Investment casting of semi-solid 6063 aluminum alloy using the GISS process

Investment casting has long been known as a process that can produce complex parts with fine details. However, it has not been used widely for parts that need to be anodized and whose nonuniform color is affected by the type of cast aluminum. Although 6063 aluminum is widely used in color anodizing for decoration purposes, it is almost entirely wrought. Some studies have used aluminum 6063 in cast aluminum, but no investment casting studies have been reported. The objective of this work is to develop a new semi-solid investment casting process using 6063 aluminum alloys to achieve a cosmetic anodized workpiece. To study the feasibility of the gas-induced superheated slurry (GISS) investment casting process, the effects of rheocasting time, mold temperature, and vacuum assistance were investigated. The results showed that the samples produced using the GISS process had a uniform microstructure and less porosity. The uniform color after anodizing and coloring also confirmed these processes. From the results, it can be summarized that the GISS process is feasible for application in aluminum 6063 investment casting for anodizing and coloring purposes.


Introduction
The investment casting process has been known for thousands of years, although it was introduced in the commercial sector only in the late 20th century [1]. The process has many advantages over other casting processes, such as its ability to produce a smooth and high-integrity surface, near net shape, and high dimension accuracy, meaning thin walls and complex shapes can be produced. Aluminum investment casting is not widely used in industry due to its high production cost and limitations in anodizing and coloring aluminum [2].
An Al-Mg-Si series alloy with good extrudability, machinability, and weldability, 6063 aluminum alloy is mainly formed using wrought manufacturing processes, especially extrusion. Anodizing 6063 aluminum alloy can improve corrosion resistance and surface appearance and also fill color below the surface, giving more options for decorative purposes [3,4]. However, it is not easy to cast 6063 aluminum alloy because of its high tendency to hot tear due to its long freezing range, its nonuniform microstructure due to liquid segregation, and the limit of fluidity in conventional casting processes [5,6]. In recent years, semi-solid metal forming has been continuously developed to solve these limitations [7,8].
The semi-solid metal forming process has been developed essentially along two routes: thixocasting and rheocasting. Thixocasting involves heating the fine grains of solid billets to a semi-solid temperature condition with an approximately 30%-50% solid fraction and then forming them into parts. Rheocasting involves converting liquid metals into a semi-solid state with an approximately 1%-20% solid fraction and then injecting them into die cavities. The disadvantage of thixocasting is the cost of the billets, metal loss when reheating the billets, and high cost of the reheating system and forming machine. It is therefore very limited in industrial applications. In contrast, rheocasting is a lower-cost process, which makes it more applicable in industry, and has been extensively researched [9,10].
Although the focus of recent studies has shifted to the rheocasting process, most have centered on cast aluminum [11][12][13][14], and few researchers have investigated the rheocasting of wrought aluminum. In most cases, they have researched the mechanical properties and thermal conductivity in permanent molds [15][16][17][18]. The surface appearance for decorative purposes in investment mold casting has not been studied. Thus, the present work attempts to study the feasibility of semi-solid investment casting 6063 aluminum alloys for a cosmetic anodizing application using the gas-induced superheated slurry (GISS) process. Obtaining cosmetic anodized pieces depends on two factors: (1) the workpiece must be complete and have sharp details, and (2) the color must be uniform and without surface defects throughout the workpiece. Various studies of elements affecting these two factors, such as the mold temperature, rheocasting time, and vacuum assistance, are reported in this research.

Investment flask casting mold
In this process, the pattern was made by injecting commercial wax into a silicone mold, and when the wax had become solid, it was removed from the mold. Figure 1(a) shows the wax pattern in this study, which used an investment flask casting mold. The pattern was attached to suitable gates and risers, as shown in figure 1(b). The metal flask was placed around the pattern. Investment powder was mixed at a ratio of 100 powder and 30 water by weight and then poured into the metal flask. The mold material was allowed to set and air-dry for 6 h. The flask-casting mold was then moved to the furnace to melt the investment pattern out of the mold. The heating rate was 80°C per hour until reaching 700°C, followed by cooling to the casting temperature. For this study, the casting temperatures were 400°C and 300°C. A schematic diagram of the investment flask casting mold preparation is shown in figure 2.

Slurry preparation and casting process
Commercial 6063 aluminum alloy was used in this study. The chemical composition, measured using an optical emission spectrometer (OES), is shown in table 1. The aluminum alloy was melted in a graphite crucible using an electric resistance furnace at 800°C and treated with commercial cleaning flux and degassing flux by purging nitrogen gas through porous graphite for 10 min For liquid casting, a steel ladle was used to scoop the molten metal from the furnace, and we waited until the metal had reached a temperature of 720°C before it was poured into the mold. In the slurry casting process, the GISS technique was used [19] by dipping the porous graphite into the melt at 664°C for 5 and 10 seconds to convert the melted aluminum into a semi-solid slurry with a different initial solid fraction [20]. The slurry was then poured into the casting mold with and without vacuum assistance. The molds were then broken with pressurized water to remove the cast part. Figure 3 shows a schematic of the GISS investment casting process. Figure 4 shows the cast part after removal from the mold. Three sample parts were produced in each condition. The parameters used in this study are summarized in table 2.

