1. Introduction

Lightweight metals such as aluminum and magnesium alloys find special applications when they contain hard ceramic particles like silicon carbide, alumina, zircon sand, and silica sand as reinforcement due to their superior mechanical and wear properties ^{[1,2,3,4,5,6]}. Silica sand which has high hardness, high thermal stability, and high compressive strength is a potential ceramic reinforcement particle for the development of aluminum matrix composites. However, for most application the effect of reinforcing phases, microstructural changes, and particle/ matrix interface on physical and mechanical properties need to be monitored carefully

The development of Al/SiO₂ composites has been conducted mostly on a trial and error basis since the 1960’s. However, these materials have not been employed in any widespread engineering applications. A coherent description of reactions in these composites and how the reactions result in changes in density and mechanical properties is lacking. With a better understanding it may be possible to predict the reactions and control properties of these composites and thereby make them more useful.

Previously, researchers have used stir mixing to incorporate SiO₂ particles into commercially pure Aluminum ^{[6,7,8,9,10,11,12]}. Rohatgi et al. developed pure aluminum matrix composites by incorporating 53-125 µm size SiO₂ particles at 720-790 °C with Magnesium added as a wetting agent ^[11,12]. Composite castings with over 6 wt% SiO₂ could be obtained when the particles were preheated to 700 °C. However, cold or untreated sand particles were rejected by the melt. This was attributed to the presence of absorbed gases or water molecules covering the SiO₂ particles and possibly poor wettability at lower temperatures. Hot spots and hard black scum were observed at temperatures greater than 1000 °C. Analysis revealed a reaction zone rich in Mg at the surface of the dispersed SiO₂ particles. The dispersed SiO₂ particles tended to be concentrated in interdendritic regions of the matrix and did not promote nucleation of α-Al. Optimal results were obtained when a maximum of 3 wt% silica sand was mixed with 3.2 wt% Mg, where the hardness increased four-fold and the abrasion resistance increased by over three times compared to the matrix. Particle incorporation became increasingly difficult and considerable surface oxidation occurred when the silica and magnesium composition ranged as high as 4 wt%. Alumina formation was reported to begin at 800 °C ^[11,12].

Powdered metallurgy (PM) techniques have also been used to create ceramic particulate composites using commercially pure aluminum and its eutectic silicon alloy mixed with up to 25 wt% of 90-300 mm SiO₂ sand ^[13]. SEM observations showed voids around sand particles indicating poor bonding at particles-matrix interface, which was confirmed by tensile testing. The ultimate tensile strength decreased from 184 to 112 MPa with the addition of 20 wt% sand particles, while the hardness remained constant. However, the composite strength increased when magnesium was added to the matrix.

From this experimental data, it would appear that Mg may play two roles in these materials. In the case of solidification processing ^[14], Mg seems to assist in wetting of the SiO₂ particles with the molten alloys. Pai et al. ^[15] suggested that wetting agents could improve wetting of SiO₂ particles by Al alloys by increasing the surface energy of solid, decreasing surface tension of liquid, and decreasing the particle/alloy interfacial energy. Mg, a reactive element, likely fulfills all three conditions for improved wetting and, in addition, it scavenges oxygen from the surface of SiO₂ particles. The second role played by Mg is to improve bonding between the matrix and the SiO₂, which seems clear from the increase in strength of the PM-produced materials. This is likely do to reactions between the Mg and SiO₂ that take place during the elevated temperatures of processing ^[16,17]. However, reactions between Mg and SiO₂ in the processing of these materials have not been thoroughly investigated and the effects of Mg concentration, SiO₂ concentration and processing time on the final composition and mechanical properties of these materials remains poorly understood.

The purpose of the present paper is to study the wetting reaction mechanism in A206 aluminum/silica sand particles composite. Magnesium additions were used to assist in the incorporation of the silica sand into the aluminum alloy. Different levels of Mg and SiO₂ additions were examined and a three-stage reaction model is proposed to describe the changes in density and hardness of this system.

---

2. Materials and Method

Silica sand with particle sizes between approximately 100 and 250 µm was wet collected from Badger Mining Corporation refinery’s dust collectors and had a composition of 99.70% SiO₂, 0.12% Al₂O₃, 0.12% K₂O, 0.04% CaO, 0.02% Fe₂O₃, 0.01% Na₂O, <0.1% MgO, and <0.1% TiO₂.

In order to achieve a precise amount of reinforcement added to the melt and also to facilitate mixing in the molten aluminum, desired quantities of silica sand particles and Mg turnings were encapsulated within thin Al foil in the form of a rolled cylinder with closed ends. The crimped ends of the foil do not form an airtight seal, allowing the removal of moisture and absorbed gases when the SiO₂/Mg packets were preheated for 30 min at 120 °C.

