Microstructural Evolution of Ni-SiC Composites Manufactured by Spark Plasma Sintering

Silicon carbide (SiC), due to its unique properties, seems to be one of the most promising materials for modern applications. Due to its wide band gap value, high breakdown electric field, good thermal and electrical conductivity, and high hardness and mechanical strength, SiC can be applied in the electronic (e.g., high power/high frequency devices) or automotive (e.g., brake discs) industry.^[1] Silicon carbide is also used as the reinforcement in metal,^[2‐8] ceramic,^[9,10] or polymer matrix composites.^[11,12] SiC addition can improve the thermal or mechanical properties,^[13] especially in the case of metal matrix composites.

The group of nickel matrix composites is of special interest. There are many works dedicated to the manufacturing and characterization of Ni-SiC composites in the form of coatings^[14‐17] and bulk materials.^[18‐20]

The pulse electrodeposition technique is the most commonly used for the fabrication of nickel-based composite coatings. In the case of layered systems nanoparticles of silicon carbide are typically used^[18,21,22] which is justified due to the low thickness of the coating. For thin films, the possible reactions depend on the relative amounts of Ni and Si available for the mutual interaction, the process temperature and atmosphere, and the impurities contained in the layers. Low processing temperature (electroforming) or short time of high-temperature exposure (laser deposition) can strongly limit or even exclude the interfacial reaction in opposition to the sintered materials.

For the synthesis of Ni-SiC in the bulk form, powder metallurgy methods can be applied. No data about Ni-SiC composites obtained via the spark plasma sintering method can be found in the literature. The main technological problem of obtaining Ni-SiC composites is the mutual reactivity of the components of the composite occurring at elevated temperatures. At temperatures exceeding 500 °C, nickel reacts with silicon carbide, which causes SiC decomposition and the creation of pure components (Si and C).^[23] As a result, free carbon is precipitated and silicon interacts with nickel, forming new Ni-Si phases. Based on the equilibrium phase diagram, nine Ni-Si compounds can be distinguished, e.g., Ni₃Si, Ni₃₁Si₁₂, Ni₂Si, Ni₃Si₂, NiSi, and NiSi₂. The authors of^[24] indicated that individual Ni-Si phases are formed as a result of melting and casting and the solid-state reaction between Ni and Si.

In the case of sintering materials showing mutual solubility, intermediate phases with varying component concentrations may appear at particular stages of the densification process. In such a case, the free energy of the sintered system is decreased not only as a result of the reduction of the pore surface area but also as a result of mutual diffusion contributing to the alignment of the material composition or leading to the formation of stable intermediate phases.^[25] Several papers analyzing the Ni-SiC behavior at elevated temperatures can be found in the literature.^[26‐30] The investigations mainly concern the post-manufacturing thermal treatment of Ni-SiC systems and its effect on the structural evolution and physical properties. There are many sophisticated experimental tools applied to describe the correlation between the material structure, microstructure, and properties vs. temperature. For example, the authors of^[14] examined the influence of the heat treatment on the tensile strength of the Ni-SiC composite. The almost 50 pct increase in the maximum strength was obtained for materials annealed in a vacuum for 24 hours at 300 °C compared to the reference sample. The increase in the temperature to 600 °C caused the drastic decrease in the material strength. It was explained by the interfacial reaction and formation of new compounds (Ni₃Si and a little amount of Ni₃₁Si₁₂) and a change in the fracture mechanisms of the material. In,^[31] the interfacial interaction and phase changes occurring during heating (600 °C to 1000 °C) of the nickel thin film deposited on the 6H-SiC substrate were examined. Based on the ternary diagram (Figure 1), it was stated that only the Ni₂Si phase was identified between 600 °C and 950 °C. Additionally, the free carbon, as the product of SiC decomposition, occurs in the structure in the form of precipitates.