Vacuum assistance
The experimental set-up without vacuum assistance is shown in figure 5(a). The flask casting mold was placed on an insulation brick at a considerable temperature. The melts were directly poured into the investment casting mold. For vacuum assistance, after pouring the melts, the steel chamber was placed and the pump opened immediately until 70 cmHg for 30 seconds to form a vacuum chamber, as shown in figure 5(b).

Castability evaluation
To analyze the castability of the samples, we propose in this study the part detail sharpness ratio (PSR). The samples were photographed using a digital camera. The images were sectioned into 10 surface area positions, as shown in figure 6, and then micrographs were taken using Image Tool Software. The PSR was calculated using the following equation:      where S is the numerical index of the pattern detail sharpness; S = 1 is complete with sharp details; S = 2 is filled, but the details are not sharp; S = 3 is filled, but the details are unclear; S = 4 is filled and without details; and S = 5 is unfilled or misrun (the representation of the S index is shown in figure 7); A is the analysis surface  area (mm 2 ); A T is the total surface area analyzed; and i is the surface area analysis position. The casting part has complete details and sharpness, the same as the wax pattern; therefore, assume that the PSR = 1.

Porosity analysis
The density of the casting samples was measured with equation (2), and equation (3) was used to calculate the percentage of porosity (h).
here D L is the bulk density of the samples (g), W d is the dry weight of the casting samples (g), W i is the weight of the casting samples in water (g), h is the percentage of porosity (%), and D s is the standard density of a 6063 aluminum alloy (D s = 2.70 g cm −3 ).

Macrostructure observation and microstructure analysis
The samples were cut into cross-sections, which were divided into three zones. The area of macrostructure observation and microstructure analysis is shown in figure 8. The samples were polished and then etched with Keller's reagent to observe the microstructures. Keller's reagent consisted of 2.5 ml HNO 3 , 1.5 ml HCl, and 1.0 ml HF in 95 ml distilled water. The microstructures were photographed with a camera to investigate the uniformity along the samples. The samples were repolished and etched with Weck's reagent to analyze the microstructures again. Weck's reagent consisted of 4 g KMnO 4 and 1 g NaOH in 100 ml distilled water. The microstructure analysis of the samples was carried out using optical microscopy (Optika model B-383 MET). The average primary α grain was measured using the linear intercept analysis method.

Blasting, anodizing, and coloring
The samples were sent to be mechanically polished by blasting with Ballotini impact media, after which, they were cleaned with ethanol reagent for 5 min They were then placed into a chemical brightness etching reagent, Eplen Al3-GC solution, at 100°C for 2 min and then washed with clean water. The anodizing process was carried out by dipping the sample into sulfuric acid solution (15% by weight of H 2 SO 4 ) with an electric current of 14 volts at 18°C for 15 min Coloring was performed using Clariant Sanodal Gold 4N gold paint, into which the sample was immersed at 45°C for 1 min The surface was sealed using Almite Sealer liquid at a temperature of 100°C for 10 min This process was conducted at GISSCO Company Limited, Thailand.

Castability
The representative photographs of the cast samples processed with different conditions are shown in figure 9. The arrow marks in figures 9(a)-(e) show that all samples processed without vacuum assistance were not completely filled in zone III.