500 gr of Aluminum Alloy A206 (Composition: 4.6% Cu, 0.25% Mg, 0.35% Mn, 0.05% Si, 0.22% Ti, and 0.10% Fe) were placed in a coated graphite crucible (ID 89 mm, Depth 127 mm) and heated using an induction furnace and held at temperatures between 850 °C and 900 °C. Various compositions and processing conditions used to prepare the composites by stir casting are listed in table 1. Once the melt reached the desired constant temperature, the measured quantities of silica particles in the Al foil packets were added to the melt. A vortex is created for various lengths of time (As listed in table 1) in the melt using a graphite stirrer coated with boron nitride revolving at 350 rpm while being held at 60° with respect to the melt surface. A steel permanent mold used to pour the melt to form castings is coated with boron nitride and preheated to 450 °C. Three measurements of density and Rockwell F Hardness were taken for each specimen and the average value is reported as the density of the composite for that specimen. For density measurements, an electronic balance with a resolution of one milligram was used while utilizing the “Archimedes Principle” and using distilled water as the auxiliary liquid to determine the density of composites and matrix alloys with respect to that of pure aluminum. SEM analysis and Energy Dispersive Spectroscopy (EDS) was carried out on selected specimens using a TopCon SM-300 Scanning Electron Microscope at 15 kV voltage with 12 mm working distance to determine morphology and EDX to determine composition.

---

3. Results and Discussion

Various Al-A206/Mg/SiO₂ compositions prepared by stir casting, their processing conditions, density, and hardness values are listed in table 1. It can be seen in figure 1 that in the absence of additional Mg there is no correlation between silica content and hardness or density. The microstructures of the different composites show the reason for this observation. Figure 2a shows that no bonding has taken place between the matrix and the SiO₂ particle. In fact voids can be seen at various locations immediately surrounding the particle. It appears that either no wetting reaction has occurred or that it is very slight. However when additional Mg is added it is clear from figure 2b that extensive reaction have taken place. As shown in Figure 3, it is clear that these reactions lead to increased density and increased hardness as SiO₂ concentration increases.

It is difficult to explain the increase in density unless it is considered that reactions between Al, Mg, and SiO₂ are converting the SiO₂ into a denser oxide. There are several possible candidates. There first possibility is that SiO₂ will be reduced by liquid Al with Si being rejected into the melt according to reaction 1:

$4A{{l}_{(l)}}~+\text{ }3Si{{O}_{2(s)}}\to \text{ }2A{{l}_{2}}{{O}_{3(s)}}+\text{ }3\left[ Si \right]$

(1)

Previous research indicates that reaction 1 occurs slowly and does not result in much Al₂O₃ production ^[1,3]. However, the presence of Mg allows other thermodynamically stable oxides to form depending on Mg concentration according to reaction 2 and 3 ^[4,5].

$2M{{g}_{(l)}}~+\text{ }Si{{O}_{2(g)}}~\to \text{ }2Mg{{O}_{(s)}}~+\text{ }\left[ Si \right]$

(2)

$2A{{l}_{(l)}}~+\text{ }M{{g}_{(l)}}~+\text{ }2Si{{O}_{2(s)}}~\to \text{ }MgA{{l}_{2}}{{O}_{4(s)}}~+\text{ }2\left[ Si \right]$

(3)

At the melt temperatures used for these experiments MgO will only be a stable reaction product if the concentration of Mg in the melt is greater than about 1.7 wt%. The spinel, MgAl₂O₄, is stable in the range from 0.04-1.7 wt% Mg and Al₂O₃ is stable for Mg concentrations below 0.04 wt%. ^[2].

The oxide conversion reaction obviously results in changing the overall density of the composite (${{\rho }_{Al-206}}$= 2.77 g/cm³, ${{\rho }_{Si{{O}_{2}}}}$= 2.3 g/cm³, ${{\rho }_{MgA{{l}_{2}}{{O}_{4}}}}$= 3.579 g/cm³, ${{\rho }_{A{{l}_{2}}{{O}_{3}}}}$= 3.9 g/cm³, ${{\rho }_{MgO}}$= 3.58 g/cm³). Such reactions are supported by the changes in the microstructure that occur from the addition of Mg as is clearly evident in figure 2b.

