is further increased to 2.0 wt.%, a continuous layer of carbides is formed, with a thickness of roughly 300–500 nm. The SEM image in Fig. 4c was taken from diamond particles extracted from the composite sample. As the carbide grains grow into the Cu alloy matrix, the transport of carbon source from the diamond is restrained and then new carbide grains nucleate between the primary carbides. Accordingly, continuous carbides are derived on the diamond surface. The thermal conductivity K of the Cu-Ti/diamond composites exhibits an apparent variation with Ti addition (Fig. 5). With

Fig. 3. XRD patterns of (a) the Cu-Ti/diamond composites produced by gas pressure infiltration and of (b) the diamond particles extracted from the Cu-Ti/diamond composites. The inset to (b) shows SEM images of a diamond particle extracted from the Cu-0.5 wt.%/diamond composite. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

increasing Ti content the K value firstly increases and then decreases, attaining a maximum value of 752 W/m K at 0.5 wt.% Ti. The unmodified Cu/diamond composite shows a low K value of 141 W/m K, owing to poor interfacial bonding between diamond and Cu. As the formed TiC connects the two phases tightly, phonon transfer across the interface is intensified and a high K value is obtained. Nevertheless, a redundant amount of Ti addition results in a K value as low as 123 W/m K at 2.0 wt.% Ti, even lower than that for unmodified Cu/diamond composite. Since TiC possesses a low K of just 21 W/m K [28], lower than both Cu matrix and diamond, the TiC layer is harmful to heat transport from diamond to Cu. Combined with the SEM images in Fig. 4, the variation of thermal conductivity is explained by the models in Fig. 6. The total interfacial thermal resistance in the transition region Ri can be assumed as a parallel of the thermal resistances RTiC and RCuTi, where the subscripts TiC and CuTi refer to TiC layer and Cu alloy matrix, respectively. When Ti content increases from 0.3 to 0.5 wt.%, the average size of the carbides at the interface is found to increase from 100 nm to 200–300 nm. The decrease in number of grain boundary is beneficial to heat transfer across the interface. When Ti content is below 0.5 wt.%, the carbides are not continuous and they are separated by the Cu alloy matrix. The remaining Cu alloy is still a continuous network and helps to maintain high ther-

192

J. Li et al. / Composites: Part A 91 (2016) 189–194

Fig. 4. Electron microscopy images showing the interfaces between diamond and Cu-xTi matrix: (a) x = 0.3 wt.%, TEM, (b) x = 0.5 wt.%, TEM, and (c) x = 2.0 wt.%, SEM.

The resultant 752 W/m K is protruding among K values reported for Cu-Ti/diamond composites: 670 W/m K via spark plasma sintering (SPS) [20,22] and 608 W/m K via pressureless sintering method [24]. However, the value is still lower than the theoretical value of 1047 W/m K, which is calculated by using a differential effective medium (DEM) model [29]:

 13 eff kc ðkd  kc Þ ð1  V d Þ ¼ eff km ðkd  km Þ

Fig. 5. Variation of thermal conductivity of the Cu-Ti/diamond composites with Ti content. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

mal conductivity. With increasing Ti content from 0.3 to 0.5 wt.%, the interfacial bonding could be improved with mechanical interlocking by larger TiC particles that are penetrated deeper into the Cu matrix. The above-mentioned two factors promote the thermal conductivity of the Cu-xTi/diamond composites. However, the gap between TiC particles decreases with increasing Ti content. This could impair the thermal conductivity of the composites because of the narrowed channel for heat transfer. By a compromise, the thermal conductivity increases from 710 to 752 W/m K with increasing Ti content from 0.3 to 0.5 wt.%. When Ti content is further increased to 2.0 wt.%, a continuous TiC layer is formed, which separates the diamond and the Cu alloy matrix completely. Since TiC has a low thermal conductivity, the thermal conductivity of the Cu-xTi/diamond composites drops dramatically.

