Abstract
The new type of Mg–8Al–1Sn (AT81) magnesium matrix composites reinforced with different volume fractions (5, 10, 15, 20, 25 and 30 vol.-%) of SiC particles (average size of 10 μm) was fabricated by powder metallurgy. With the increasing volume fraction of SiC particles (SiCp), the particles gradually show more homogeneous distribution. Compared with the AT81 alloy, the yield strength (YS) and ultimate compressive strength of the SiCp/AT81 composites are improved simultaneously. With the increasing SiCp from 0 to 30 vol.-%, the YS and ultimate compressive strength increase from 69 to 239 MPa and 286 to 385 MPa respectively, while the corresponding fracture strain (ε) decreases from 19·3 to 4·8%. The improvement of the YS and ultimate compressive strength of the SiCp/AT81 composites benefits from the more homogeneous microstructure due to the increase in the SiC particles.
Introduction
The interest in developing SiC particles (SiCp) reinforced magnesium metal matrix composites (MMCs) has increased substantially in recent years.1–3 The SiCp reinforced magnesium MMCs exhibit more superiorities over monolithic magnesium and magnesium alloys, such as high elastic modulus, high strength, superior creep and wear resistances at elevated temperatures, and thus numerous studies have been made on SiCp reinforced magnesium MMCs. 4 Pure Mg, Mg–Al–Zn, Mg–Al–Si, Mg–Zn–Zr, Mg–Zn–Ca and Mg–Zn–Y systems are the most commonly used matrix materials in magnesium MMCs.5–10 Moreover, the addition of Sn to Mg alloys has been given much attention.11–13 The room and elevated temperature mechanical properties can be improved with the appropriate addition of Sn to Mg alloys. 14 The Mg–Al–Sn system shows better strength and elongation.15,16 Therefore, the Mg–Al–Sn system also has great potential as matrix in magnesium MMCs.
Particles reinforced magnesium MMCs had been produced by a variety of methods, such as stir casting, 1 mechanical alloying, 17 squeeze cast, 18 powder metallurgy,19,20 etc. In the powder metallurgy process, the MMCs are fabricated from powders without passing through a fully melting state and can attain more uniform distribution of particulates in the matrix without or with less excessive reactions between the matrix and reinforcement. 21 However, SiCp reinforced magnesium MMCs manufactured by powder metallurgy have not been investigated thoroughly. 22 Therefore, this work focuses primarily on the microstructure and compressive properties of a new type of Mg–8Al–1Sn magnesium MMCs with different volume fractions (5, 10, 15, 20, 25 and 30 vol.-%) of SiCp fabricated by powder metallurgy.
Experimental procedure
In this study, the starting materials were made from commercial powders of magnesium (>99·9 wt-% purity, ∼74 μm), aluminium (≥98·0 wt-% purity, ∼29 μm), tin (≥99·0 wt-% purity, ∼38 μm) and silicon carbide (≥99·0 wt-% purity, ∼10 μm). Elemental powder blends corresponding to nominal Mg–8Al–1Sn (wt-%) alloy and SiCp/AT81 composites with different volume fractions of SiC particles (5, 10, 15, 20, 25 and 30 vol.-%) were mixed sufficiently at a rotational velocity of 50 r min−1 for 7–8 h in a ball mill (Model: PMQW2L, China) under the protection of high purity argon gas with the mass ratio of ball to powder of about 10:1. No lubricant was used during ball-milling process. Then, the blends were cold pressed into cylindrical compacts (30 mm in diameter and 45 mm in height) at pressures ranging from 70 to 75 MPa to obtain densities of 70–80% theoretical density by using a stainless steel die at room temperature in air. After that the cylindrical compacts were heated to 753K and held at this temperature for 1·5 h in a self-made vacuum hot pressing and sintering furnace (vacuum degree: 0·01 MPa). The cylindrical compacts were subsequently pressed under 45 MPa for 10 min and cooled down to the ambient temperature.
The microstructures were characterised by a scanning electron microscope (SEM, ZEISS EVO18, Germany) equipped with an energy dispersive spectrometer (EDS) analyser (INCA-X-Max, UK). The phase constituents of the alloys were examined by X-ray diffraction (XRD, Model D/Max 2500PC Rigaku, Japan) with Cu Kα radiation at the voltage of 40 kV at a scanning speed of 4° min−1. The cylindrical specimens with a diameter of 4 mm and a height of 8 mm were used for compressive tests. The uniaxial compressive tests were carried out on a servohydraulic material testing system (MTS, MTS 810, USA) at a constant strain rate of 1·0×10−1 s−1. At least two samples were tested and the average values are presented. Samples for microstructure observations were firstly ground with 2000 mesh SiC papers, followed by buffing with 0·5 μm diamond pastes, and then chemically etched in acetic picral solution (1 g picric acid, 1 mL acetic acid, 2 mL distilled water, 15 mL ethanol) for about 20 s.
Results and discussion
Table 1 shows the theoretic density, actual density and porosity of AT81 alloy and 5–30 vol.-% SiCp/AT81 composites, respectively. The theoretic density and actual density gradually increase with the increasing SiC particles. It is indicated that 1·2% porosity exists in the AT81 alloy (Table 1). No porosity exists in the 5–25 vol.-% SiCp/AT81 composites and only 0·2% porosity exists in the 30 vol.-% SiCp/AT81 composite, which shows that SiCp/AT81 composites are much dense (Table 1). The porosity can be influenced by the solid diffusion and the compaction load used in the hot pressed treatment. 21 Moreover, it is noteworthy that the actual densities are greater than the theoretic ones in some composites (5, 15 and 20 vol.-% SiCp/AT81), which may be attributed to the volatilisation of a small amount of magnesium during the fabrication progress and therefore, the actual content of SiC particles is greater than the theoretic one.
Theoretical densities, actual densities and porosities of AT81 alloy and composites reinforced with different volume fractions of SiCp
The SEM microstructures of the SiCp/AT81 composites with different volume fractions of SiC particles are shown in Fig. 1. As shown in Fig. 1a, most of the SiC particles are distributed at grain boundaries in the 5 vol.-%SiCp/AT81 composite, which was also observed in SiCp/AZ91 composites. 23 With the increasing SiC content, the SiC particles increase and the distribution of the particles becomes more homogeneous (Fig. 1b–f). Figure 2a shows high magnification SEM micrograph of the 10 vol.-% SiCp/AT81 composite. The result of EDS line scan analysis shows that a small amount of oxygen is present in the matrix and interfacial reaction may occur between the SiC particle and the matrix (Fig. 2b). To prove the interfacial reaction further, the XRD analysis is provided in Fig. 3. The XRD result shows that some MgO and Mg2Si phases are also formed in the composites. Note that Mg has a high tendency to form oxide layers on its surface during the fabrication process. The Mg2Si phase comes from the interfacial reaction between the matrix and the SiCp, which has been reported by Bochenek and Braszczyńska 24 in SiCp reinforced Mg–Al composite.

Images (SEM) of a 5 vol.-%SiCp/AT81, b 10 vol.-%SiCp/AT81, c 15 vol.-%SiCp/AT81, d 20 vol.-%SiCp/AT81, e 25 vol.-%SiCp/AT81 and f 30 vol.-%SiCp/AT81

a high magnification SEM micrograph of 10 vol.-%SiCp/AT81 composite and b EDS line scan along line in a

Patterns (XRD) of a AT81 alloy and b 20 vol.-%SiCp/AT81 composite
The engineering stress–strain curves of the AT81 alloy and SiCp/AT81 composites are plotted in Fig. 4 and average engineering compressive properties are presented in Table 2. The SiCp/AT81 composites show higher strength than the AT81 alloy. With the increasing SiC particles from 0 to 30 vol.-%, the yield strength (YS) and ultimate compressive strength increase (UCS) from 69 to 239 MPa and 286 to 385 MPa, respectively, while the corresponding fracture strain (ε) decreases from 19·3 to 4·8%. The increase in particle contents could induce high dislocation density due to the large difference coefficient of thermal expansion between particles and Mg matrix, which leads to the improvement of the YS and ultimate compressive strength.25,26 It has also been reported that particle distribution significantly affected the mechanical properties of the composites and the uniform particle distribution often resulted in good mechanical performance. 25 The applied load can effectively transfer to the uniformly distributed strong particles. 26 Therefore, the more homogeneous distribution of the SiC particles is more beneficial to the enhancement of the YS and ultimate compressive strength of the SiCp/AT81 composites with the increasing SiC particle contents.

Engineering stress–strain curves with different SiC contents
Yield strength (YS), ultimate compressive strength (UCS) and fracture strain (ε) of AT81 alloy and composites reinforced with different volume fractions of SiCp
Figure 5 shows the SEM fractographs of AT81 alloy, 15 and 30 vol.-%SiCp/AT81 composites, respectively. It can be seen that the AT81 alloy and the composites present typical cleavage fracture, which indicates brittle fracture.2,27 With the increase of SiC particle contents, the fracture microstructures become more homogeneous in Fig. 5b and c, leading to the enhancement of the yield and compressive strength of the SiCp/AT81 composites.

Fractographs (SEM) of a AT81 alloy, b 15 vol.-%SiCp/AT81 composite and c 30 vol.-%SiCp/AT81 composite
Conclusions
The Mg–8Al–1Sn magnesium matrix composites with different volume fractions (5, 10, 15, 20, 25 and 30 vol.-%) of SiC particles are successfully synthesised by powder metallurgy. A little interfacial reaction occurs between the SiC particles and the matrix. The distribution of the particles becomes more homogeneous with the increasing SiC particle content. With the increasing SiCp from 0 to 30 vol.-%, the YS and ultimate compressive strength increase from 69 to 239 MPa and 286 to 385 MPa respectively, while the corresponding fracture strain (ε) decreases from 19·3 to 4·8%. The more homogeneous distribution of the SiC particles is more beneficial to the enhancement of the YS and ultimate compressive strength of the SiCp/AT81 composites with the increasing SiC particle contents.
Footnotes
Acknowledgements
Financial supports from the Natural Science Foundation of China (NSFC) (Grant No. 51271086 and No. 51301074) and The Doctoral Fund of Ministry of Education of China (Grant No. 20120061110031) are greatly acknowledged. Partial financial supports come from The Natural Science Foundation of Jilin Province (Grant No. 201115010) and The National Training Programs of Innovation and Entrepreneurship for Undergraduates (Grant No. 2013A43127).