Visual appearance without vacuum assistance
At different liquid casting mold temperatures, Liquid-M3, as shown in figure 9(a), had less filling capacity than Liquid-M4, as shown in figure 9(b). The PSR of Liquid-M3 was 0.38, which was lower than Liquid-M4 (PSR = 0.43).
The flow behavior and the filling capacity are the main factors that make parts incomplete. To understand how these factors can affect the part detail sharpness of liquid molten alloys, Flemings [21] showed that the flow stoppage in the thin channels of an alloy is caused by broken dendrite accumulation near the tip when the solid fraction is critical (f crit ); at this point, the solid particles near the tip develop coherency or interlocking to such a degree that the viscosity dramatically increases and the flow is stopped. Figure 10(a) shows an illustrated representation of liquid metal flowing in the fluidity channel and the mode of stoppage. Arnberg et al [22] suggested that dendrites start impinging and form a network that prevents further flow at a solid fraction of approximately 30-35 pct. Thus, the fluidity length (L f )of the liquid metal cast at its melting point can be written as where v is the velocity of the metal flow front, C is a constant, r is the density of the liquid metal, H f is the heat of fusion of the metal, f crit is the critical solid fraction, T M is the metal temperature, T 0 is the mold temperature, and ( / V A) is the volume per surface area of the casting. At the same liquid metal temperature, T , M it shows that the fluidity length, L , f increases with the increasing mold temperature. In the case of semi-solid slurries, the results of the microstructure characterizations show uniformly dispersed primary solid particles in the slurries before and during the flow in the channel [23]. When the solid particles grow to a size and fraction that yield sufficiently high viscosity, the slurry front stops flowing. Figure 10(b) shows this flow behavior schematically. The fluidity of semi-solid slurries with an initial solid fraction of f 0 can be written as Therefore, for casting semi-solid slurries with the same initial solid fraction of f , 0 it is shown that the fluidity length, L , f increases with increasing mold temperature. The rheocasting time is longer for a higher initial solid fraction, f , 0 and more viscosity makes the fluidity length, L , f lower. Therefore, samples with a lower rheocasting time have better fluidity and a higher PSR. Figure 9(e) shows a photograph of the samples in the G5-M4-Vac conditions that has a PSR of 0.76. No misrun defect was identified on the sample at zone III. In summary, only the sample produced under these conditions fully filled the mold cavities. Campbell [24] described the influence of vacuum assistance on filling a narrow plate of thickness 2r. Assuming that the melt does not wet the mold, filling is possible if the pressure at the interface exceeds the resistance caused by surface tension, as the following equation indicates:

Visual appearance with vacuum assistance
where γ is the surface tension of the melts, P a is the atmospheric pressure, r is the density of the melts, g is the acceleration of gravity, and h is the head of the metal.
In the other hand, without vacuum assistance, the equation is given as follows: where P m is the back pressure due to the residual gasses present in the mold cavity. For the relation indicated in equations (6) and (7), it was found that for vacuum assistance, the atmospheric pressure was allowed to act on the metal via the running system, with the vacuum efficiently drawing the molten metal into the mold.

Microstructure uniformity
The representative microstructures throughout the sample of different conditions are shown in figure 11. For the conventional investment casting, the microstructures of Liquid-M3 and Liquid-M4 were not uniform. It was found that zone III had a larger primary α grain than zones I and II, as shown in figures 11(a) and (b), respectively. At a mold temperature of 400°C, Liquid-M4 had more liquid segregation at zone III than Liquid-M3, as shown by the arrow mark in figure 11(b); these factors are because of the lower cooling rate near to the sprue and gate.
Figures 11(c)-(d) show the microstructures of G5-M4 and G10-M4, the semi-solid castings with different rheocasting times at the same mold temperature of 400°C. As is evident, the microstructures were uniform and there was limited liquid segregation. Only small liquid segregation in G10-M4 was found, as shown by the arrow mark in figure 11(d). For semi-solid casting with vacuum assistance, G5-M4-Vac, the microstructure was uniform throughout the sample, as shown in figure 11(e). In summary, the results show that the primary α grain size of the three zones of the semi-solid slurry was finer and narrower than that of the liquid casting. This may be because the solid particles in the slurries flowed uniformly in the channel, yielding uniform and limited phase segregation in the slurry cast sample. These observations were expected and had been reported in the literature [23]. The uniformity of the primary α grain size affecting the oxide layer in the anodizing process has previously been reported [25]; therefore, the uniformity also affects the reflection of the color. For the color to be uniform along the sample, a uniform microstructure is needed.

Porosity
As shown in figure 14, the percentage porosity of Liquid-M3, Liquid-M4, G5-M4, G10-M4, and G5-M4-Vac was 4.86, 6.59, 1.76, 1.84, and 1.50, respectively. It is clear that the samples produced using liquid casting had more porosity than the semi-solid casting samples, which may have been caused by the turbulent flow of the liquid casting when pouring into the mold. The representative microstructures of Liquid-M4 are shown in figure 14(a). However, with the different rheocasting time and vacuum assistance of the semi-solid casting, the

Polishing marks
Polishing marks are defects that are found in the anodizing process [26,27]. There are many reasons why they are caused during polishing or blasting. One of the reasons is that the blasting media opens the aluminum surface, exposes the internal porosity, and allows the media to be embedded. Both embedded media and porosity are clearly visible as defects after the anodizing and coloring process. A schematic of the embedded media in the surface during the blasting process is shown in figure 15. Figure 16(a) shows a polishing mark defect on the surface before and after coloring in the Liquid-M4 condition. It was found that porosity was distributed as defects all over the surface. These defects were significantly lower in the semi-solid casting because its percentage porosity was lower than that of the liquid casting. Figure 16(b) shows fewer polishing marks on the surface before and after coloring in the G5-M4-Vac condition.
In summary, polishing mark defects from the porosity found in liquid investment casting can be reduced by using semi-solid investment casting.