eff

with kd ¼

kd kd 1 þ ah c

ð1Þ

where kc, kd and km are the thermal conductivities of composite, dispersed reinforcement and matrix, respectively. The a and Vd are the radius and volume fraction of spherical reinforcement, respectively, and the hc is interfacial thermal conductance. Accordingly, the concentration of Ti addition requires further fine optimization to achieve higher thermal conductivity in the Cu-Ti/ diamond composites. This job is left for future study. Fig. 7(a) shows the CTE values for the Cu-0.5 wt.%Ti/diamond and unmodified Cu/diamond composites. The CTE of the Cu0.5 wt.%Ti/diamond composite was 6.50
106/K at 323 K, which is compatible with the CTE of semiconductor materials [30] and is lower than reported values for Cu/diamond composites [8,31,32]. The unmodified Cu/diamond composite has a higher CTE of 7.26
106/K at 323 K. When the bonding between the matrix and the diamond is weak, the Cu matrix would dominate the CTE value of the composite. On the contrary, in the Cu0.5 wt.%Ti/diamond composite, both the Cu matrix and the diamond would contribute to the CTE of the composite because of the sound interface bonding. The CTE value increases with increasing temperature, which is consistent with the Grueneisen theory [33]. The CTE value increases a little bit slower after 473 K. In the fabricated Cu/diamond composites, the thermal stress exhibits such as tensile stress in the Cu matrix and compressive stress in the diamond. When the specimens are heated from room temperature, the residual stresses generated during fabrication are

193

J. Li et al. / Composites: Part A 91 (2016) 189–194

Fig. 6. Interfacial thermal resistance in the Cu-Ti/diamond composites: (a) schematic illustration of Cu-TiC-diamond transition region and (b) a parallel analogue of thermal resistance in the transition region. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

released gradually and this is helpful to increase CTE; at higher temperature, the promotion to CTE of residual stress does not occur any more. Therefore, the increasing rate of CTE is lowered between 473 K and 573 K. Fig. 7(b) shows both the measured and predicted CTE values for the Cu/diamond composites. Two existing models are applied to calculate the theoretical values of CTE [34,35]: (1) The Turner model

ac ¼

V r K r ar þ V m K m am V r Kr þ V mKm

(2) The Kerner model

ac ¼ V r ar þ V m am þ V r V m

Fig. 7. (a) Coefficient of thermal expansion of the Cu-Ti/diamond composites and thermal linear expansion curve for a heating and cooling cycle of the Cu-0.5 wt.% Ti/diamond composite; (b) comparison of the CTEs between predictions and experimental results. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

ð2Þ ðar  am ÞðK r  K m Þ K r V r þ K m V m þ 3K4Gr Kmm

ð3Þ

where a, V, K and G are the CTE, volume fraction, bulk modulus and shear modulus, and c, r, m are indexes of composite, reinforcement and matrix, respectively. In this study, V diamond = 0.61 and V Cu = 0.39. The other parameters [17] used for calculation are adiamond = 2.30
106/K, aCu = 16.42
106/K, K diamond = 580 GPa, K Cu = 140 GPa, Gdiamond = 360 GPa and GCu = 49 GPa. Unlike the Turner model, the Kerner model takes shear strain into consideration. In Fig. 7(b) the two Kerner curves are obtained by reversing the role of copper matrix and diamond reinforcement, since the diamond and the copper have comparable volume fractions in the composites. Without regard to the effect of the third component (alloying element or interphase) on the CTE, the calculations for the Cu/diamond composites are obtained with the three models. As shown in Fig. 7(b), the experimental results are all in good agreement with the values predicted by the Kerner copper matrix model, confirming the presence of shear effects in the composites. The measured CTEs for Cu/diamond composites are apart from the Kerner diamond matrix model. The deviation is understood by the fact that the diamond particles do not form a continuous phase, as shown in Fig. 2(a), although they possess a higher volume fraction than copper phase. The measured CTE for the Cu-0.5 wt.%Ti/diamond composite is below the Kerner copper matrix line. This could be ascribed to the effect of Ti alloying on the shear modulus of the copper matrix.

