Spheroidization and Coarsening of Eutectic Si Particles in Al-Si-Based Alloys

/e present study was carried out on three Al-Si cast alloys viz., 319, 356, and 413 alloys, solidified at 8°C/s. Samples from 319 and 413 alloys were solution heat-treated at 510°C, whereas samples from 356 alloy were solutionized at 550°C, for up to 1200 h. /e results reveal that complete spheroidization of eutectic Si particles in terms of achieving individual spherical particles cannot be achieved inmost Al-Si-Cu-Mg alloys even after a solutionizing time of 1200 h which contradicts with the existing theory. Addition of Sr to Cu-free 356 alloy could lead to complete spheroidization after 1200 h at 550°C if the alloy was solidified at 8°C/s. Besides the dissolution theory of Ostwald, coarsening of Si particles can as well take place by impingement, fusion, and agglomeration. Increasing the Si content makes it difficult to achieve spheroidization, i.e., fragmentation and coarsening. Results obtained from observations of deeply etched samples (3D) contradict those obtained from polished samples (2D).


Introduction
William Ostwald [1,2] suggested the dissolution of small crystals or solid particles and the redeposition of the dissolved species on the surfaces of larger crystals or solid particles. e ripening process occurs because larger particles are more energetically favored than smaller particles, giving rise to an apparent higher solubility for the smaller ones [3] as schematically shown in Figure 1.
In the case of unmodified Al-Si alloys (for the majority of alloys), coarsening of Si particles takes place by a combination of particle fragmentation mechanism along with the silicon particle dilution mechanism, similar to that proposed by Ostwald. Smaller silicon particles dilute and diffuse to cause some silicon particles to grow laterally as presented in Figure 2 [4].
In nonmodified Al-Si-Mg alloys, the coarsening rate of eutectic silicon was found to follow the Lifshitz-Slyozof-Wagner (LSW) theory, which corresponds to a zero volume fraction approximation. e LSW model takes place after an initial time delay during which the fragmentation of the silicon plates takes place [5,6]. e coarsening rate can be calculated using equation (1) following the diffusion-controlled growth model [7,8], also known as the LSW model: where r 3 denotes the final average radius, r 0 3 is the average radius at time (t) � 0, and K LSW is the coarsening rate constant.
Colley [9] suggested that the roundness relates to the sharpness or smoothness of the particle's corners or edges, whereas the sphericity (or circularity) determines how the overall shape of the particles approaches that of a sphere (3D) or a circle (2D). e equivalent circular diameter is the diameter of a circle that has the same area (A) as the Si particle and may be calculated using the following formula ( Figure 3): equivalent circular diameter (d) � ���������������� 4 * particle area (A) π .
e aspect ratio of the particle is calculated by drawing an imaginary best fit ellipse to the particle projection; by measuring the major and minor axes of the ellipse, the aspect ratio can be calculated according to the following relation: aspect ratio (A.R.) � major axis length (L) minor axis length (B) .
e sphericity or circularity of the Si particles is a function of the perimeter (P) and the particle area (A); its value lies between 0 and 1, where the maximum value of 1 refers to a perfect circle. e sphericity is quantified as follows: sphericity � 4 * π * particle area (A) Perimeter (P) 2 .
Finally, the roundness of the particle's edges is a function of the particle area (A) and the length of the best fit ellipse major axis (L) and is given by: roundness � 4 * particle area (A) π * major axis length (L) 2 .
In the case of modified Al-Si alloys, prior to heat treatment, the addition of Sr changes or "modifies" the acicular morphology to a fine, fibrous form. Later, solution heat treatment (SHT) helps in spheroidizing the eutectic silicon in either nonmodified or modified alloys [10] (Figure 4). e change in the eutectic Si morphology occurs in two stages: (i) dissolution or fragmentation of the eutectic Si branches after necking and (ii) granulation or spheroidization of the fragmented branches as their average size (a) Figure 1: Schematization of the Ostwald dissolution mechanism: (a) before solutionizing; (b) dissolving of the small particles; (c) the large particles growing due to dissolution of the small ones (broken arrow) and/or by merging (black arrow) [3].

Atoms of Si
As cast Fragmentation and coarsening Figure 2: Schematic presentation of the new mechanism for the development of eutectic silicon particles during solution heat treatment [4].

Particle area (A)
Equivalent circular area Equivalent diameter (d) Figure 3: Schematic representation showing the projection of the Si particle as the equivalent circle with the same area A and the equivalent circular diameter d.

2
Advances in Materials Science and Engineering decreases, allowing eventual spheroidization. Many authors [11,12] have stated that fragmentation of acicular silicon is facilitated by previous Sr treatment so that the resulting fibrous eutectic silicon would be spheroidized in a shorter time; this behavior was attributed to the larger interfacial area of the fibrous structure and driving force for morphological change [6,13,14]. On the other hand, nonmodified structures will take a longer time to be spheroidized. Figure 4 shows the sequence of spheroidization of the eutectic silicon particles in both nonmodified and modified Al-Si alloys [5,6,15,16]. e processes of spheroidization and coarsening of eutectic silicon particles result mainly from the reduction in the surface energy associated with the interface between the silicon particles and the Al matrix [13,[17][18][19][20][21]. At elevated temperatures, the size and frequency of surface perturbations increase at the interface, leading to the formation of near-spherical Si particles due to the breakdown of the eutectic silicon particles. Subsequently, these near-spherical particles tend to coarsen to further reduce the interfacial area; the coarsening occurs by the dissolution of the smaller Si particles into the larger (more stable) particles following the Ostwald ripening mechanism.
Ogris et al. [17] studied the fragmentation of eutectic silicon rods into spheres in the Al-Si-Mg alloy by considering the model shown in Figure 5, where the cylindrical rod in the top half of the model represents the interconnected modified silicon particles. e authors concluded that the fragmentation time for eutectic silicon rods during solution treatment is highly dependent on the initial rod radius and the solutionizing temperature. e main objective of the present work is to examine the abovementioned theories in Al-Si-Mg and Al-Si-Cu-Mg alloys using polished (2D) as well as deeply etched (3D) samples (solidified at 8°C/s) subjected to solutionizing time of up to 1200 h at different solutionizing temperatures.

Experimental Procedure
e as-received 354 alloy ingots were cut, dried, and melted in a 70 kg capacity SiC crucible using an electric resistance furnace. e Sr-modified alloy series was obtained by adding ∼200 ppm Sr using Al-10 wt.% Sr master alloy, whereas the grain refinement was accomplished through the addition of Al-5 wt.% Ti-1 wt.% B master alloy in the form of rods in order to achieve a level of ∼0.2 wt.% Ti in the final alloys. e melt was kept at a temperature of 800 ± 5°C. e degassing process was carried out using a rotary graphite impeller that rotates at ∼150 rpm for 15-20 min; pure dry argon was pumped inside the melt through the rotating impeller at a constant rate of 20 m 3 /h. It should be mentioned here that Sr was added 5 min before the end of degassing to minimize its oxidation. After degassing, the melt was carefully skimmed to remove the oxide layers from the melt surface. e melt was then poured into the 2 preheated permanent molds. Each permanent mold preheated at 450°C employed was preheated in order to remove all traces of moisture from the mold. An ASTM B-108-type permanent mold was used to prepare castings from which the standard tensile test bars were obtained (Figure 6(a)). e rest of the molten alloys were poured into a metallic mold heated at 350°C (Figure 6(b)) [22]. e solidification rate was ∼8°C/s in both cases. Table 1 lists the chemical composition of the three alloys used in the present work.
Tensile testing at the ambient temperature was carried out using an MTS servo-hydraulic mechanical testing machine at a strain rate of 4 × 10 −4 ·s −1 , for the as-cast and heattreated test bars. e attachable extensometer (strain gauge) was used to measure the deformation that takes place in the samples during the test, and the data acquisition system attached to the machine converts it to an accurate measure of the percentage elongation. e data acquisition system provides the tensile properties in terms of ultimate tensile strength (UTS), yield strength (YS), and percentage elongation to fracture (%El). Five test bars for each alloy/condition were tested, and the average values of ultimate tensile strength (UTS), 0.2% offset yield strength (YS), and percentage elongation to fracture (%El) were reported to represent the tensile properties of the corresponding alloy/ condition.
Samples for metallographic examination were sectioned from the cast bars (25 mm × 25 mm × 5 mm) and polished following standard procedures. Eutectic Si characteristics were evaluated using a Clemex image analyzer coupled with an optical microscope. Measurements were carried out at 100x as shown in Figure 7. For each sample, at least 100 fields were scanned to enhance the accuracy of the measurements. In order to better understand and visualize the morphological changes occurring in the silicon particles during extended solution treatment, the samples were deep etched using Keller's etchant to dissolve the aluminum matrix and expose the eutectic silicon particles. ese samples were then examined by a Hitachi-SU8000 field-emission scanning electron microscope (FESEM) using the backscattered electron (BSE) mode. Tables 2-5 shows that the average surface area of the silicon particles of alloy 319 increases with solutionizing treatment, both for modified and unmodified samples. us, there is a growth of the thickness of the silicon particles which may be accompanied by an increase in length for the unmodified samples. For Srmodified samples, the silicon particles undergo fragmentation, followed by coarsening.

Image Analysis Results. A study of
us, particle density is increased in these samples. Solution heat treatment has the effect of reducing the density of the Si particles in the samples.
erefore, there is dissolution of certain silicon particles which feed the growth of other silicon particles. Roundness is higher for samples that contain strontium (62.6% in amount). Roundness for such samples also increases with SHT. In the case of high Si-containing alloy, i.e.,   Figure 5: Schematic illustrating a rod-shaped eutectic particle that fragments into a series of spherical particles [17]. 413 alloy, the rate of fragmentation of silicon particles is lower compared to that observed for 319 and 356 alloys, which allows for an increase in the thickness of the silicon particles, to maintain or increase the surface area of the particles. As in the case of other alloys, roundness, for samples containing strontium, also increases with SHT (76.6% in amount).

Optical
Microstructure. Figures 8-10 represent the microstructure of the present alloys under different working conditions. Figure 8(a) shows the typical microstructure of a 319 alloy before SHT, which is composed of needle-like eutectic silicon and eutectic CuAl 2 precipitates. After SHT (Figure 8(b)), it is observed that there is a dissolution of the CuAl 2 precipitates (black arrows) [23,24] and the fine silicon particles are observed within the latter. ere is also an increase in the thickness and length of the needle-like silicon particles. e circle in Figure 8(b) reveals the fragmentation of a silicon particle. e inset in this figure reveals the presence of ultrafine Si particles within the Al-CuAl 2 eutectic indicating the end of solidification. Figures 8(c) and 8(d) depict the microstructures of the Sr-modified samples, before and after SHT for 24 h at 510°C, respectively. As can be seen, the addition of Sr allows for almost complete dissolution of the CuAl 2 phase during the SHT process. A comparison between Figures 9(a) and 9(b) clearly demonstrates the thickening of the Si particles following solutionizing for 24 h at 550°C (a temperature that is 40°C above that used for 319 and 413 alloys). e microstructure of alloy 413 before SHT is shown in Figure 10(a). It is composed of primary silicon, eutectic silicon, and the α-Al 15 (Mn,Fe) 3 Si 2 intermetallic phase (Chinese script). e latter is obtained due to the presence of elements such as Mn and Cr. ese Chinese script particles, in unmodified alloys, are found between the particles of eutectic silicon. Solutionizing the as-cast alloy at 510°C for 24 h, Figure 10(b), caused an increase in the thickness and length of the silicon particles. Another observation that can be made for these samples is that the addition of strontium causes a shift in the Al-Si eutectic point, leading to the disappearance of the primary silicon particles.
e Srmodified sample, after SHT (Figure 10(d)), confirms the presence of the α-Al 15 (Mn,Fe) 3 Si 2 intermetallic inside the α-aluminum dendrites (black arrows) [20]. Apparently, solution heat treatment of the sample (Figure 10(d)) has no effect on the size of the silicon particles.

Etched Samples.
According to Fick's first law [25], atoms throughout the solid are constantly vibrating from thermal energy. At a given temperature, each atom in the slice has a frequency at which it will jump to the adjacent slice, ] with equal probability of jumping in either direction; therefore, the frequency of jump from one slice to the next in a particular direction is ]/2 (Figure 11(a)). Whether an impurity atom occupies a substitutional or interstitial position, the atom is trapped in a periodic potential defined by the lattice.
e probability of an atom jumping from one position to the next increases exponentially with the increasing temperature (Figure 11(b)) [26]. LeClaire and Lidiard [27] proposed a model to explain impurity-diffusion in the fcc system as shown in Figure 12. Figure 13 presents the variation in the size and distribution of the Si particles as a function of heating at 550°C for different solutionizing times. As can be seen, in the nonmodified condition, the eutectic Si precipitated in the form of thin platelets with sharp angles (white circle in Figure 13(a)). Fragmentation was observed after 24 h, Figure 13(b) similar to that depicted in Figure 8(b) and continued with the increase in heating time up to 100 h (see Figure 13(c)) (see arrows), with traces of fragmentation after 400 h-white arrows in Figure 13(d). Figure 13(e) is a highmagnification electron image of Figure 13(d) revealing the clustering of several small platelets in order to form one particle. Partial spheroidization was observed after 1200 h solutionizing at 550°C as displayed in Figure 13 Tables 2-5. Solutionizing for 1200 h led to practically spherical particles (Figure 14(f)). It should be kept in mind that the addition of Sr depresses the eutectic temperature coupled with a shift in the eutectic point to a higher Si content [28].
In the case of 413 alloy that contains 11.7% Si, the rates of fragmentation and coarsening are expected to be much slower than those exhibited by 356 alloy. In addition, the maximum solutionizing temperature is about 510°C compared to 550°C for the 356 alloy. Figure 15 depicts the variation in the Si particle morphology and size with the applied heat treatment. After 1200 h, the unmodified alloy still showed elongated particles (see white arrow in Figure 15(c)), with more spheroidization in the modified alloy (Figure 15(d)). Increasing the solutionizing temperature to 565°C resulted in incipient melting in both 356 and 413 alloys even after a short solutionizing time of 8 h as shown in Figure 16 [29].
Based on intensive SEM examination of all conditions, it seems that there is more than one mechanism for Si particle coarsening as illustrated in Figure 17, before reaching the same microstructure viewed by the optical microscope. Most of the Si particles show the presence of pinholes (broken arrows in Figure 17(f)) that go through the entire particle (broken arrow in Figure 17(e)) or conical cavities (solid arrow in Figure 17(e)) which may need further sectioning to reveal their cause.

Applications.
e ambient-temperature tensile properties of 356-type alloys solutionized up to 400 h are plotted in Figure 18(a) (nonmodified) and Figure 18(b) (Sr-modified) [22]. Both alloys show the same trend in strength variation for    e ductility behavior is also similar in the nonmodified, unmodified, and Sr-modified alloys. Ductility values improve significantly after solution treatment for 8 hours at 550°C compared to the ductility values in the ascast condition for both alloys. e ductility of the unmodified alloy remains constant between 8 and 50 hours of solutionizing, followed by further enhancement to achieve a peak value of 9.7% after 100 hours of solution treatment and remains almost constant up to 200 hours. On the other hand, ductility values of Sr-modified alloy show a continuous enhancement up to 50 hours of solutionizing where the ductility reaches its peak value of 10.5%. e longest duration of solution treatment (i.e., 400 hours) results in reduction in the ductility values of both alloys to about 8%. Figures 18(c) and 18(d) exhibit the shape and distribution of the eutectic Si particles sectioned from a Sr-modified tensile bar solutionized at 550°C for 50 h revealing the difference in 2D (before deep etching) and 3D (same sample after deep etching). As can be seen, due to the collision of the Si particles in Figure 18(d), it is difficult to determine their individual sphericity or the exact particle size (e.g., section A-A or B-B in Figure 18(d)). In addition, the joint particles in Figure 18(d) are not forming a straight line. erefore, it is also difficult to measure their actual length (thick white broken line). Since the 2D micrographs represent the cross-sections of such particles, they appear nearly spherical which explains the reported high standard deviations in Tables 2-5. e tensile curves of Al-Si-Cu-Mg 319-type alloys could be divided into four regions across the range of solution temperatures used.
is is shown schematically in Figure 19 [29,30]. Region I corresponds to the change in tensile properties of the alloy on going from the as-cast to the solution heattreated condition. Region II represents the recommended solution treatment temperature range. Region III represents a continuation of region II until peak properties are reached (spheriodization/coarsening), following which incipient melting begins to occur, while Region IV corresponds directly to the progress of incipient melting with the increase in the solution temperature. Figure 20 shows the microstructure corresponding to the early stages of incipient melting when the tensile bars are solutionized at a temperature just above region III (520°C) revealing the presence of a large number of pores (Figure 20(a)). Figure 20(b) is a high magnification of     Figure 11: (a) Diffusion from a concentration gradient [25]. (b) Impurity atom diffusing along with a periodic potential [26].
(1) A host atom (nearest neighbor to an impurity) jump which does not "dissociate" the impurity from the vacancy (5)    the transient temperature differences within the bar, establishing a system of residual stresses simultaneously (region IV).

Conclusions
Based on the results documented in the present work, the following conclusions may be drawn: (1) Complete spheroidization of eutectic Si particles in terms of achieving individual spherical particles cannot be obtained in most Al-Si-Cu-Mg alloys even after a solutionizing time of 1200 h which contradicts with the theory of Paray and Gruzleski [14] (2) Addition of Sr to 356 alloy (Cu-free) could lead to spheroidization after 1200 h/550°C if the alloy is solidified at a high rate, i.e., 8°C/s (3) Beside the dissolution theory of Ostwald, coarsening of Si particles can as well take place by impingement, fusion, and agglomeration (4) Increasing the Si content makes it difficult to achieve spheroidization, through fragmentation and coarsening (5) Results obtained from observation of deeply etched samples (3D) contradict the observations made from polished samples (2D)

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.