Available literature data indicated how many technological factors influence the material composition. Additionally, the type (understood as shape, morphology and particle size) of starting materials and the method of densification can affect the behavior of the material during sintering. The knowledge of the fundamental phenomena occurring in the material's structure during the densification process has a crucial role in designing new materials solutions. This is of great importance in the case of newly developed technologies, such as SPS, with high application potential.

The authors of^[23] presented the calculation of the Gibbs free energies for different possible reactions in the Ni-Si-C system. It was stated that due to the strong affinity between Si and Ni and the strong Si-C bond, the formation of nickel silicades is possible only at elevated temperatures when the thermal energy can overcome the activation barrier.

Our work seems to be the first attempt to analyze the relationship between the SPS sintering condition and microstructural changes in the Ni-SiC bulk composite. Special attention was paid to the metal-ceramic interface, responsible for the quality of composite materials. The samples were sintered by the spark plasma sintering (SPS) method, allowing for relatively quick densification in conditions of limited temperature compared to conventional sintering.^[30] In the SPS method, direct electric current flow is a heating source for sintering of powder material. Additionally, application of a uniaxial mechanical load supports the diffusion processes, allowing for material densification. Our motivation was to determine the influence of the temperature, sintering time, electric current, and mechanical load on the course of changes in the structure of the sintered Ni-SiC material. It should be emphasized that the performed investigation is the first original analysis dedicated to the manufacturing of Ni-SiC composites via the Spark Plasma Sintering method. Obtained results seem to be a good reference for researchers using SPS for the manufacturing of advanced materials.

Ni-SiC composites were prepared via the powder metallurgy route. For the starting material, nickel powder (NewMet Koch, 5 µm, 99.9 pct) and SiC powders (Saint-Gobin, 80 µm, 99.99 pct) were selected as the matrix and the reinforcement, respectively. The morphology of the starting powders is presented in Figure 2.

In order to provide a uniform distribution of elements in the powder mixture, the mixing process was conducted in a planetary ball mill (Pulverisette 6, Fritsch). A nickel-based composite mixture containing 20 vol pct silicon carbide was prepared in the tungsten carbide (WC)/cobalt (Co) container using the following mixing parameters: rotational speed ω = 100 rpm, mixing time t = 4 hours, ball-to-powder ratio = 5:1, WC/Co milling balls—10 mm in diameter. The densification process was provided in the SPS apparatus. The sintering conditions are presented in Table I. The process conditions were selected based on the preliminary sintering tests for pure nickel (Sample 7), assuming a high densification rate of the obtained samples. The system was heated at a rate of 100 °C/minute to 700 °C (800 °C in the case of Sample 6), and then the heating rate was slowed down to 50 °C until the desired sintering temperature was reached. After the holding time, the samples were naturally cooled to room temperature with an average cooling rate of 10 °C/minute. The external load was released from 50 to 0 MPa directly after the desired holding time. The diversification of the process parameters was conditioned by obtaining a different degree of interaction between the components of the composite.

The Ni-SiC composites were sintered in graphite die and punches. As a result, samples of 25 mm in diameter were obtained. As received materials were cut, ground, and polished for further investigations. The density of the Ni-SiC composites was measured using the hydrostatic method based on Archimedes’ principle. Additionally, the relative density ρ_rel was determined as the ratio of the measured bulk density ρ_m and the theoretical density ρ_t of the Ni-SiC composite. For calculations, the density of Ni and SiC was assumed to be ρ_Ni = 8.9 g/cm³ and ρ_SiC = 3.2 g/cm³, respectively.

Next, the detailed microstructural analysis including scanning electron microscopy with energy dispersive spectroscopy and transmission electron microscopy was performed. The machined surface (ground and polished) of the samples was analysed using the scanning electron microscope Auriga produced by the Zeiss company (Oberkochen, Germany). The detailed microstructural observations of the interface forming between the Ni matrix and SiC particles were carried out on a high resolution scanning transmission electron microscope (STEM) Hitachi HD-2700 operating under an accelerating voltage of 200 kV. Structural investigations were combined with advanced energy dispersive X-ray (EDX) point and mapping analyses. The selected area electron diffraction (SAED) patterns were taken using a transmission electron microscope (TEM) JEOL JEM-1200EX. Thin foils with a thickness of ∼ 100 nm for STEM and TEM investigations were prepared by the focused ion beam (FIB) lift-out technique using a FIB-SEM Hitachi NB5000 system.

