Influence of Continuous Casting Speeds on Cast Microstructure and Mechanical Properties of an ADC14 Alloy

To improve the mechanical properties of the casting alloys, various attempts have been made to use alternative casting technologies. The Ohno continuous casting (OCC) process is a unidirectional solidification method, which leads to high-quality cast samples. In this study, the Al-Si-Cu-Mg alloy was cast at casting speeds of 1 mm/s, 2 mm/s, and 3 mm/s, by the OCC process. The aim of this study is to investigate the effects of the casting process parameters, such as casting speeds and cooling conditions, on the crystallization characteristics and mechanical properties of OCC-Al-Si-Cu-Mg alloy. Particularly, secondary dendrite arms spacing of α-Al dendrites in OCC samples significantly decreases with increasing casting speed. Moreover, the mean tensile strength of the samples, produced at the highest casting speed of 4.0 mm/s, is significantly higher than that for the samples produced at a casting speed of 1.0 mm/s.


Introduction
The reductions in automotive exhaust greenhouse gases, such as carbon dioxide (CO2) and nitrogen oxides (NOx), are strongly required for environmental reasons. Light-weighting of vehicles presents an opportunity for cutting greenhouse gas emissions. Aluminum (Al) alloys are one of the light weighting automotive materials, which meets vehicle safety and performance requirements [1]. Therefore, automotive parts made of heavy steel have been replaced with Al alloys. Particularly, the hypereutectic Al-Si alloy family is widely used in the automotive industry because of its high strength, good castability, and low density.
The Al-Si-Cu-Mg alloy, used in this study, is particularly the commercial hypereutectic Al-Si16-Cu4-Mg0.6 alloy (JIS (Japanese Industrial Standards) ADC14). This Al-Si-Cu-Mg alloy has been also widely used for various automotive parts, such as linerless engine blocks, pistons, pumps, and compressors. Furthermore, there has been increasing use in the cast Al-Si-Cu-Mg alloy for automotive parts, over the world. The cast Al-Si-Cu-Mg alloy microstructure consists of mainly coarse α-Al, primary silicon (Si), and needle-shaped eutectic phases, which significantly affects the mechanical properties of Al-Si-Cu-Mg alloy [2]. In particular, it should be noted that the mechanical properties of Al-Si-Cu-Mg alloy are also directly affected by α-Al dendrite size, as indicated by dendrite arms spacing (DAS). However, the use of this Al-Si-Cu-Mg alloy, to replace heavy steel in automotive industry, has been restricted because of their lower strength and lower ductility.
Al-Si alloy family has been investigated using a unique continuous casting technique proposed by Ohno, which is known as the Ohno continuous casting (OCC) process. The OCC process is a unidirectional solidification method, and this casting process provides phase control and texture control [3], which can lead to easy control of microstructure size [4] and crystal orientation [5]. Some researchers have investigated the mechanical properties of OCC-Al alloys, showing excellent tensile and fatigue properties. This can be explained by that their unidirectional microstructures, low defect density, and uniformly oriented lattice structure [6][7][8]. Therefore, it is believed that the OCC process is useful in the automotive industry.
These days, a better understanding of how alloy solidifies at different casting conditions, involving cooling rates is required in order to improve mechanical properties of cast Al-Si-Cu-Mg alloys as automotive parts. However, very little investigation has been reported, which includes the interpretation of cooling conditions and mechanical properties for the OCC-Al-Si-Cu-Mg alloy on a metallurgical basis.
Therefore, the Al-Si-Cu-Mg samples were cast at casting speeds of 1 mm/s, 2 mm/s, 3 mm/s, and 4 mm/s via the OCC process in this study. In addition, the influences of microstructures, such as α-Al phases, primary Si, and DAS on the mechanical properties of cast Al-Si-Cu-Mg alloys were investigated. The aim of this work is to investigate the effects of casting speeds and cooling conditions on the α-Al dendritic grains-growing and tensile properties of OCC-Al-Si-Cu-Mg alloy. Table 1 shows the chemical compositions of Al-Si-Cu-Mg alloy used in this study. The Al-Si-Cu-Mg samples were prepared by the OCC process. In this study, the samples were also prepared by the gravity cast (GC) process for comparison with the OCC process. Figure 1a shows a schematic drawing of a horizontal-type OCC arrangement, consisting of a melting furnace, a heated graphite mold with a diameter of 5 mm, a graphite crucible, a cooling device, and a dummy rod for withdrawal of the cast sample. Approximately 0.4 kg of Al-Si-Cu-Mg ingot was placed in the graphite crucible for melting, with the graphite mold heated to approximately 910 K, which is just above the liquidus of the Al-Si-Cu-Mg alloy. The melted Al alloy in the crucible was fed continuously into the mold through a runner. A schematic drawing of conventional GC arrangement, consisting of an electric furnace, a crucible, and a metal mold, is shown in Figure 1b for comparison. The Al-Si-Cu-Mg ingot was also placed and melted in the crucible at approximately 910 K for 1 h, using the electric furnace. Then the molten alloy was solidified in the wide rectangular mold.