The reactions thought to be involved in creating this conversion are listed in Table 2 and shown schematically in Figure 4. Initially the Mg concentration in the molten alloy is greater than 1.7 wt% and Mg reacts with the SiO₂ forming MgO. This is responsible for the initial wetting of the particle. In Stage 1 the Mg, being the most highly reactive species, is quickly used up and the Mg concentration falls to below the 1.7 wt% level. In Stage 2 the MgO is no longer stable and Al and Mg from the melt begin to react with it and the SiO₂ to form MgAl₂O_4. This reaction further reduces the Mg concentration in the melt until it reaches 0.04 wt% where it is then in equilibrium with the MgAl₂O₄. This causes the onset of stage 3 in which Al directly reacts with SiO₂ to form Al₂O₃. The mechanism is still a reduction of SiO₂ by Al, but because of the Mg/MgAl₂O₄ equilibrium only Al is allowed to react. While this can occur by an internal precipitation reaction at the MgAl₂O₄/SiO₂ interface, the large difference in densities between Al₂O₃ and SiO₂ makes it likely that cracking will occur in the scale allowing direct contact between the melt and SiO₂ which will act to increase the reaction rate. These various stages are apparent in Figure 2b in which incomplete conversion of SiO₂ to Al₂O₃ and MgAl₂O₄ are present in two particles, while complete conversion has occurred in the other three particles. Notice that cracking has occurred in all of the completely converted particles. Figure 5 shows an SEM micrograph where the indicated regions were examined by EDX. The reinforcement seems to consist of anAl₂O₃ center encapsulated by a magnesium-rich outer surface, which is in turn surrounded by an Al-Si matrix. The matrix is completely lacking in Mg, indicating that the greater than 2 wt% Mg originally present is now mostly present as reaction product surrounding the reinforcement particles. The XRD graph of Al-A206-3Mg-13SiO₂ after 17 minutes mixing time is shown in figure 6 which indicated the presence of Al₂O₃ and MgAl₂O₄ in the final composite.

The SEM and EDX analysis for the composite with the highest SiO₂/Mg ratio shows that the reaction product consists mainly of MgAl₂O₄ with some Al₂O₃. Eutectic Si structures were clearly seen in the matrix away from the reinforcement particles. This is clearly an indication that the SiO₂ first is transformed into MgAl₂O₄ until the Mg level is too low and thereafter Al reacts with SiO₂ to form Al₂O₃, while both reaction reject Si into the melt. It also an indication that the SiO₂/Mg ratio can be an important factor to the density or hardness.

Based on this three stage reaction model it is possible to speculate on the behavior of this system. Stage 3, if limited to an internal precipitation reaction would most likely be the slowest reaction and would therefore be rate limiting. However, if cracking occurs, a more constant conversion with time would be expected. Given enough time all the SiO₂ will convert to MgAl₂O₄ and Al₂O₃ which are much denser oxides. Following from this it would be expected that the density of the composite will increase with increased reaction time until all SiO₂ has been converted to Al₂O₃. Figure 7 shows the change in density with reaction time for differing SiO₂ and Mg concentrations and clearly shows an approximately linear increase in density with reaction time for most cases. Furthermore, it is clear that increasing the Mg concentration results in composites with lower densities for the same reaction time. This is likely due in part to the creation of more MgAl₂O₄, which is the less dense phase.

---

4. Conclusion

Silica sand particles were dispersed in the matrix of cast A206 alloy, modified with Mg as a wetting agent using stir casting, and the effects of the reaction time and SiO₂ content were investigated. Measurements for various aluminum/ silica sand compositions show that in the absence of additional Mg there is no correlation between silica content and hardness or density. This behavior is attributed to weak or lack of bonding between the matrix and the SiO₂ particle. Also, voids are present in the particle matrix interface.

Addition of magnesium leads to increased density and increased hardness of the Al-A206 based composite as SiO₂ concentration increases. Reactions between Al, Mg, and SiO₂ which converts the SiO₂ into a denser oxide is proposed as an explanation for the overall change in density of the composites. A three-stage reaction mechanism is proposed. At the first stage Mg reacts quickly with the SiO₂ forming MgO until the Mg concentration falls to below the 1.7 wt% level. In Stage 2, the melt react with MgO and SiO₂ to form MgAl₂O₄ until the Mg concentration in the melt reaches its equilibrium concentration with the MgAl₂O₄. At stage three, where Mg and MgAl₂O₄ are in equilibrium Al₂O₃ forms as a result of direct reaction between Al and SiO₂. This likely occurs through direct contact of Al and SiO₂ due to cracking of the particles as a result of differences in oxides densities. Based on the proposed reaction model, changes in both physical and mechanical properties of A206-Mg-SiO₂ metal matrix composites may be explained in terms of the base Alloy / SiO₂ / Mg chemistry and reaction times.

---

Acknowledgements

This research is supported by the U.S. Army through grant TACOM - W56HZV-04-C-0784. Reference herein to any specific commercial product, process or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation or favoring by United States Government or Department of Army (DoA). The opinions of the authors expressed herein do not necessarily reflect those of the United States Government or the DoA, and shall not be used for advertising or product endorsement purposes.

---

Conflict of Interest

The authors declare that there is no conflict of interest regarding the publication of this manuscript.

---