The structural changes were analysed using X-ray diffraction and Raman spectroscopy. Ex-situ X-ray powder diffraction measurements were performed using the Rigaku SmartLab 3 kW diffractometer equipped with a high-speed 1D silicon strip detector D/teX Ultra 250 and filtered Cu K_α radiation. The powder diffraction patterns were measured in the θ/2θ Bragg–Brentano geometry in the angular range of 10°\(\le 2\theta \le 100\)° with a scanning step of 0.01° and a speed of 2°/minute. Qualitative analysis was performed using a database PDF-4 + 2021 and the Integrated X-ray Powder Diffraction Software PDXL2 supplied by Rigaku. Experimental data were also analyzed using the quantitative Rietveld refinement method using the PDXL2 software.

Additional information on the composition of the material inside the particles was obtained by performing Raman spectroscopy measurements with a Renishaw spectrometer using a 532 nm (2.33 eV) Nd:YAG laser and an Andor Newton CCD detector. The laser beam was focused using a lens with 100 times magnification. The system allowed spectra to be collected in the spectral range from 0 to 2800 cm⁻¹ from the areas of interest.

Due to the fact that there are almost no literature references concerning the densification of Ni-SiC composite materials via the SPS method, our original results should be discussed in relation to other available techniques used. The only paper,^[38] where authors used Spark Plasma Sintering to densify Ni-SiC is concentrated on the mechanical milling process on the reinforcement distribution in the metallic matrix. In the case of using SiC nanoparticles, no phase changes after sintering at 1150 °C for 30 minutes were reported. In the most cases, the investigations are performed for nickel-silicon thin films system.^[34,35] According to the reported data, the microstructural changes at the nickel-silicon carbide interface starts after exceeding a temperature of about 500 °C.^[39] On the direct contact area, the solid-state reaction between Ni and SiC occurs. Due to the nine silicades forming in the Ni-Si system, there is a variety of possible combinations of the final products. The decisive role in creating the final microstructure morphology and structure composition play the conditions of formation process, e.g., reaction temperature, time of annealing, atmosphere, external load, etc. Several reports have been dedicated to the analyses of microstructural and phase changes in correlation to the process parameters. The analytical investigation devoted to the possible course of reaction based on the estimation of the Gibb’s energies indicated that the most commonly formed phases are NiSi₂, NiSi, and Ni₂Si.^[37,40]

In our investigation, the SPS method has been used to manufacture bulk Ni-SiC composites. It is a relatively new method of materials densification, but it is used more and more frequently due to numerous advantages. In particular, the limitation of the possible grain growth due to the lower sintering temperature, higher heating rate, and shorter sintering time in comparison to other methods makes it more and more popular.^[36,41,42] We modified the process conditions to observe the microstructure and phase evolution of the sintered material. It should be emphasized that the relative density of the material is comparable (except for Sample 5, sintered in lower pressure assistance), which was the goal of the experiments, but the structure is significantly different. This diversity has a crucial effect on the macroscopic properties (e.g., mechanical or thermal). The observed changes were very dynamic despite the very short time in the maximum sintering temperature. The transformation process occurring in the microstructure during sintering can be divided into specific stages:

With the appearance of new Ni-SiC contact surfaces, the decomposition reaction of silicon carbide takes place rapidly [Figure 7(a)]. This early stage continues until the ceramic phase is completely consumed [Figure 7(c)]. It should be noticed that the total time of SiC decomposition last only a few minutes when the global sample temperature does not exceed 850 °C. Due to the specific nature of the SPS process, the local temperature increase in the contact points of individual grains may appear, and additionally, the flow of electric current may accelerate the diffusion of atoms. After 1 minute sintering at 850 °C, a reactive zone covering tens of microns appeared and two subphases can be distinguished (Sample 1—Figure 11). It confirms that the front of the reaction moves very quickly to the center of the SiC grains and a new structure is formed. The morphology of the as-reacted material is not uniform at this stage. At the beginning, the amorphous structure of the Ni-Si solid solution is formed due to the diffusion and incorporation of silicon atoms into the nickel crystallographic unit. This is evidenced by a slight shift of the peaks representing nickel from their nominal positions (Figure 3). At the same time, the polycrystalline structures of Ni₃₁Si₁₂ and Ni₃Si are detected by the detailed TEM analyses. It should be noticed that range of the reactive zone in Sample 1 does not exceed 10 µm, and due to this fact, it was not confirmed by the XRD results. At the interphase, small amounts of nickel are present (subphase I—Figure 11) to support the further decomposition of the rest of the silicon carbide. The appearance of the amorphous material was reported by Rao et al.^[43] where a high dose of Si ions during the implantation process of the Ni substrate caused its partial amorphization. Further annealing at temperatures exceeding 600 °C caused the crystallization of the Ni₃Si phase. As it was reported in,^[43] the good crystallinity of Ni₃Si and good alignment to the nickel substrate was observed. On the other hand α-Ni-Ni₃Si eutectic irradiated with Ni-ions and annealed at 620 °C can be transformed to the Ni₃₁Si₁₂ intermetallic phase of the hexagonal structure.^[44] In our investigation, the appearance of the amorphous structure at the beginning of interphase transformation can be caused by the current flowing across the material (similar effect to the ion implantation), but with further annealing, a fully crystalline structure is formed. Similarly, in^[45] the original Ni-Si solid solution appears, which is then transformed subsequentially to Ni₃Si, Ni₃₁Si₁₂ to finally Ni₂Si phase. This phenomenon occurred after 1 minute of annealing at 550 °C, but the range of observed changes is about 100 nm. In our case, the range of observed changes is significantly larger after 1 minute of sintering in comparison to the results presented in.^[45]

The observed silicon carbide decomposition process is very dynamic. It can be attributed to the phenomenon accompanying the SPS method. The effect of the solid mutual diffusion could be caused by the occurrence of the strong potential gradient related to the electric field. The DC power supply that was used in the SPS equipment may intensify the effects of diffusion, especially in the case of conducive materials. Rudinsky et al.^[46,47] performed the SPS sintering process for Cu-Ni to study the effect of the electric current on the progress of diffusion. They observed that there is a strong diffusion effect at temperatures over 950 °C. It was attributed to the loss of nickel ferromagnetism in the presence of the electric current at the tested temperatures. It is caused by the rearrangement of electrons in the structure of the nickel band and the increase in the conductivity band resulting from the application of the current, which causes the chemical potential repulsion between copper atoms and nickel clusters. In our case, the range of structural evolution is much greater, which can be caused by the increased movement of atoms due to the flow of the current.^[48] The applied SiC particles with a size of 40 to 80 µm completely react in the SPS process within a few minutes. Moreover, the observed speed of the structural changes in the area of the SiC grains is determined only by the diffusion rate of atoms in the contact area. In the SPS technique, the stabilization of the temperature inside the SiC particle compared to the contact area of the two particles in contact takes place in just a few milliseconds.^[49] In our investigation, the SiC particles finally disappeared with the extension of the sintering time (Sample 2 and Sample 3 with sintering times of 5 and 15 minutes, respectively). The formation of the Ni₃₁Si₁₂ compound is also in good agreement with the TEM investigation provided by Kwon et al.,^[15] where Ni₃₁Si₁₂ phase was identified as the reaction product of nickel film deposited on the 6H-SiC substrate and heat-treated at 550 °C.

