Effect of Silicon Addition on Microstructure, Hardness and Young’s Modulus of Injection Molded AZ91D Alloy

Abstract

Magnesium chips were coated with silicon powder using a binder, and injection molding was performed using silicon-coated chips as the raw material. Microstructural analysis of the products revealed that all added silicon (Si) precipitated as Mg₂Si particles, with no voids present at the interfaces between the Mg₂Si particles and the matrix. Furthermore, the hardness and Young’s modulus of the products increased with higher Si content. The Young’s moduli followed the rules of mixtures.

 

This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 87 (2023) 186–191. Abstract and keywords were slightly modified. The captions of Fig. 3, Fig. 4, and Fig. 6 are slightly modified.

Fig. 8 Effect of Si addition on Young’s modulus with the rules of mixtures (Parallel model, Series model and Mori-Tanaka model).

1. Introduction

Concerns due to increasing environmental damage have led to an increased demand for weight reduction in various components. In particular, magnesium alloys are becoming increasingly prominent as structural materials as they have the lowest density among all commercially available metals, coupled with their high recyclability and resource abundance.^1–4) Among these, the AZ91D alloy, (Mg–9 mass% Al–1 mass% Zn alloy) is a popular choice for commercial applications. This alloy is favored for its relatively high strength and commendable corrosion resistance. While AZ91D alloys are typically molded using the die casting method, the use of the injection molding technique for these alloys has gained interest. This method does not require flame-proof gas and facilitates the production of thin-walled, near-net molding.

The ratio of the strength or Young’s modulus of a structural material to its density plays an important role in reducing its weight. Notably, magnesium alloys display a lower Young’s modulus compared to other metals. Consequently, designs that increase the cross-sectional area become necessary to compensate for this bending stiffness. Therefore, it is imperative to improve the Young’s modulus of magnesium alloys in order to reduce the size and weight of structural components.

Several magnesium-based composites, which incorporate highly rigid second-phase dispersed particles, have been explored for improving the Young’s modulus of magnesium alloys.⁵⁾ Among these, the intermetallic compound Mg₂Si is a promising second-phase particle for strengthening composite materials. Its notable properties include a hardness of 600–700 HV,⁶⁾ Young’s modulus of 108 GPa,⁷⁾ and density of approximately 1.99 g/cm³. Magnesium alloys dispersed with Mg₂Si are produced through various methods such as casting, molten metal impregnation using preforms, and solid-phase synthesis via powder metallurgy. These methods lead to the production of what is known as Mg₂Si-dispersed magnesium-based composite material.^6,8–19) However, there have been no reports on improving the Young’s modulus of these composited through the injection molding method. The use of the injection molding technique for producing Mg₂Si-dispersed magnesium-based composites could offer numerous advantages over existing production methods, paving the way for broader industrial applications of these composites.

The injection molding method that uses graphite-dispersed magnesium-based composites has been previously reported.²⁰⁾ This method involves coating the raw material of the composite with magnesium alloy chips and graphite powder using a binder. Despite graphite’s poor wettability with molten magnesium,²¹⁾ the resultant composites exhibit a good interface between the matrix and graphite.

The previously developed technique²⁰⁾ can be applied to any powder particles for manufacturing high-performance magnesium-based composites. In this study, we fabricated Mg₂Si-dispersed magnesium-based composites to enhance the stiffness of magnesium alloys. Our method included coating AZ91D magnesium alloy chips with silicon powder and then injection molding using these coated magnesium alloy chips as the raw material. We anticipated that Mg₂Si would disperse within the matrix through an in-situ reaction of silicon powder and molten magnesium in the molding machine’s cylinder. To evaluate the impact of silicon addition on the Young’s modulus, we conducted microstructural analyses, hardness tests, and tensile tests on the injection molded products.

2. Experimental Method

2.1 Preparation of injection molded products

The experiments were performed using AZ91D magnesium alloy chips (dimensions: 0.5 × 1 × 4 mm, supplied by Nippon Material) with chemical compositions listed in Table 1. The silicon used was sourced from silicon powder (average particle size: 5 µm and 16 µm, purity: 3N, from Kojundo Chemical Laboratory). We used paraffin wax as a binder for coating the chips with silicon powder. The coating process involved mixing the wax dissolved in isopropyl alcohol, with the AZ91D magnesium alloy chips and silicon powder in a container, followed by applying heat and agitation.²⁰⁾ The wax was removed by heating the coated chips at 673 K under atmospheric conditions. To quantify the silicon, some of the coated chips were ultrasonically cleaned in acetone to remove the silicon powder, and the mass difference before and after cleaning was measured. The chips coated with silicon powder, having an average grain size of 5 µm, contained silicon addition of 20 mass%, whereas those with an average grain size of 16 µm had 7.0 mass%. For injection molding, the chips coated with an average silicon particle size of 5 µm were mixed with uncoated AZ91D chips, adjusting the silicon content to 1.8, 3.6, 7.8, 9.0, 12, 15, and 20 mass%. These samples were used as raw materials for injection molding. Meanwhile, the chips coated with an average silicon particle size of 16 µm were used directly as the raw material for injection molding.

Table 1 Chemical composition of AZ91D alloy (mass%).

For the injection molding process, specifically to mold a flat plate of 2.5 mm thickness as shown in Fig. 1,²⁰⁾ we used a magnesium injection molding machine (JLM75MG, Japan Steel Works). The temperature of the cylinder was set to a range of the liquidus temperature with a tolerance of 0–20 K. This setting was based on the liquidus temperature from the calculated phase diagram. Under these conditions, the solid phase ratio of the molded product was assumed to be less than 5%. The molding process was performed at an injection speed of 1.0 m/s and a mold temperature of 483 K.²²⁾

Fig. 1

Schematic drawing of the products and tensile specimens. Gray dashed line and gray arrow represent polished surface for microscopic observation and observation direction, respectively.²⁰⁾

2.2 Evaluation of injection molded products

To examine the cross-sectional microstructure of the injection molded products, small pieces were cut from the center of the products and examined using an optical microscope (GX-53, Olympus) and a scanning electron microscope (SEM). Prior to microscopic observations, specimens were embedded in phenol resin, and their surfaces were polished with water-resistant abrasive paper (up to #1200) followed by diamond paste with grain diameters of 6, 3, and 1 µm. We conducted particle analysis of the optical microscopic images of the molded products using image analysis software (Stream, Olympus). This analysis was performed at 500× magnification to distinguish individual particles and eliminate casting porosities as much as possible. Additionally, some samples underwent scanning transmission electron microscopy - energy dispersive X-ray spectrometry (STEM-EDS) analysis using STEM equipment (Talos F200X TEM, Thermo Fisher Scientific). For this, thin samples were cut from the products using a focused ion beam (DB235, FEI). Observations were carried out at an acceleration voltage of 200 kV. We obtained high-angle annular dark-field STEM (HAADF-STEM), bright-field STEM (BF-STEM) images and elemental maps that contained compositional information. The phases present in the products were identified using X-ray diffraction (XRD). The products were polished to the thickness of approximately 1 mm from the surface, and XRD measurements were conducted on these polished surfaces. The XRD analysis utilized an X-ray diffractometer (X’Pert-MRD, Philips) with a Cu tube (CuKα, λ = 1.54056 Å), a tube current of 60 mA, a tube voltage of 40 kV, a scanning speed of 0.01°/s, and a diffraction angle 2θ ranging from 20° to 50°.

We measured the Vickers hardness at a position corresponding to the center of the thickness of the molded part, the same location as for the optical microscopy observations. These measurements were conducted using a Vickers hardness tester (FLV-50ARS-F, Future Tech) under a load of 49 N and a holding time of 10 s. For each sample, we took measurements at five different points and obtained their average to represent the hardness value.

For tensile tests, we cut the specimens from a specific area of the products, as illustrated in Fig. 1. These specimens were equipped with a strain gauge (KFEL-5-120-C1, Kyowa Dengyo) and tested using a tensile testing machine (AG-50kN, Shimadzu) at a tensile speed of 0.008 mm/s. We determined Young’s modulus based on from the stress-strain curves obtained from these tests.

3. Results and Discussion

3.1 Effect of silicon addition on microstructure of injection molded products

Optical micrographs depicting the cross sections of the molded product are presented in Fig. 2. These images show the AZ91D magnesium alloy with varying silicon additions: 0 mass% (pure AZ91D), 1.8 mass%, 3.6 mass%, 7.8 mass%, 9.0 mass%, 12 mass%, 15 mass%, and 20 mass% Si (average particle size: 5 µm), as well as 7.0 mass% Si (average particle size: 16 µm). The small black spots in the cross-sectional microstructure of the AZ91D magnesium alloy (Fig. 2(a)) are casting porosities, which are attributed to solidification shrinkage. These casting porosities appeared with similar sizes and frequencies across all the molded products, irrespective of the silicon content, average grain size of the silicon powder, and the total amount of silicon added. In the molded product, containing 1.8 mass% Si (average particle size: 5 µm), we observed polygonal second phases with diameters ranging from a few µm to a dozen µm, exhibiting gray contrasts, along with very fine lineation (Fig. 2(b)). Tang et al. conducted a study on the as-cast microstructure of Mg–4%Si alloys and identified the polygonal phase as primary Mg₂Si and the fine linear phase as eutectic Mg₂Si.²³⁾ Our results showed that fine linear eutectic Mg₂Si decreased as the amount of silicon added increased. This trend was marginally observed in the 3.6 mass% Si (average particle size: 5 µm) added moldings and absent in the 7.0 mass% Si (average particle size: 16 µm, Fig. 2(c)) and in the 7.8 mass% Si (average particle size: 5 µm, Fig. 2(e)) added moldings. Conversely, the quantity of polygonal primary Mg₂Si particles increased with the addition of Si. Figures 3 and 4 display the median diameter and total area fraction (%) of the primary Mg₂Si particles, as determined by particle analysis of the observed surface. The median diameter of Mg₂Si particles tended to slightly decrease with larger amounts of silicon added, and it stabilized in the range of approximately 2 to 3 µm. Notably, the median diameter of the primary Mg₂Si particles in the molding with 7.0 mass% Si (average particle size: 16 µm) was similar to that in the moldings using silicon powder with the average particle size of 5 µm. Meanwhile, the total area fraction of Mg₂Si particles grew proportionally with the Si addition, reaching 62.5% — more than half of the observed surface area — in the molding with 20 mass% Si (Fig. 2(i)). Despite some errors, these results suggest that all the added Si precipitated as Mg₂Si within the magnesium alloy matrix.

Fig. 2

Optical micrographs of (a) AZ91D alloy and the Mg alloy with (b) 1.8 mass% Si (5 µm), (c) 3.6 mass% Si (5 µm), (d) 7.0 mass% Si (16 µm), (e) 7.8 mass% Si (5 µm), (f) 9.0 mass% Si (5 µm), (g) 12 mass% Si (5 µm), (h) 15 mass% Si (5 µm) and (i) 20 mass% Si (5 µm).

Fig. 3

Effect of Si addition on the median diameter of Mg₂Si.

Fig. 4

Effect of Si addition on total area fraction of Mg₂Si particles.

Figure 5 presents the XRD patterns obtained from the molded products with a 20 mass% Si addition. In these patterns, diffraction peaks corresponding to α-Mg and β-Mg₁₇Al₁₂ were identified, along with a peak attributable to Mg₂Si. The presence of the Mg₂Si phase was confirmed by the presence of its (111) plane diffraction peak detected at a 2θ angle of approximately 24.3°. Notably, the intensity of the strongest Mg₂Si peak surpassed that of the strongest α-Mg peak, corroborating the observation that Mg₂Si constituted more than half of the observed area in Fig. 2(i). Moreover, the absence of any diffraction peaks originating from silicon implies that the entire 20 mass% of silicon powder reacted with magnesium, forming Mg₂Si particles.

Fig. 5

XRD pattern of the Mg alloy with 20 mass% Si (5 µm).

Figure 6 shows the results of STEM-EDS analysis conducted on the product molded with the addition of 12 mass% Si. The images reveal the absence of voids at the interface between the primary Mg₂Si particles and the α-Mg phase (Fig. 6(a)) and β-Mg₁₇Al₁₂ phase (Fig. 6(b)). This observation indicates that primary Mg₂Si precipitated effectively through an in-situ reaction between molten magnesium and silicon powder in the cylinder of the injection molding machine, thereby establishing a robust interface between the matrix and Mg₂Si. Furthermore, the similarity in median diameters of the primary Mg₂Si particles in the molded products, regardless of the silicon powder diameters used (Fig. 3), suggests that these primary Mg₂Si particles precipitated directly from the molten metal.

Fig. 6

STEM-EDS analysis of the interfaces between (a) the Mg₂Si particle and α-Mg phase, and (b) Mg₂Si particle and combined α-Mg/β-Mg₁₇Al₁₂ phases in the Mg alloy with 12 mass% Si (5 µm). “α” and “β” represent α-Mg phase and β-Mg₁₇Al₁₂ phase, respectively.

Particles several tens of nanometers in size were detected at the interface area between the primary Mg₂Si particles and the matrix (Fig. 6). The presence of oxygen, indicated by the elemental maps in Fig. 6, suggests that these nanoparticles are likely oxides. This oxide formation could be attributed to the oxygen sourced from the oxide film on the magnesium chip surfaces. We theorized that this oxygen diffused into the liquid phase when the chip melted, and these oxides precipitated at the interfaces between the primary Mg₂Si particles and the matrix. This process likely occurred during solidification in the cooling phase of the molding process.

3.2 Effect of silicon addition on hardness and Young’s modulus hardness of injection molded products

Figure 7 shows the effect of Si addition on the hardness of the molded products. The hardness of the molded products increased linearly with the addition of Si, reaching 186 HV for the product with 20 mass% Si, compared to 73 HV for the AZ91D magnesium alloy molded product. Since Mg₂Si particles possess a hardness in the range of 600–700 HV,⁶⁾ the hardness of the products is presumed to increase with the total area fraction of the Mg₂Si particles. The average particle size of the added silicon powder appeared to have no discernable impact on the hardness.

Fig. 7

Effect of Si addition on hardness.

Tensile test-derived effects of Si addition on the Young’s modulus are shown in Fig. 8. The Young’s moduli of the molded products increased with higher silicon content. Notably, the product containing 20 mass% Si reached a Young’s modulus of approximately 75 GPa, which is about 1.7 times higher than that of the AZ91D alloy, measured at 43 GPa.

Fig. 8

Effect of Si addition on Young’s modulus with the rules of mixtures (Parallel model, Series model and Mori-Tanaka model).

The Young’s moduli obtained from our experiments were compared with theoretical values derived from the rules of mixtures. The theoretical values were calculated based on the series and parallel models,²⁴⁾ representing the upper and lower limits for simple composite behavior, and the Mori-Tanaka model, a well-known method for predicting the average behavior of composite materials,^25,26) as shown in Fig. 8. When applying these models to illustrate the behavior of Mg₂Si composites, the volume fractions of Mg₂Si in the molded products are crucial. For our calculations, we estimated the volume fractions of Mg₂Si based on the amount of added silicon, assuming complete incorporation of silicon into Mg₂Si molded products (Fig. 4). The Young’s modulus and Poisson’s ratio for the AZ91D magnesium alloy were 43 GPa and 0.35, respectively, whereas those for Mg₂Si were 108 GPa and 0.187, respectively. The Young’s moduli of the molded products were plotted near the curve of the series model and the Mori-Tanaka model for silicon additions of 12 mass% or less and for silicon additions of 15 mass% and 20 mass%, respectively. Thus, the experimental Young’s modulus values align well with the predictions of the rules of mixtures. For simplicity, the edge component in the rules of mixtures was assumed to be the AZ91D magnesium alloy (Young’s modulus: 43 GPa), regardless of the amount of silicon added. However, realistically, the magnesium in the AZ91D alloy is consumed by the reaction with silicon. Therefore, with increasing silicon addition, the chemical composition of the matrix (α-Mg and β-Mg₁₇Al₁₂ phases) tends to lower magnesium concentration, relatively high aluminum concentration, and an increased volume fraction of the β-Mg₁₇Al₁₂ phase compared to those of the original AZ91D magnesium alloy.

The observed increase in Young’s modulus in line with the rules of mixtures, coupled with the findings regarding the interface between Mg₂Si particles and the matrix, suggests that Mg₂Si particles can be effectively integrated into magnesium alloys using the injection molding method. This method holds significant promise for industrial applications due to its compatibility with conventional equipment and technologies. While there are challenges yet to be addressed, such as minimizing casting porosities through optimization of molding conditions, this technique has the potential to replace existing metallic materials with stiffer magnesium materials, ultimately reducing the weight of the components.

4. Conclusion

In this study, we utilized AZ91D magnesium alloy chips coated with silicon powder using a binder as the raw materials for the injection molding method. Our investigation focused on the effects of silicon addition on the microstructure, hardness, and Young’s modulus of the molded products, leading to the following conclusions:

1. (1)    Cross-sectional observations and X-ray diffraction analysis of the molded products indicated that all the added silicon precipitated as Mg₂Si in the magnesium alloy matrix.
2. (2)    No voids were detected at the interface between the primary Mg₂Si particles and matrix.
3. (3)    Vickers hardness of the molded products increased linearly with increasing silicon content.
4. (4)    The Young’s modulus of the molded products tended to increase with higher Si content, aligning with the rules of mixtures. Notably, the molded product containing 20 mass% Si exhibited a Young’s modulus of 73 GPa.

REFERENCES

• 1)   Y.  Kojima: J. JILM 58 (2008) 526–548. doi:10.2464/jilm.58.526
• 2)   S.  Saikawa: J. JILM 60 (2010) 571–577. doi:10.2464/jilm.60.571
• 3)   Y.  Chino,  X.  Huang and  K.  Suzuki: J. Japan Inst. Met. Mater. 81 (2017) 49–54. doi:10.2320/jinstmet.J2016059
• 4)   H.  Somekawa: J. Japan Inst. Met. Mater. 83 (2019) 65–75. doi:10.2320/jinstmet.J2018067
• 5)   G.  Sasaki,  M.  Yoshida and  H.  Fukunaga: Materia Japan 39 (2000) 519–525. doi:10.2320/materia.39.519
• 6)   K.  Kondoh,  H.  Ogimuma and  T.  Aizawa: Mater. Trans. 44 (2003) 524–530. doi:10.2320/matertrans.44.524
• 7)   Y.  Li,  T.  Ma,  Y.  Ren,  T.  Liu and  X.  Zou: Mater. Res. Express 7 (2020) 036533. doi:10.1088/2053-1591/ab7e4a
• 8)   K.  Kondoh,  H.  Ogimuma,  E.  Yuasa and  T.  Aizawa: Mater. Trans. 42 (2001) 1293–1300. doi:10.2320/matertrans.42.1293
• 9)   K.  Kondoh,  H.  Ogimuma,  R.  Tuzuki and  T.  Aizawa: Mater. Trans. 44 (2003) 611–618. doi:10.2320/matertrans.44.611
• 10)   K.  Kondoh,  H.  Ogimuma,  A.  Kimura,  S.  Matsukawa and  T.  Aizawa: Mater. Trans. 44 (2003) 981–985. doi:10.2320/matertrans.44.981
• 11)   H.  Muramatsu,  K.  Kondoh,  E.  Yuasa and  T.  Aizawa: JSME Int. J. Ser. A 46 (2003) 247–250. doi:10.1299/jsmea.46.247
• 12)   R.  Tsuzuki,  M.  Ishihara,  K.  Kondoh and  E.  Yuasa: J. Jpn. Soc. Powder Metallurgy 51 (2004) 736–740. doi:10.2497/jjspm.51.736
• 13)   K.  Kondoh,  R.  Tuzuki,  W.  Du and  S.  Kamado: Materia Japan 43 (2004) 275–280. doi:10.2320/materia.43.275
• 14)   K.  Kondoh: J. JILM 54 (2004) 187–191. doi:10.2464/jilm.54.187
• 15)   M.  Kamakura,  K.  Tokaji and  Y.  Uematsu: J. Soc. Mater. Sci., Japan 55 (2006) 55–60. doi:10.2472/jsms.55.55
• 16)   K.  Asano and  H.  Yoneda: Mater. Trans. 48 (2007) 1469–1475. doi:10.2320/matertrans.F-MRA2007834
• 17)   M.A.  Malik,  K.  Majchrzak and  K.N.  Braszczyńska-Malik: Archives of Foundry Engineering 12 (2012) 109–112. doi:10.2478/v10266-012-0116-8
• 18)   K.N.  Braszczyńska-Malik and  M.A.  Malik: Materials 13 (2020) 3591. doi:10.3390/ma13163591
• 19)   P.P.  Seth,  O.  Parkash and  D.  Kumar: RSC Adv. 10 (2020) 37327–37345. doi:10.1039/D0RA02744H
• 20)   Y.  Hideshima,  F.  Maeda,  T.  Fukuta and  K.  Ozaki: J. Japan Inst. Met. Mater. 86 (2022) 237–244. doi:10.2320/jinstmet.J2022019
• 21)   S.  Shinozaki,  J.  Morita and  K.  Wasai: J. JILM 55 (2005) 310–314. doi:10.2464/jilm.55.310
• 22)   S.  Akazawa,  T.  Fukuda,  K.  Obunai,  K.  Ozaki and  K.  Murakami: J. JILM 65 (2015) 113–117. doi:10.2464/jilm.65.113
• 23)   S.  Tang,  J.  Zhou,  C.  Tian and  Y.  Yang: Trans. Nonferrous Met. Soc. China 21 (2011) 1932–1936. doi:10.1016/S1003-6326(11)60952-7
• 24)   I.  Kimpara: High Polymer, Japan 31 (1982) 1081–1087. doi:10.1295/kobunshi.31.1081
• 25)   T.  Mori and  K.  Tanaka: Acta Metall. 21 (1973) 571–574. doi:10.1016/0001-6160(73)90064-3
• 26)  K. Togo: Zairyo Kyodo Kaiseki-gaku, (Uchida Rokakuho Publishing Co., Ltd., Tokyo, 2004).