Diamond/Copper composite material, break the limit!
With the continuous miniaturization, integration, and high performance of modern electronic devices, including computing, 5G/6G, batteries, and power electronics, the increasing power density leads to severe joule heat and high temperatures in the device channels. This is followed by performance degradation and device failure. Efficient heat dissipation is becoming an important problem in electronic products. To mitigate this problem, the integration of advanced thermal management materials on electronic devices can significantly improve their heat dissipation capabilities.
Diamond has excellent thermal properties, the highest isotropic thermal conductivity of all bulk materials (k= 2300W/mK), and has an ultra-low coefficient of thermal expansion at room temperature (CTE=1ppm/K). Diamond particle reinforced copper matrix (diamond/copper) composites, as a new generation of thermal management materials, have received great attention due to their potential high k value and adjustable CTE.
However, there are significant mismatches between diamond and copper in many properties, including but not limited to CTE (a clear difference in order of magnitude, as shown in Figure (a)) and chemical affinity (no solid solution, no chemical reaction, as shown in Figure (b)).
Significant performance differences between copper and diamond (a) coefficient of thermal expansion (CTE) and (b) phase diagram Source: Paper
These mismatches inevitably result in low bond strength and high thermal stress at the diamond/copper interface inherent in the high temperature manufacturing or integration process of diamond/copper composites. As a result, diamond/copper composites will inevitably encounter interface cracking problems, and the heat conductivity will be greatly reduced (when diamond and copper are directly combined, its k value is even much lower than pure copper (< 200W/mK)).
At present, the main improvement method is to chemically modify the diamond/diamond interface through metal alloying or surface metallization. The transitional interlayer formed on the interface will improve the interface bonding force, and the relatively thick interlayer is more conducive to withstand the interface cracking. As mentioned in the references, to achieve bonding, the thickness of the interlayer needs to be hundreds of nanometers or even micrometers. However, transitional interlayers on the diamond/copper interface, such as carbides (TiC, ZrC, Cr3C2, etc.), have lower intrinsic thermal conductivity (<25W/mK, several orders of magnitude smaller than diamond or copper). From the point of view of improving the interface heat dissipation efficiency, it is necessary to minimize the thickness of the transition sandwich, because according to the thermal resistance series model, the interface thermal conductivity (G copper-diamond) is inversely proportional to the thickness of the sandwich (d) :
The relatively thick transitional interlayer is conducive to improving the interface bonding force of diamond/diamond interface, but the excessive thermal resistance of the interlayer is not conducive to the interface heat transfer. Therefore, a major challenge in integrating diamond and copper is to maintain a high interfacial bonding strength while not introducing excessive interfacial thermal resistance when adopting interfacial modification methods.
The chemical state of the interface determines the interfacial bonding strength between heterogeneous materials. For example, chemical bonds are much higher than van der Waals forces or hydrogen bonds. On the other hand, the thermal expansion mismatch between the two sides of the interface (where T refers to CTE and temperature, respectively) is another key factor in determining the interface bonding strength of diamond/copper composites. As shown in Figure (a) above, the coefficient of thermal expansion of diamond and copper is clearly different in order of magnitude.
In general, thermal expansion mismatches have been a key factor affecting the performance of many composites, as the density of dislocations around the fillers increases significantly during cooling, especially in metal-matrix composites reinforced with non-metallic fillers. Such as AlN/Al composites, TiB2/Mg composites, SiC/Al composites and diamond/copper composites studied in this paper. At the same time, the diamond/copper composite is prepared at a higher temperature, usually greater than 900°C in traditional processes. The obvious thermal expansion mismatch is easy to generate thermal stress in the tensile state of the diamond/copper interface, resulting in a sharp decline in interface adhesion and even interface failure.
In other words, the interfacial chemical state determines the theoretical potential of the interfacial bond strength, and the thermal mismatch determines the degree of decline of the interfacial bond strength after the high temperature preparation of the composite material. Therefore, the final interface binding force is the result of the game between the above two factors. However, most current studies focus on improving the interface bonding strength by adjusting the chemical state of the interface, such as by the type, thickness and morphology of the transition interlayer. However, the decrease of interface bond strength caused by serious thermal mismatch has not been paid enough attention.
Concrete experiment
As shown in Figure (a) below, the preparation process consists of three main stages. First, an ultra-thin Ti coating with a nominal thickness of 70nm was deposited on the surface of diamond particles (Model: HHD90, mesh: 60/70, Henan Huanghe Cyclone Co., LTD., China) at 500°C by RF magnetron sputtering deposition method. The high purity titanium plate (purity: 99.99%) is used as the titanium target (source material), and argon (purity: 99.995%) is used as the sputtering gas. The thickness of the titanium coating is controlled by controlling the deposition time. In the deposition process, the substrate rotation technology is used to expose all faces of the diamond particles to the sputtering atmosphere, and the Ti element is uniformly deposited on all surface planes of the diamond particles (mainly including two facets: (001) and (111)). Secondly, 10wt% alcohol is added in the wet mixing process to make the diamond particles evenly distributed in the copper matrix. Pure copper powder (purity: 99.85wt%, particle size: 5 ~ 20μm, China Zhongnuo Advanced Material Technology Co., LTD.) and high-quality single crystal diamond particles are used as the matrix (55vol%) and reinforcement (45vol%), respectively. Finally, the alcohol in the prepressed composite is removed with a high vacuum of 10-4Pa, and then the copper and diamond composite is densified by powder metallurgy (spark plasma sintering, SPS).
(a) Schematic diagram of the preparation process of diamond/copper composites; (b) Different sintering processes in SPS powder metallurgy preparation
In the SPS preparation process, we innovatively proposed a low temperature high pressure (LTHP) sintering process and combined it with the interface modification of an ultra-thin coating (70nm). In order to reduce the introduction of thermal resistance of the coating itself, an ultra-thin interface modified layer (70nm) was used in this work. For comparison, we also prepared the composites using the traditional high temperature low pressure (HTLP) sintering process. The HTLP sintering process is a traditional formulation that has been widely used in previously reported works to integrate diamond and copper into dense composites. This HTLP process typically uses a high sintering temperature of > 900°C (close to the melting point of copper) and a low sintering pressure of ~ 50MPa. However, in our proposed LTHP process, the sintering temperature is designed to be 600°C, well below the melting point of copper. At the same time, by replacing the traditional graphite mold with a cemented carbide mold, the sintering pressure can be greatly increased to 300MPa. The sintering time of the above two processes is 10 minutes. In the supplementary materials, we have made a supplementary explanation on the optimization of LTHP process parameters. Detailed experimental parameters for different processes (LTHP and HTLP) are shown in Figure (b) above.
Conclusion
The above research aims to overcome these challenges and elucidate the mechanisms for improving the heat transfer performance of diamond/copper composites.
1. A new integrated strategy was developed to combine ultra-thin interface modification with LTHP sintering process. The obtained diamond/copper composite achieves a high k value of 763W/mK and a CTE value of less than 10ppm/K. At the same time, a higher k value can be obtained at a lower diamond volume fraction (45%, compared to 50%-70% in traditional powder metallurgy processes), which means that costs can be significantly reduced by reducing the content of diamond fillers.
2. Through the proposed strategy, the fine interface structure is characterized as a diamond /TiC/CuTi2/Cu layered structure, which greatly reduces the transition interlayer thickness to ~ 100nm, far less than the hundreds of nanometers or even a few microns previously used. However, due to the reduction of thermal stress damage during the preparation process, the interfacial bond strength is still improved to the covalent bond level, and the interfacial bond energy is 3.661J/m2.
3. Due to the ultra-thin thickness, the carefully made diamond/copper interface transition sandwich has low thermal resistance. At the same time, MD and Ab-initio simulation results show that the diamond/titanium carbide interface has good phonon property matching and excellent heat transfer capability (G>800MW/m2K). Therefore, the two possible heat transfer bottlenecks are no longer the limiting factors at the diamond/copper interface.
4. The interface bond strength is effectively improved to the level of covalent bond. However, the interfacial heat transfer capacity (G= 93.5MW/m2K) was not affected, resulting in an excellent balance between the two key factors. The analysis shows that the simultaneous improvement of these two key factors is the reason for the excellent thermal conductivity of diamond/copper composites.
