Effect of Matrix Particle Size on Densification Behavior, Microstructure, and Mechanical Properties of an Al/FMG/SiC Hybrid Composite

A study was conducted to examine the impact of matrix particle size on the density, microstructure, and mechanical properties of hybrid aluminum matrix composites reinforced with Fe-based metallic glass (FMG) and SiC particles. The composites were manufactured using the spark plasma sintering (SPS) process. The results showed that increasing matrix particle size led to an increase in porosity content. The distribution of reinforcing particles in the matrix was found to be more uniform when smaller particles were used. However, the phase properties and stored dislocation density remained unchanged with varying matrix particle sizes. On the other hand, the yield compressive strength and ductility of the composites significantly decreased with increasing particle size. The yield strength of a composite with a matrix particle size of 20 μm was approximately 80% higher than that of a composite with a matrix particle size of 63 μm. Therefore, for the synthesis of Al/FMG/SiC hybrid composites, the optimal outcome for different mechanical properties was observed with a lower matrix particle size.


Introduction
Aluminum matrix composites (AMCs) that are discontinuously reinforced with hard particles have gained widespread use in various industrial applications due to their isotropic properties, low density, and superior wear and corrosion resistance [1][2][3][4][5]. Typically, AMCs are reinforced with a single type of particle from a specific class of engineering materials, such as ceramics, polymers, carbon-based materials, or metallic glasses. However, recent attention has been focused on hybrid AMCs, in which more than one type of particle reinforces the matrix simultaneously [6][7][8][9]. In these composites, each reinforcement type contributes its own positive effects to the properties of the composite, leading to a potentially unprecedented amount of reinforcement due to synergistic effects. Over the past few decades, various types of hybrid AMCs have been developed, including those reinforced with two different ceramic particles (e.g. A/SiC/ B4C [10]), those reinforced with ceramic and carbon-based materials (e.g. Al/SiC/graphite composites [11][12][13]), and those reinforced with ceramic and metallic materials (e.g. Al/SiC/FMG [14] and Al/Cr3c2/Ni-Cr [15]). For example, research has shown that the simultaneous presence of SiC and FMG in an Al matrix can improve both its strength and ductility [7]. Metallic glasses are highly suitable for reinforcing metal matrix composites, owing to their metallic properties [16,17]. As a result, their combination with ceramic or carbon-based reinforcement particles has the potential to enhance the mechanical properties of the composite. Another study found that the wear properties of an Al matrix in a hybrid AMC were significantly improved due to the increased hardness from steel particles and the lubricating effect of graphite reinforcements [18]. Numerous studies have investigated the physical and mechanical properties of hybrid composites. The impact of various parameters, such as the type and size of reinforcements, the interfacial conditions between the matrix and reinforcements, and the addition of special elements to the matrix, as well as processing procedures such as heat treatment and shot peening, on these properties have been examined. Studies that investigated the effects of reinforcement particle size have 1 3 consistently indicated that reducing the size of reinforcement particles (by using nanoparticles) leads to an increase in the strength and hardness of the hybrid composite. For example, Chang et al. [19] found that microhardness was enhanced in an (Al4SiC4 + SiC) hybrid-reinforced Al composite as the particle size of SiC decreased. Lemine et al. [20] found that using more nanoparticles of B4C as reinforcing particles in an (NBC + B4C)/Al composite led to a remarkable improvement in compressive ultimate strength due to the activation of the Orowan strengthening mechanism. The quality of the interface between the reinforcements and the matrix plays a crucial role in determining the extent of load transfer to the reinforcements under external stresses [21]. Various studies have shown that heat treatments, such as age-hardening, can improve the mechanical properties of hybrid composites. For example, Liu et al. [22] found a remarkable increase in tensile strength in a (SiCp + Ti)/7075Al hybrid composite due to precipitation hardening of the matrix alloy resulting from aging heat treatment. Another study [23] found that the formation of Mg2Si precipitates as a result of artificial aging treatment in a (GNPS + B4C)/6063Al hybrid composite positively influenced the tensile strength of the composite.
However, the impact of matrix particle size on the microstructure and mechanical properties of hybrid AMCs has not been thoroughly researched and to date, research in this field has primarily concentrated on composites containing a single reinforcing particle. Therefore, it is necessary to explore the impact of matrix particle size in composites that feature hybrid reinforcements. This aspect is crucial to consider when manufacturing hybrid AMCs using powder metallurgy, as the matrix particle size can affect the densification process and distribution of reinforcement particles in the composite during consolidation. To address this knowledge gap, the present study aims to investigate the effect of matrix particle size on the microstructural, phase, and mechanical properties of a hybrid Al/FMG/SiC composite prepared using powder metallurgy.

Experimental Procedure
The study reported in this paper builds upon previous research [24] in which hybrid aluminum matrix composites reinforced with FMG and SiC particles with volume fractions of 3% and 7% respectively were developed successfully. The current study aims to examine the impact of matrix particle size on the properties of these composites. The matrix material used was pure elemental aluminum powder with average particle sizes of 20 μm, 45 μm, and 63 μm. These matrix materials are referred to as HC20, HC45, and HC63 in the study, respectively (Fig.1).
The FMG particles used in the composite had a composition of Fe75Si15B5Zr5 and were produced through mechanical alloying via high-energy planetary ball milling of Fe, Si, B, and Zr elemental powders. The fabrication process for these FMG particles is described in detail in previous work [24]. To create the hybrid composite powders, the appropriate amounts of pure aluminum and reinforcement particles (SiC and FMG) were mixed in a horizontal mixer at a low energy input of 100 rpm for 2 h. Figure 2 shows FESEM micrographs of the FMG and SiC particles, which have average sizes of 16 μm and 18 μm, respectively, and are close to each other in size. The average size of all the particles was determined through analysis with Clemex image analyzing software.
The composite powders were placed in a graphite die with an inner diameter of 25 mm and underwent sintering in a Nanozint 10i SPS setup. The process was conducted at 550 °C for 10 min with a pressure of 40 MPa and a heating rate of 50 °C/min, in a vacuum environment. Pure Al powders of similar particle size were also sintered under the same conditions and were labeled as P20, P45, and P63 for comparison.
The bulk density of the sintered samples was determined through the application of Archimedes' principle. Microstructure was analyzed using an FE-SEM (ZEISS Sigma 300) and an EDS device. The samples were mechanically polished, etched with Keller reagent, and characterized in terms of phase composition via X-ray diffraction (XRD) using a diffractometer with a Cu anode, scanned over a range of 20°-90°.
The stored dislocation density in the matrix of the composites was calculated using the Williamson-Hall equation, which is based on XRD patterns [25,26]. This equation allows for the determination of both the crystallite size (D) and the lattice microstrain (ε): where B is the full width at half maximum (FWHM) of a peak, θ is the diffraction angle, K is a constant, and λ is the wavelength of the X-rays. The expression B cos θ/λ was plotted along the Y-axis and 2 sin θ/λ was plotted along the X-axis for the crystalline planes (111), (200), (220), and (311). A linear fit of the data was then performed, and the microstrain and crystalline size were estimated from the slope and y-intercept. The stored dislocation density was calculated using the following equation, incorporating the calculated microstrain and crystalline size [27]: where ρ is the dislocation density and b is the Burgers vector.
Compressive strength was determined on specimens with a height of 12 mm and a diameter of 8 mm, according to the ASTM E9 standard. Three tests were performed for each sample. The hardness of the sintered specimens was measured using the Brinell scale, with an applied load of 31.25 kg and a 2.5 mm indenter. The results of each test were the average of five consecutive indentations, with standard deviations included. Figure 3 shows the porosity content of the composite and pure Al specimens. As expected, the composite specimens had a higher porosity content compared to the pure Al specimens. This is because the reinforcing particles interfere with the rearrangement of the powder particles and the contact between neighboring matrix particles during the initial phase of the sintering process, leading to decreased consolidation. The figure shows that both the composite and pure Al samples had an increase in porosity with an increase in matrix particle size.

Porosity Content
The densification of pure Al during the SPS process has been described in the literature [28,29] as occurring in three steps. These steps include rearrangement of the powder particles, local plastic deformation and creation of the sintering neck, and extensive sintering through mass transport. The particle size can significantly impact almost all of these steps. In the first step, a reduction in particle size increases the number of contact points per unit volume, making it easier for the particles to rearrange. In the second step, the effect of particle size on the formation of the sintering neck and local plastic deformation needs to be evaluated.
The tensile strength of the sintering neck is a key indicator of its extension between adjacent particles. The following equations can be used to calculate the tensile strength of the sintering neck as a function of particle size [29]: where R' is the equivalent radius of curvature, R1 and R2 are the radii of the contacted particles, is the tensile stress of the sintering neck, and is the surface tension. The ratio of R' for pure Al particles with average sizes of 20 μm, 45 μm, and 63 μm is approximately 1:2:3. As a result, the tensile strength ratio is estimated to be 3:2:1, meaning that the tensile stress of the sintering neck between small particles is three times higher than that between large particles. This higher tensile strength of the small particles is believed to contribute to their lower porosity content, as the growth and expansion of the sintering neck shrink the pores. During sintering in the SPS process, the application of an electric field enhances mass transport through Joule heating [30]. Thus, the influence of particle size on Joule heating should be evaluated in the third step of sintering. According to the Joule equation: The amount of heat generated (Q) is directly proportional to the square of the current (I), resistance (R), and time. Larger particles result in a greater pore volume between particles, which decreases the cross-sectional area available for the flow of electrons (i.e., current). Consequently, the amount of heat generated is lower in larger particles than in smaller ones. This suggests that a decrease in matrix particle size, except for the first step, leads to a lower porosity content.
For composite specimens, the ratio of matrix to reinforcement particle size (PSR) is a crucial factor affecting the densification behavior during sintering. Research has shown that a lower PSR value, particularly close to unity, results in a higher density of the composite specimens [31]. As the grain size of the matrix increases, the PSR for the reinforcing particles (both SiC and metallic glass particles) and the matrix decreases.
The porosity content of HC20 and HC45 samples only differs slightly, but a noticeable increase in porosity content is observed in the HC63 sample. Given the size of the matrix and reinforcing particles, the PSR for the HC20 and HC45 specimens falls within the optimal range of 1/3 to 3 [32].  (Figs. 4a and b), while the HC63 sample displays a heterogeneous distribution of SiC and FMG particles in the matrix. Additionally, some microscopic voids are visible at the interface between particles and the matrix.

Distribution of the Reinforcing Particles
Previous studies [31] have shown that the size difference between the reinforcing particles and the matrix plays a critical role in particle distribution within the matrix. When the size of the reinforcing particles is significantly smaller than the size of the matrix powder, the particles tend to clump together within the matrix [33].  Figure 4 also highlights that agglomeration and the formation of voids at the interfaces are more likely to occur in SiC particles for all samples. This can be attributed to the metallic characteristics of FMG particles, which exhibit better interfacial properties compared to SiC particles, resulting in stronger FMG/matrix bonds that allow for a more efficient distribution of FMG particles within the metallic matrix of the composites. Prior studies in the field of metallic glass reinforced composites have demonstrated the high quality of the interface between the glass particles and the matrix [17,34,35]. A quantitative evaluation of particle distribution can be conducted using the inter-particle distance parameter (λinter), which represents the average distance of an individual particle to its nearest neighbor. To measure the degree of distribution homogeneity, a statistical parameter known as the coefficient of variation (COV) [7] was employed. This parameter is defined as the ratio of the standard deviation (σm) to the mean (m). COV values can range between  Figure 5 presents the COV values of different composite samples, calculated from the corresponding FESEM images (Fig. 4). Figure 5 highlights that the HC20 sample exhibits the most favorable particle distribution, as evidenced by the lowest COV value. This finding is in line with the qualitative analysis of FESEM images.
The X-ray diffraction (XRD) patterns of the composite samples are presented in Fig. 6. The XRD pattern shows sharp Bragg peaks that correspond to aluminum and silicon carbide (SiC) (Fig. 6(a)). Figure 6(b) displays a broad halo at 2θ = 44° in the XRD patterns, which is characteristic of the amorphous structure of the Fe-based metallic glass (FMG) particles. No peaks corresponding to interfacial products are detected in the XRD patterns of the composite samples. There is also no significant difference in terms of the intensity and width of the crystalline peaks or their positions. This indicates that the dislocation densities stored in the matrix of all composite samples are similar to each other (Fig. 7).
The stored dislocation density in the matrix of the composite upon cooling from the processing temperature is influenced by the type and size of the reinforcing particles, which indirectly determine the type of matrix/interface region [36][37][38]. As a result, when the same reinforcement conditions are applied to all composite samples, the stored dislocation density in the matrix is almost the same. The patterns observed in the samples show sharp Bragg peaks with similar intensity belonging to aluminum and SiC, and a broad halo for FMG particles around 2θ = 44°. Additionally, there were no peaks indicating the presence of interfacial products in the patterns of the sintered samples. This can be attributed to the brief contact period between the reinforcing powder particles and the matrix during the Spark Plasma Sintering (SPS) process. In fact, the SPS method offers a distinct advantage over other sintering techniques, as it utilizes faster heating rates, which result in shorter holding times for the production of sintered parts with full or nearly full density [39,40].

Mechanical Properties
The results of the Brinell hardness tests on the various samples are presented in Fig. 8. As expected, the hybrid The findings also reveal that the hardness of the pure aluminum samples decreases with an increase in the size of the matrix powder particles. The hardness of sample P63 is significantly lower than the other two pure aluminum samples. Among the composite specimens, the hardness values of HC20 and HC45 are comparable; however, the hardness of HC63 is significantly lower than the other two composite specimens. Figures 9 and 10 depict the typical stress-strain relationships under compressive mode for the pure aluminum and composite samples, respectively. The corresponding data for the stress-strain curves are presented in Table 1. For the pure aluminum specimens, the trend in the hardness, yield strength, and compressive ductility values is similar and is positively impacted by the reduction in the particle size of the matrix powder.
The slope of the stress-strain curves for all composite specimens after yielding is similar, indicating that they have the same strain-hardening capacity. There is a positive correlation between the strain hardening rate and the dislocation density of metallic materials [41], which suggests that this behavior is a result of the small difference in stored dislocations after sintering and before the compression test (as shown in Fig. 7). The yield compressive strength and ductility for HC63 were significantly lower, by about 41% and 19%, compared to HC20. Since all composite specimens have similar composition and phase characteristics (as shown in Fig. 6), these differences in mechanical properties can only be attributed to two factors:

Porosity Content
The significantly higher porosity content of HC63 compared to the other two composite specimens (as shown in Fig. 3) provides a preferential path for crack propagation under load, leading to a decrease in the ductility of the composite [42]. This porosity also negatively impacts the mechanism of load transfer to the FMG and SiC particles, leading to a decrease in the strength and hardness of the composite specimen.

Distribution of the Reinforcing Particles
The non-uniform distribution of the reinforcing particles in the matrix composite is the second factor that affects the mechanical properties of HC63. This non-uniform distribution reduces the average spacing between adjacent reinforcing elements, promoting the growth of interfacial nucleated cracks and increasing the risk of fracture.

Conclusions
In this study, the impact of matrix particle size on the densification behavior, microstructure, and mechanical properties of Fe-based metallic glass (FMG) particles and SiC particle composites was examined. The primary conclusions that can be drawn are as follows: 1. Larger matrix particle size resulted in a more heterogeneous distribution of reinforcing particles and higher porosity content in the composite matrix. 2. XRD analysis showed that the characteristics of the constituent phases remained unchanged with changes in the matrix particle size. 3. The stored dislocation density was not affected by the matrix particle size. 4. The hardness, yield strength, and strain to fracture of the composite specimen with the smallest matrix particle size (HC20) were improved by 5 BHN, 41 MPa, and 11, respectively, compared to the composite specimen with the largest matrix particle size (HC63).