The evolution of the free carbon formation process, as a product of SiC decomposition, was also observed in our investigation. Carbon does not react with silicon and nickel and does not form a stable phase with them.^[50] At the beginning of the sintering process, carbon is present in the reactive zone in a dispersed form as small single precipitates. Their size and amount increase as the SiC decomposition process progresses. The TEM analyses indicated the forms of carbon at the early stage of sintering: individual precipitates across the whole interphase area and a carbon layer at the boundary between the pure nickel and the interphase zone. It is in good agreement with the studies mentioned above,^[15] which indicated the grouping of carbon precipitates in particles, its movement to the external surface from the reaction zone, and the final formation of the thin film of graphite. The explanation of such a behavior can be found in,^[15] where the authors indicated that the location of free carbon layer at the grain boundary reduces the interfacial energy.

Finally, the described mechanisms of interfacial interaction at the nickel-silicon carbide boundary is presented in Figure 14.

The Raman spectroscopy results confirmed the presence of carbon in the form of graphite after the SPS process. Although the detailed analyses especially in the case of peak below 1400 cm⁻¹ can indicated for the presence of amorphous carbon. Similar observation are presented by Lim et al.^[40] who distinguished the presence of two forms of carbon, graphite (peak G—1580 cm⁻¹) and amorphous carbon (peak D—1355 cm⁻¹), based on the Raman spectroscopy. On the other hand, this mode may be related to the damage in the structure of the hexagons of carbon in graphite.^[51,52] With the progress of sintering, the intensity of the carbon peaks increases and the initial individual precipitates form more complex clusters (see Figures 4, 5 and 6). The TEM results confirmed their growth from an initial 100 nm (Sample 1) to even 700 nm (Sample 4). The SAED analysis indicated a partially amorphous structure, which can be related to the amorphization of carbon nanoprecipitates (see Figure 12). A similar effect has been reported by the authors of;^[15] amorphous carbon layers were also detected.

External load plays significant role for the progress of materials densification during sintering process. It impacts the sintering mechanisms by intensifying the diffusion process and increasing the progress of sintering. Additionally, the external load allows for the easier regrouping of particles at the beginning and effects on speed of pore elimination processes in the final stage of sintering. In the case of Sample 5, where the load was at a low level (5 MPa), the residual porosity was about 20 pct. It was generally localized between the nickel particles in the area of concentrated SiC particles. Although the contact areas between adjacent grains were limited, almost all SiC particles were decomposed. The Ni-Si structure was not uniform, with visible precipitates of carbon appearing accidentally across the reactive zone.

During materials densification via the SPS method, many independent factors influence the sintering mechanisms. Besides the temperature, time, and external pressure, the flowing current can directly influence on diffusion processes. With the progress of sintering, the number of intergrain contacts increases, accelerating SiC decomposition and finally resulting in the ordering of the system within the silicade area. With the increasing sintering temperature and time, more thermodynamically stable phases appeared. In our case, the dominant phase was Ni₃Si (Sample 6). Lim et al.^[40] revealed the transformation from the Ni₃Si₂ phase to only the Ni₂Si compound after 2 hours of annealing at 1050 °C. Local melting areas were identified after 6 hours heat treatment at 1100 °C, where Ni₃Si and Ni₅Si₂ phases were identified as the reaction products.^[53,54] In contrast, only the presence of Ni₂Si was noticed for Ni/4H-SiC wafers after annealing at 700 to 1000 °C.^[55] In the higher range of temperatures (1245 °C), Ni₂Si and Ni₅Si₂ were present in the structure after 60 minutes annealing time.^[36,56] In our case, there was no evidence of the presence of a liquid phase, even in the case of an increased sintering temperature of up to 1000 °C (Sample 6). The final product composition is dependent on the amount of the decomposed silicon. In our case, 20 vol pct of silicon carbide allowed for the identification of two Ni₃₁Si₁₂ and Ni₃Si phases. It should be emphasized that all available silicon atoms have been bonded in the silicade phases, and the residual nickel remained in the materials in an unchanged form (Ni-fcc).

Based on provided investigation the following conclusion can be draw: