Effect of Ca and P on the Size and Morphology of Eutectic Mg₂Si in High-Purity Al-Mg-Si Alloys

Abstract

Controlling the size and morphology of primary and eutectic Mg₂Si phase via modification and refinement (i.e., Ca and P addition) in Al-Mg₂Si-based alloys containing higher percentages of Si and Mg is of great necessity to ensure their high performance. The present investigation is mainly aimed to elucidate the effect of Ca and P on the nucleation and growth pattern of eutectic Mg₂Si in high-purity Al-13.1 wt.% Mg₂Si near-eutectic alloys with and without the additions of P and Ca. It was found that purity levels of raw starting materials exert a significant influence on the size and morphology of the near-binary eutectic Al-Mg₂Si. The presence of P suppresses the precipitation of ternary eutectic Al-Mg₂Si-Si. The presence of low levels of Ca promotes the faceted growth of Mg₂Si, while at sufficiently higher levels of Ca, Al₂Si₂Ca intermetallic is formed and the morphology of the Mg₂Si phase is changed from fibers/or globules to blocks or plates. The current investigation is of great importance for the further development of high-performance Al-Mg₂Si-based cast alloys.

1. Introduction

The 6xxx (Al-Si-Mg) wrought alloys with small additions of Si (0.7–1.3 wt.%) and Mg (0.6–1.2 wt.%) under different temper designations are some of the most widely used Al alloys due to their excellent deformation ability and mechanical properties. However, for aluminum foundry alloys, approximately 90% are based on Al-Si-Mg with higher additions of Si (about 7–12 wt.%) but low levels of Mg (0.3–0.7 wt.%) due to their excellent castability and good balance between mechanical properties and costs. Eutectic Si and β-series precipitates (i.e., Mg₂Si) formed during the ageing treatment are mainly responsible for their strengthening effects. Apart from the eutectic Si in Al-Si-based alloys, eutectic Mg₂Si in Al-Mg₂Si-based alloys has also been widely studied to develop new Al-Mg₂Si-based alloys containing higher percentages of Si and Mg with superior mechanical properties. Al, together with Mg and Si, forms a simple binary system with quasi-eutectic occurring at ~13.1 wt.% Mg₂Si. The vertical section constructed (in ternary system) through the Al-rich corner to the point demonstrates that the composition of the compound Mg₂Si is quasi-binary [1,2]. The alloys located on this line form a near-eutectic binary Al-Mg₂Si. It has been reported that the Mg₂Si phase is an intermetallic compound, in which the ratio of Mg/Si (about 1.73) can be adjusted to form Mg₂Si [3,4]. The eutectic composition of binary Al-Mg₂Si alloys has been determined [1,5,6,7]. However, the exact eutectic composition is still not yet well established. Previous investigations [8,9] suggested that the binary eutectic composition lies on the Mg-rich side. The Al-Mg-Si ternary system has a very narrow range for a binary Al-Mg₂Si eutectic. The variations of the composition outside this range cause the formation of either the ternary eutectic of Al-Mg₂Si-Si or Al-Mg₂Si-Mg₁₇Al₁₂ along with binary eutectic [10], which causes the degradation of mechanical properties at high temperatures because of the fact that the ternary eutectic of Al-Mg₂Si-Si or Al-Mg₂Si-Mg₁₇Al₁₂ has a lower melting temperature and thereby cannot be used at elevated temperatures. The composition of the ternary eutectic and its freezing temperature was also determined and documented [1,6,11]. However, the freezing temperature of these ternaries is, so far, not yet well-established. It has been reported [9] that the Al-Mg₂Si eutectic solidified in a non-faceted-faceted manner, in which Mg₂Si is a faceted phase while Al is a non-faceted phase. It has also been reported [12], that, using Cu wedge mold, the variation of cooling rates affects the size and morphology of the primary Mg₂Si and Al-Mg₂Si eutectic in a hypereutectic Al-20 wt.% Mg₂Si alloy. Primary Mg₂Si particles with a polygonal shape and refined Al-Mg₂Si eutectic cells were observed. Furthermore, increasing the additions of mischmetal metal (MM) reduces the size of primary Mg₂Si particles and also changes the morphology of eutectic Mg₂Si from flakes to fibers in in situ Al-15 wt.% Mg₂Si composites [13,14]. More interestingly, additions of P changed the morphology of primary Mg₂Si from dendritic or equiaxed to polygonal shape and thereby reduced the size of the particles [15]. These polygonal shape particles are advantageous for both mechanical and corrosion properties [16]. Chong et al. [17] reported that the addition of TiB₂ particles in Al-12.67Mg-10.33Si alloy reduced the size and changed the morphology of primary Mg₂Si from dendritic to polygonal shape because TiB₂ particles acts as nuclei for primary Mg₂Si due to their small lattice mismatch of only ~5%. On this basis, it can be concluded that most of the data are available on the modification and refinement of the primary Mg₂Si phase, but limited research works are available on hypo- and near-eutectic compositions. The effect of P additions on commercial-purity Al-Mg-Si near-eutectic alloys has been reported by Pabel et al. [18,19]. The additions of P in commercial-purity Al-Mg-Si near-eutectic alloys change the morphology of the eutectic Mg₂Si phase from lamellar to globular and remarkably enhanced the breaking strength of the alloy. It should be noted that although a detailed investigation on the effect of purity levels and P additions on eutectic Si in Al-Si alloys has been reported [20,21], but the effect of P additions on eutectic Mg₂Si phase in high-purity Al-Mg-Si near-eutectic alloys (in contrast to commercial-purity Al-Mg-Si near-eutectic alloys [18,19]) is still lacking. In terms of Ca addition, Moussa et al. [22] reported that the addition of 0.3 wt.% of Ca can effectively modify the morphology and refine the size of the primary Mg₂Si and produced best wear properties in hypereutectic Mg-5 wt.%Si alloy. It was also proposed [23] that the addition of Ca refine the eutectic Mg₂Si by the creation of Ca-based compounds which acted as nucleation sites for Al-Mg₂Si eutectic in hypoeutectic Al-Mg₂Si alloy. Another study [24] showed that the presence of Ca refines both the primary and eutectic Mg₂Si which resulted in the improvement of tensile properties in Mg-Mg₂Si composites. A literature survey suggested that the previous studies on Ca are mainly focused on the hypo and hypereutectic region. However, the effect of P and Ca on the nucleation and growth of eutectic Mg₂Si is not well understood in quasi-binary Al-Mg₂Si eutectics. This investigation is an attempt to further explain and clarify the effect of P and Ca on the size and morphology of near-eutectic Al-Mg₂Si in a high-purity Al-Mg-Si alloy system, which is of great importance for the further development of high-performance Al-Mg₂Si-based alloys.

2. Materials and Methods

Al-13.1 wt.% Mg₂Si eutectic alloys with and without additions of P and/or Ca, were prepared using 99.999 wt.% (5N) super purity electrolytically refined Al (Hydro Aluminium High Purity GmbH, Grevenbroich, Germany) and 99.999 wt.% (5N) purity Si (Siltronic AG GmbH, Munich, Germany) and 99.93 wt.% (3N3) Mg obtained from SAG GmbH (Lend, Austria). Additions of P and Ca were made using Al-19 wt.% Cu-1.4 wt.%P and Al-10 wt.%Ca master alloy rods, respectively. It should be noted that when P was added, Cu was also added, but the possible effect of Cu on the eutectic Mg₂Si phase are not included here. The alloys were first prepared by melting Al and Si together in a clay-bonded graphite crucible in an electric resistance furnace. The crucible was coated from the inside by ceramic material (BN) to avoid any possible reaction between the molten material and the crucible. The calculated amount of Mg was subsequently added. All alloys were gravity die cast in a specially designed permanent cast iron mold at a temperature of ~720 °C. The mold temperature was kept at 350 °C. The melt temperature was measured using a portable K-type Chromel-Alumel thermocouple connected to a digital meter, which is able to measure temperatures up to 1300 °C. Graphite spray was applied inside the mold prior to pouring to increase heat transfer through mold wall and facilitate the removal of the cast from the mold. A filter was placed at the location shown in Figure 1a to reduce the effect of bi-films and to ensure the smooth entrance (with a minimum turbulence) of the melt into the mold. Figure 1 shows the pictures of the permanent mold, an actual cast part and the die dimensions (unit: mm), respectively. Alloys with additions of 50, 200 and 300 ppm P levels were also manufactured using the same procedures. Quantitative chemical analysis was obtained from OES-spark analysis whilst the trace P content was measured using glow discharge mass spectroscopy (GDMS) (SHIVA Technologies, EAG Lab., Toulouse, France). The measured compositions of all manufactured alloys are listed in

Table 1. Measured composition of five alloys manufactured for this investigation.

No.	Mg wt.%	Si wt.%	P (ppm)	Fe (ppm)	Ca (ppm)	Cu wt.%	Cr (ppm)	Ni (ppm)	Ti (ppm)	Al
1	8.61	4.58	0.4	<5	27	<1	<3	<4	<4	Bal.
2	8.45	4.60	50	35	8	0.045	<3	<4	<4	Bal.
3	8.64	4.54	200	40	6	0.273	<3	<4	<4	Bal.
4	8.45	4.60	300	44	6	0.366	<3	<4	<4	Bal.
5	8.50	4.60	0.4	35	500	<1	<3	<4	<4	Bal.

. In addition, a high-purity Al-13.1 wt.% Mg₂Si eutectic alloy containing 500 ppm Ca (only 0.4 ppm P) was also prepared by using the same experimental procedure in order to elucidate the effect of high Ca addition levels on the eutectic Mg₂Si phase (also listed in Table 1).

Scheil simulation using Thermo-Calc software (2022b, Stockholm, Sweden) with a TTAL5 database was performed to predict the phases formed during solidification. Their formation temperatures and relevant mole or weight fractions of all solid phases or each phase were also calculated. A Quik-Cup obtained from Heraeus Electro-nite GmbH (Bitterfeld-wolfen, Germany) was employed to analyze the thermal response of the alloys. For DSC analysis, samples were machined in the forms of small discs and with weights of ~50 mg. Samples were heated from 350 °C to 620 °C at a heating rate of 20 °C/min in a power-compensated PerkinElmer (Waltham, MA, USA) diamond DSC apparatus. Samples were held at this temperature (620 °C) for 10 min to ensure the homogenization of 620 °C and then cooled from 620 °C to 350 °C with the same cooling rate (20 °C/min.)

Sample preparation for microstructure characterization followed the standard methods, including grinding and polishing. For optical microscopy, ZEISS Imager AXIO was used, while SEM analysis was performed on a FEI QUANTA 200 equipped with an EDX (Oxford Instruments INCA x-sight, Abingdon, UK) system. For image analysis, a NIS elements BR 3.0 SP3 Hotfix 4 build 472 basic research software (NIKON, Japan) was employed. The measurement of the distribution and shape factor of Mg₂Si was carried out with at least three different micrographs and a minimum of 50 readings were taken each time. The obtained results can be expected to be typical and reliable. In order to examine the three-dimensional (3D) morphology of eutectic Mg₂Si, samples were deeply etched using 30% NaOH solution at a temperature of 60 °C for 10 min [25]. Thin foil specimens for TEM investigations were prepared by a Model 691 Gatan PIPS™ (Precision Ion Polishing System, Gatan, Pleasanton, CA, USA) with an angle of 4° and a voltage of 4 KV. TEM was performed using a Philips CM12 (Philips, Eindhoven, The Netherlands) operated at 120 KV and an image-corrected JEOL-2100 (JEOL Ltd., Tokyo, Japan) operated at 200 KV.

3. Results

3.1. Scheil Simulation

Scheil simulation using Thermo-Calc software with a TTAL5 database was performed by adding Mg = 8.6 wt.%, Si = 4.5 wt.% and the remaining aluminum, respectively. The simulation was used to predict the formation temperature of eutectic, weight fraction and solidification range of the solid phases evolved during solidification. Figure 2 illustrates the mole fraction of solid phases and weight fraction of Mg₂Si phase in near-eutectic Al-13.1 wt.% Mg₂Si alloy without adding any other chemical elements. The liquidus and eutectic temperatures of the manufactured alloys were predicted to be ~598 °C and ~594 °C, respectively, as shown in Figure 2a. The labeled numbers 1–4 show the solidification path of the phases. The presence of the Si phase at the end of solidification along with Al and Mg₂Si was also predicted, as shown in Figure 2b. The black curve in Figure 2a was predicted under the equilibrium conditions.

3.2. Thermal Analysis

Figure 3a shows the cooling curve (black) of the Al-13.1wt.% Mg₂Si near-eutectic alloys without the addition of chemical elements. Two distinct peaks can be observed on the superimposed 1st derivative dT/dt curve (blue). The first peak corresponds to the nucleation of the Al-Mg₂Si eutectic (indicated by a sharp increase in the 1st derivative dT/dt curve), while the second peak signifies the nucleation of ternary eutectic Al-Mg₂Si-Si at a temperature of ~548 °C. Experimentally, the liquidus temperature of the alloy was measured as ~600 °C (close to the calculated value (598 °C) by Thermo-Calc) and the eutectic temperature was measured as ~594 °C (the same as the calculated value (594 °C) by Thermo-Calc). This indicates that the experimental and simulation results are in a good agreement. Figure 3b shows the DSC solidification exotherms for the Al-Mg₂Si near-eutectic alloy without the addition of chemical elements. The first sharp exotherm A represents the nucleation of Al-Mg₂Si eutectic cells and a smaller exotherm manifests the nucleation of ternary eutectic (i.e., Al-Mg₂Si-Si) with the nucleation temperature onset of ~548 °C. It should be noted that the exotherm A is not sharp, which can be attributed to the higher sample masses (~50 mg) employed in the DSC pan in order to clearly observe the exotherm B.

P content in the Al-13.1 wt.% Mg₂Si alloy without the addition of chemical elements was measured as 0.40 ppm by using the GDMS technique. Figure 4a,b shows the cooling curves (black) for the alloys containing 200 and 300 ppm P, respectively, with the superimposed 1st derivative dT/dt curves (blue). In both cases, only one sharp increase in the 1st derivative dT/dt curve (i.e., the nucleation of the Al-Mg₂Si eutectic) was observed, while no further variations in cooling rate were recorded. Figure 4c shows the DSC solidification exotherms of Al-Mg₂Si alloys with different addition levels of P. In the first thermogram (with 50 ppm P content), a small ternary eutectic peak was evident, while in thermograms 2 and 3, no other peak was apparent except the large exotherm A. This may be due to the fact that the addition of P restrains the nucleation of the ternary Al-Mg₂Si-Si eutectic.

3.3. Microstructural Characterization

Figure 5 shows the optical micrograph of the Al-13.1 wt.% Mg₂Si near-eutectic without any addition at lower and higher magnifications when the mold was heated to 350 °C prior to pouring. The microstructure mainly consists of fine Al-Mg₂Si eutectic cells. Mg₂Si, with fibrous or cylindrical shape, along with visible irregular and dog bone-shaped morphology, can be seen in Figure 5d at a higher magnification. Figure 6 shows the micrographs of fine Al-Mg₂Si eutectic grains at a mold temperature of 350 °C. The detailed analysis of the micrographs was performed using image analysis software (NIS elements BR 3.0 SP3 Hotfix 4 build 472). Various geometrical features of the Mg₂Si phase were calculated, including area (µm²), perimeter (µm), equivalent diameter (µm) and phase count. It should be noted that mean values of the number of Mg₂Si fibers in the selected region of interest (ROI), i.e., 10 × 10 µm² (see Figure 6b), are 110, while average equivalent diameter [26,27] was approximately ~0.3732 μm when the mold temperature was kept at 350 °C. Figure 6d shows a ternary eutectic which formed in the vicinity of or at the eutectic cell boundaries in the alloy containing only 0.4 ppm P. Figure 7a shows an SEM backscattered (BS) image in which three phases (i.e., Al, Mg₂Si and Si) can be distinguished by color contrast, as indicated by arrows on the micrograph. Figure 7b shows corresponding elemental maps from the selected region. The color contrast illustrates the presence of Al, Mg and Si, respectively. Figure 8 shows optical and SEM micrographs for the alloy containing 300 ppm P and no ternary eutectic was apparent in the vicinity of or at the eutectic cell boundaries; however, a few primary Mg₂Si particles with a regular shape were observed. Figure 9 shows deep-etched SEM BS images of the Al-13.1 wt.% Mg₂Si-300 ppm P alloy. The 3D fibrous morphology of Mg₂Si is visible on both micrographs along with a few polygonal-shaped Mg₂Si particles. Figure 10 shows a SEM secondary electron (SE) image taken from a deeply etched sample and an EDX spectrum taken from the selected point, which confirms the Mg₂Si phase composition. In order to further elucidate the shape of the eutectic Mg₂Si phase, Figure 11a–c illustrates the bright-field TEM image and the corresponding selected area diffraction pattern of the Al-13.1 wt.% Mg₂Si alloy. The faceted Mg₂Si phase can easily be identified on the micrograph.

In addition, in order to elucidate the effect of Ca on the shape of the eutectic Mg₂Si phase, Figure 12a shows the presence of the Al₂Si₂Ca intermetallic in the Al-Mg-Si alloy system containing 500 ppm Ca. Figure 12b–e shows the corresponding elemental maps, while Figure 12f represents the EDS spectrum taken from the location, as shown in Figure 12a. It should be noted that, at the higher levels of Ca, the morphology of Mg₂Si is changed from fibers or globules to a block/or plate. Figure 12g shows a schematic diagram, which indicates the presence of Al₂Si₂Ca at the growing interfaces of the Mg₂Si phase. Furthermore, Figure 13 shows the plate-shaped Mg₂Si phase. No Ca-rich layer was observed along the plate-shaped Mg₂Si in the [011] zone axis within the resolution in the present investigation. Figure 14 shows the globular-shaped Mg₂Si phase. No Ca-rich layer was observed along the globular-shaped Mg₂Si in [011] zone axis within the resolution in the present investigation. A detailed analysis concerning thermal analysis, DSC and microstructure characterization is discussed below.

4. Discussion

4.1. Microstructure without and with Additions of Ca

The microstructure of the Al-13.1 wt.% Mg₂Si near-eutectic consists of very fine eutectic grains (see Figure 5 and Figure 6). Two-dimensional (2D) fine Mg₂Si globules should be fibers in 3D with a short length. For the formation of the Mg₂Si compound, the solidification theory suggests that the solute redistribution during solidification increases the local concentration of both Si and Mg in the melt, which leads to the formation of intermetallic Mg₂Si. This occurs because there is a strong chemical affinity between Si and Mg. The nucleation and growth mechanism for a high-purity Al-13.1 wt.% Mg₂Si near-eutectic is proposed below, in spite of a lack of strong experimental support.

4.1.1. Nucleation Mechanism

The nucleation during solidification will always be aided by the presence of inherent impurities in the melt. However, the alloys in the present investigation were manufactured using high-purity materials; furthermore, in-mold filtration was employed to minimize the effect of oxides. In this scenario, the presence of fine eutectic grains in high-purity melts can be explained as follows: firstly, during an initial eutectic reaction, Al, as the major phase, will nucleate first at some unknown substrate or independently (homogeneously at a larger undercooling), which then acts as a nucleating agent for Mg₂Si. This occurred simultaneously in eutectic cells, initiating in melt volume at multiple locations. In each eutectic cell, a so-called decoupled growth then proceeds separately until it terminates at the cell boundaries with the creation of irregular and faceted Mg₂Si. Secondly, the independent or homogeneous nucleation of both Al and Mg₂Si occurred at a high undercooling at multiple locations in a given melt volume, and eutectic will then grow in a decoupled manner. Thus, a globule and fine structure may form due to the absence of favorable nucleation sites, but the growth proceeds extremely slowly. The proposed nucleation sequence is shown schematically in Figure 15.

As discussed above, it can be concluded that impurities present in the raw starting materials may have a significant influence on the morphology, size and distribution of Mg₂Si in the Al-Mg₂Si near-eutectic alloys. However, the exact nucleation and growth mechanisms involved remain unclear and still need further investigations. DSC traces and thermal analysis (see Figure 3a,b) also confirm the presence of ternary eutectic. The freezing point of ternary eutectic was found to be 548 °C (apparent on DSC traces), which is less than in previous findings [6,11]. The recorded lower freezing temperature for ternary eutectic nucleation might be due to the use of ultra-pure materials for alloy preparation and thus an absence of favorable nucleation sites. However, the exact composition of the ternary eutectic was not determined in the present investigation. The Al-Mg-Si ternary system has a narrow range for the binary Al-Mg₂Si eutectic [10]. The presence of a ternary eutectic (i.e., Al-Mg₂Si-Si (see Figure 6d) at the eutectic cell boundaries might suggest that there is still a slight variation in the composition of the selected alloy, i.e., Al-13.1 wt.% Mg₂Si. Due to this, the excess Si is available in the melt at the end of solidification, alongside the required Si for the formation of Mg₂Si, and thus nucleates and grows simultaneously to Al and Mg₂Si.

4.1.2. Growth Mechanism

In high-purity Al-Mg-Si near-eutectics, the presence of fibrous Mg₂Si suggests that the growth of the Mg₂Si phase is isotropic. A similar growth pattern of Si was also observed in Al-Si hypoeutectic alloys with the addition of impurity elements, such as Sr and Na [20,21,28], which are well-known as the impurity-induced twinning mechanisms (IIT), but no inherent twins were observed in Mg₂Si in the present investigation under the given growth conditions. Furthermore, at the eutectic cell boundaries, towards the end of solidification, the growth pattern of Mg₂Si changes from fiber to irregular flake shapes, which can be seen clearly in deep-etched samples (Figure 9 and Figure 10). This indicates that the growth pattern changes from a non-faceted to faceted mode (growth takes place along certain crystallographic directions). Jackson [29,30,31] proposed a relationship to distinguish between faceted and non-faceted materials in terms of a factor $\mathit{\alpha }$, which can be expressed as:

$$\mathit{\alpha }=\left(\frac{∆{\mathit{S}}_{\mathit{m}}}{{\mathit{R}}_{\mathit{g}}}\right)\left(\frac{{\mathit{C}\mathit{N}}_{\mathit{S}}}{{\mathit{C}\mathit{N}}_{\mathit{I}}}\right)$$

(1)

where ${\mathsf{\Delta }\mathit{S}}_{\mathit{m}}$ is the entropy of fusion in J/mol and ${\mathit{R}}_{\mathit{g}}$ is the ideal gas constant measured in J/mol.K. ${\mathit{C}\mathit{N}}_{\mathit{s}}$ is the surface coordination number, ${\mathit{C}\mathit{N}}_{\mathit{I}}$ is the internal coordination number. Surprisingly, for Mg₂Si (nucleating at a high undercooling), $\mathit{a}$ is less than 2, and for Al, $\mathit{a}$ is also less than 2. The calculation of $\mathit{\alpha }$ factor is given in the Supplemental Materials (SI). This suggests that the eutectic will grow in a non-faceted/non-faceted manner (regular eutectic). However, TEM image (Figure 11) confirms the presence of a faceted eutectic Mg₂Si phase. A question that arises here concerns why the growth pattern of Mg₂Si changes from the non-faceted to the faceted mode. It should be noted that a noticeable amount of Ca (see Table 1) is still present, although the alloys were prepared using high-purity Al. The key point is that during the electrolytic refining of aluminum by the so-called “three-layer method”, Ca is added into the electrolyte as CaF₂ (calcium fluorite) to increase its conductivity. From there, Ca is deposited at the cathode along with Al [32] and became an impurity within high-purity Al. The presence of the Ca plays an important role here, which may change the growth pattern, as discussed below. In the first case, when Ca is present in small quantities, the rejection of Ca (since Ca has no solid solubility [33] in Al) ahead of the growing fronts of Mg₂Si may produce a fine layer of either Al₂ Si₂Ca, Ca₂Si or Ca_xMg_ySi (x and y describe the possibilities of the formation of different stoichiometric compounds) intermetallic [29,34], and this Ca-rich layer might act as a precursor to increasing the Jackson $\mathit{\alpha }$ factor (i.e., $\mathit{\alpha }$ > 2), which is just sufficient to force the microscopic appearance of Mg₂Si from non-faceted to faceted, as shown in TEM images (Figure 11a,b). However, at this level of Ca, no Ca-rich layer was observed along the fibers and plates of Mg₂Si. In the second case, when Ca is present in large quantities, the morphology of eutectic Mg₂Si is changed from a fiber to a plate and/or block, as shown in Figure 12a. Ca is proposed to play a dual role. Firstly, instead of forming a fine Ca-rich layer at growing interfaces, it may substitute inside Mg₂Si as (Ca, Mg)₂Si and increase its entropy of fusion, which alters the growth pattern of Mg₂Si from non-faceted to faceted. The argument is based upon the EDS map of Ca (Figure 12e), in which Ca and Mg overlap. It is important to note that the entropy of the fusion of a compound is related to the degree of disorder in the system, which can be affected by changes in the composition or structure of the compound. In the case of (Mg, Ca)₂Si, if Ca replaces Mg in Mg₂Si (Ca distribution is evident in Mg₂Si crystal, see Figure 12), the degree of disorder in the system may increase. This is because of the fact that Ca and Mg have different ionic radii and chemical properties, which can lead to changes in the crystal structure of the compound. However, it should also be noted that the exact magnitude of the increase in the entropy of fusion depends upon the amount of Ca that is incorporated into the Mg₂Si lattice and the resulting changes in the crystal structure. Secondly, the formation of the Al₂Si₂Ca intermetallic, at sufficiently higher levels of Ca, appears to nucleate on Mg₂Si and forces the eutectic Mg₂Si to grow as blocks and/or plates, as shown in the SEM image Figure 12a and schematically in Figure 12g. The EDS shown in Figure 12f validates the appearance of Al₂Si₂Ca. Moreover, no Ca-rich layer was observed along both plate and fiber shape Mg₂Si in the Al-Mg-Si alloy containing 500 ppm Ca. The TEM images (Figure 13a–c), taken by tilting toward the zone axis [011], confirm the absence of a Ca-rich layer along the plate shape Mg₂Si within the resolution in the present investigation. Similarly, the TEM images (Figure 14a–c), taken by tilting toward the zone axis [011], of globular-shaped Mg₂Si also show no Ca-rich layer along the globular-shape Mg₂Si within the resolution in the present investigation. On the basis of the above evidence, it can be concluded that when Ca is present in higher quantities, it either forms a separate Al₂Si₂Ca or substitutes inside Mg₂Si.

4.2. Effects of P on Nucleation and Growth of Ternary Eutectic

Cooling curves (Figure 3a) and DSC solidification thermograms (Figure 3b) revealed the presence of the ternary eutectic phase. However, with the addition of 200 ppm to 300 ppm of P, the exotherm B, which corresponds to the solidification of the ternary eutectic (i.e., Al-Mg₂Si-Si), is absent. This effect of P, observed in high-purity near-eutectic Al-Mg-Si alloys, is in accordance with the previously published patent work by Pabel et al. [18] in the commercial-purity Al-13.1 wt.% Mg₂Si near-eutectic alloys. It has also been reported that additions of Li produced similar results [30], but much higher quantities (Li) are required in comparison to P. Ternary eutectic (i.e., Al-Mg₂Si-Si in the present investigation) has a melting point of only ~548 °C and forms at the lateral stages of solidification and nucleates at eutectic cell boundaries. The presence of this ternary eutectic in the alloy causes the degradation of mechanical properties at high temperatures due to the fact that the ternary eutectic has a low melting temperature [10].

The suppressing action of P on the nucleation of the ternary eutectic can be explained as follows. On the one hand, if primary Al dendrites are formed during the initial decoupled eutectic reaction, they may act as nucleating substrates for AlP. This is also true for hypoeutectic Al-Si alloys, where Al is formed first at a low P content and under non-equilibrium conditions. It is well-established that the AlP compound is a potent nucleation site for Si in hypoeutectic Al-Si alloys [35,36,37,38]. Similarly, in Al-Mg₂Si alloys with a low P content, some AlP may be available to nucleate excess Si. On the other hand, in the case of eutectic and hypereutectic Al-Mg₂Si alloys at a higher P content, AlP formed first and then Mg₂Si nucleates directly on AlP. Ultimately, no or less (if any) AlP is available to nucleate excess Si. The remaining Si may then form an Al-Si binary eutectic or appear as free Si alone if a higher amount of Si is available. This free Si might form at the eutectic cell boundaries. Furthermore, another important aspect is the interaction between Ca and P. Thermodynamics revealed that, during solidification, the binary compound calcium phosphide, i.e., 1/2Ca₃P₂ is formed preferentially, since the free enthalpy ${∆\mathit{G}}_{\mathbf{298.15}}$ of calcium phosphide Ca₃P₂ is lower than the free enthalpy of AlP [39]. The reaction 3/2Ca + 2P → 1/2Ca₃P₂ suggested that Ca is the limiting reactant while P is an excess reagent. After the consumption of all Ca, the remaining P is then available to form AlP, which indicates that fewer nucleation sites (in the form of AlP) are available to nucleate both Mg₂Si and Si. This is fully consistent with our previous work [18,19], where it was assumed that P combines with Ca and forces the formation of eutectic Mg₂Si by supercooling.

The fine fibrous and globular morphology of the Mg₂Si phase indicates that the formation of AlP compounds may also act as a nuclei for eutectic Mg₂Si, since physical and structural similarities exist between Mg₂Si and Si (AlP is a potential substrate for Si) [13,40]. However, it should be noted here that although there are the structural similarities, atomic mismatch between AlP (substrate) and Mg₂Si (nucleating crystal) is approximately ~14%, which is higher than in the case of AlP for Si. The atomic mismatch between the substrate and nucleating crystals can be calculated through a useful equation proposed by Turnbull [41], as shown below:

$$\mathit{\delta }=\frac{\left|{\mathit{a}}_{\mathit{s}}-{\mathit{a}}_{\mathit{c}}\right|}{{\mathit{a}}_{\mathit{c}}}\times \mathbf{100}$$

(2)

where ${\mathit{a}}_{\mathit{s}}$ and ${\mathit{a}}_{\mathit{c}}$ are the lattice parameters of substrate and nucleating crystals. In a deeply etched sample containing 300 ppm P (Figure 9b), a few polygonal-shaped Mg₂Si particles were observed. The structure consists of squares and hexagons, indicating that it is a geometrical shape identical to a tetrakaidecahedron. This kind of Mg₂Si morphology was also previously reported by Qin et al. [15] at a very high P content, i.e., 0.41 wt.% P. However, in the present investigation, this morphology was observed at a much lower P content (300 ppm P). Both suggest that at higher levels of P, AlP forms first, which may act as a nucleating agent for primary Mg₂Si. Due to the fact that no evidence of the attachment of AlP was found, both for eutectic Mg₂Si and primary Mg₂Si particles during the TEM investigation, it is therefore impossible to fully clarify the effects of P on the nucleation and growth of the ternary eutectic (Al-Mg₂Si-Si). A more detailed composition analysis with a higher resolution (i.e., atom probe tomography) is needed. In addition, considering that fact that the spheroidized morphology is effective in enhancing mechanical as well as corrosion properties, further investigations on these properties are also needed.

5. Conclusions

We have investigated the effect of Ca and P on the nucleation and growth pattern of eutectic Mg₂Si in high-purity Al-13.1 wt.% Mg₂Si near-eutectic alloys without and with the additions of P and Ca. We found that purity levels of the starting materials exert a great influence on the final microstructure of near-eutectic Al-Mg-Si alloys. We proposed that the inherent low levels of Ca, in raw starting materials, promote the faceted growth of Mg₂Si by forming a Ca-rich layer ahead of the growing fronts. However, at sufficiently higher levels of Ca, it may either substitute inside Mg₂Si as (Ca, Mg)₂Si or form Al₂Si₂Ca intermetallic. The Al₂Si₂Ca intermetallic appears to nucleate on the eutectic Mg₂Si and forces the Mg₂Si to grow in the form of blocks and/or plates instead of fibers. P additions suppress the formation of the non-equilibrium ternary eutectic (Al-Mg₂Si-Si). With the additions of higher levels of P, AlP forms first and acts as a nucleating agent for primary and/or eutectic Mg₂Si. The present investigation is of great importance for the further development of high-performance Al-Mg₂Si-based alloys.