194

J. Li et al. / Composites: Part A 91 (2016) 189–194

The thermal expansion curve is shown in the inset to Fig. 7(a). A residual strain (ep) of 0.40112
103 is derived after a thermal cycle from 323 K to 640 K. The low ep value is acceptable for practical applications of electronic packaging materials.

[12]

[13]

4. Conclusion The Cu-xTi/diamond composites (x = 0.3, 0.5, 2.0 wt.%) were produced by the gas pressure infiltration method, with a thin and discontinuous layer of TiC formed at the interface. The presence of TiC is favorable to the significant enhancement of interfacial thermal conductance, giving a high thermal conductivity of 752 W/m K at x = 0.5 wt.%. The Cu-0.5 wt.%Ti/diamond composite displays a coefficient of thermal expansion of 6.50
106/K that is compatible with semiconductors. The finding suggests that Ti alloying in the metal matrix is an effective method to achieving high thermal conductivity in Cu/diamond composites.

[14]

[15]

[16]

[17] [18]

[19]

Acknowledgements

[20]

This work was financially supported by the National Natural Science Foundation of China (No. 51271017, 51301018) and International Science and Technology Cooperation Program of China (No. 2014DFA51610).

[21]

[22]

References [1] Zweben C. Advances in high-performance thermal management materials: a review. Covina, CA, ETATS-UNIS: Society for the Advancement of Material and Process Engineering; 2007. [2] Zweben C. Thermal materials solve power electronics challenges. Power Electron Technol 2006;32:40–7. [3] Zhang C, Wang R, Cai Z, Peng C, Feng Y, Zhang L. Effects of dual-layer coatings on microstructure and thermal conductivity of diamond/Cu composites prepared by vacuum hot pressing. Surf Coat Technol 2015;277:299–307. [4] Muhammad Z, Raza K, Khalid FA, Mabrouki T. Numerical investigation of the effect of interfacial thermal resistance upon the thermal conductivity of copper/diamond composites. Mater Des 2015;86:248–58. [5] Ren S, Hong Q, Chen J, He X, Qu X. The influence of matrix alloy on the microstructure and properties of (flake graphite + diamond)/Cu composites by hot pressing. J Alloy Compd 2015;652:351–7. [6] Rape A, Gott K, Kulkarni A, Singh J. Simulation of matrix conductivity in copper–diamond composites sintered by field assisted sintering technology. Comp Mater Sci 2015;110:29–33. [7] Li J, Zhang H, Zhang Y, Che Z, Wang X. Microstructure and thermal conductivity of Cu/diamond composites with Ti-coated diamond particles produced by gas pressure infiltration. J Alloy Compd 2015;647:941–6. [8] Kang Q, He X, Ren S, Liu T, Liu Q, Wu M, Qu X. Microstructure and thermal properties of copper–diamond composites with tungsten carbide coating on diamond particles. Mater Charact 2015;105:18–23. [9] Chu K, Jia C, Guo H, Li W. Microstructure and thermal conductivity of Cu– B/diamond composites. J Compos Mater 2012;47:2945–53. [10] Grzonka J, Kruszewski MJ, Rosin´ski M, Ciupin´ski Ł, Michalski A, Kurzydłowski KJ. Interfacial microstructure of copper/diamond composites fabricated via a powder metallurgical route. Mater Charact 2015;99:188–94. [11] Abyzov AM, Kruszewski MJ, Ciupin´ski Ł, Mazurkiewicz M, Michalski A, Kurzydłowski KJ. Diamond–tungsten based coating–copper composites with

[23] [24]

[25]

[26] [27]

[28] [29]

[30] [31]

[32] [33] [34] [35]

high thermal conductivity produced by Pulse Plasma Sintering. Mater Des 2015;76:97–109. Sinha V, Spowart JE. Influence of interfacial carbide layer characteristics on thermal properties of copper-diamond composites. J Mater Sci 2013;48:1330–41. Shen X, He X, Ren S, Zhang H, Qu X. Effect of molybdenum as interfacial element on the thermal conductivity of diamond/Cu composites. J Alloy Compd 2012;529:134–9. Ren S, Shen X, Guo C, Liu N, Zang J, He X, Qu X. Effect of coating on the microstructure and thermal conductivities of diamond–Cu composites prepared by powder metallurgy. Compos Sci Technol 2011;71:1550–5. Schubert T, Ciupinski L, Zielinski W, Michalski A, Weissgaber T, Kieback B. Interfacial characterization of Cu/diamond composites prepared by powder metallurgy for heat sink applications. Scr Mater 2008;58:263–6. Weber L, Tavangar R. On the influence of active element content on the thermal conductivity and thermal expansion of Cu–X (X=Cr, B) diamond composites. Scr Mater 2007;57:988–91. Yoshida K, Morigami H. Thermal properties of diamond/copper composite material. Microelectron Reliab 2004;44:303–8. Zhang C, Wang R, Cai Z, Peng C, Wang N. Low-temperature densification of diamond/Cu composite prepared from dual-layer coated diamond particles. J Mater Sci: Mater Electron 2015;26:185–90. Rosinski M, Ciupinski L, Grzonka J, Michalski A, Kurzydlowski KJ. Synthesis and characterization of the diamond/copper composites produced by the pulse plasma sintering (PPS) method. Diam Relat Mater 2012;27–28:29–35. Che QL, Chen XK, Ji YQ, Li YW, Wang LX, Cao SZ, Jiang YG, Wang Z. The influence of minor titanium addition on thermal properties of diamond/copper composites via in situ reactive sintering. Mater Sci Semicon Proc 2015;30:104–11. Rape A, Liu X, Kulkarni A, Singh J. Alloy development for highly conductive thermal management materials using copper-diamond composites fabricated by field assisted sintering technology. J Mater Sci 2013;48:1262–7. Che QL, Zhang JJ, Chen XK, Ji YQ, Li YW, Wang LX, Cao SZ, Guo L, Wang Z, Wang SW, Zhang ZK, Jiang YG. Spark plasma sintering of titanium-coated diamond and copper–titanium powder to enhance thermal conductivity of diamond/copper composites. Mater Sci Semicon Proc 2015;33:67–75. Froumin N, Frage N, Polak M, Dariel MP. Wetting phenomena in the TiC/(Cu– Al) system. Acta Mater 2000;48:1435–41. Chung C, Lee M, Tsai M, Chu C, Lin S. High thermal conductive diamond/Cu–Ti composites fabricated by pressureless sintering technique. Appl Therm Eng 2014;69:208–13. Monje IE, Louis E, Molina JM. Aluminum/diamond composites: a preparative method to characterize reactivity and selectivity at the interface. Scr Mater 2012;66:789–92. Standing R, Nicholas M. The wetting of alumina and vitreous carbon by copper-tin-titanium alloys. J Mater Sci 1978;13:1509–14. Che ZF, Zhang Y, Li JW, Zhang HL, Wang XT, Sun C, Wang JG, Kim MJ. Nucleation and growth mechanisms of interfacial Al4C3 in Al/diamond composites. J Alloy Compd 2016;657:81–9. Williams W. The thermal conductivity of metallic ceramics. JOM 1998;50:62–6. Tavangar R, Molina JM, Weber L. Assessing predictive schemes for thermal conductivity against diamond-reinforced silver matrix composites at intermediate phase contrast. Scr Mater 2007;56:357–60. Cengel YA, Klein S, Beckman W. Heat transfer: a practical approach. Boston: WCB McGraw-Hill; 1998. Fan YM, Guo H, Xu J, Chu K, Zhu XX, Jia CC, Yin FZ, Zhang XM. Pressure infiltrated Cu/diamond composites for LED applications. Rare Met 2011;30:206–10. Hu H, Kong J. Improved thermal performance of diamond-copper composites with boron carbide coating. J Mater Eng Perform 2014;23:651–7. Nix FC, MacNair D. The thermal expansion of pure metals: copper, gold, aluminum, nickel, and iron. Phys Rev 1941;60:597–605. Turner PS. Thermal-expansion stresses in reinforced plastics. J Res NBS 1946. Kerner E. The elastic and thermo-elastic properties of composite media. Proc Phys Soc Sect B 1956;69:808.