Materials
To obtain various microstructural characteristics, the casting operation of the OCC process was carried out at different speeds, from 1.0 mm/s to 4.0 mm/s. In this study, the round-rod samples were produced via the OCC process with a diameter of 5 mm and a length of approximately 1 m.

Metallurgical Analysis
Some of the OCC and GC samples were sectioned to observe microstructures using optical microscope (OM). The cut surfaces were first metallographically polished using up to a 4000 gritsized silicon carbide paper and were subsequently buff-polished using an Al2O3 powder with an average diameter of 0.3 μm and a colloidal SiO2 solution.
The DAS were also investigated from optical micrographs of cross-section of the cast samples. In particular, the secondary dendrite arms spacing (SDAS) was determined by several optical micrographs and applying an image analysis software. The total spacing from the first to the last arm in a certain area was firstly measured to calculate the SDAS. Consequently, the SDAS was obtained from the following equation: where L is the measuring length in μm, n is the number of existing dendrite arms in a certain area, and V is the micrograph magnification. Then, an average value of the results was calculated after repeated measurements in different areas.
In addition, some of the cast rod was sectioned parallel to the withdrawal direction and scanning electron microscope (SEM)-based electron back-scattered diffraction (EBSD) analysis was performed to develop a more quantitative view of the microstructures. The cut surfaces were also buff-polished and the orientation data obtained from the EBSD scans were analyzed using an Orientation Imaging Microscopy (OIM TM ) analysis software (version 7.2, EDAX Inc., Mahwah, NJ, USA). Moreover, after the tensile tests, the fracture surfaces were also observed to analyze the material defects, using SEM.

Mechanical Properties
The hardness of longitudinal cross-sections of OCC samples was measured using a Vickers hardness tester at a load of 9.8 N for a holding time of 15 s. The hardness of GC samples was also measured under the same conditions. Nine different points were measured for each sample. The highest and lowest values were discarded, and the hardness of each sample was determined using the average of the remaining seven values.
Moreover, the tensile tests were carried out using an electro-servo-hydraulic system with a crosshead speed of 1 mm/min in air at room temperature. Dumbbell-shaped round specimens (4 mm in diameter and 30 mm in length) were used as shown in Figure 2. The three specimens were prepared and tested for each tensile test. Load and strain were measured using a load cell attached to the machine and a foil-type strain gage attached to the gage section of the specimens, respectively. The tensile strength and strain of the specimens were obtained from the tensile stress-strain curve.    Figure 4 shows middle parts of cross-sectioned OCC samples. In this study, no sharp difference is observed between upper and lower parts of the samples at any of the casting speeds, although upper parts were cooled directly by spray water and cooled at the fastest rate (Figure 1a). It can be explained by that the diameter of cast rod is too small to show large difference in the cooling rate, between upper and lower parts. In addition, optical micrographs of longitudinal cross-sections are also shown in Figure 5. It is clear from the results at higher casting speeds (Figure 5b-d) that the α-Al dendrite cells grow in the withdrawal direction. No sharp difference was also observed between upper and lower parts of cast samples, regardless of the casting speeds. The grain structure transition is also observed in longitudinal cross-sections with changes in casting speeds. The quantitative analysis of dendritic structures can be conducted by measurements of DAS. As shown in Figure 5, it is clearly observed that the dendritic structures are refined with increasing casting speed. It is considered that higher cooling rates during solidification induce a higher degree of refinement in dendritic array.

Microstructural Characteristics
EBSD analysis was also carried out in order to analyze morphologies of the dendritic structures, including SDAS. All observations were obtained from longitudinal cross-sections of the samples as shown in Figure 6. It can be more clearly observed from both image quality (IQ) maps and corresponding grain boundary (GB) maps that SDAS of α-Al dendrites significantly decreases with increasing casting speed. SDAS is one of the most important factor, which affects the mechanical properties of Al-Si alloy family. The effects of SDAS and grain size on mechanical properties are similarly obtained for the related Al-Si alloys [9]. Figure 7 shows the relationship between the casting speeds and SDAS. SDAS of α-Al dendrites in OCC samples significantly decreases with increasing casting speed. Moreover, SDAS of OCC samples is smaller than that in GC samples, regardless of the casting speed. The mean SDAS is approximately 4.0 μm for the high casting speed of 4.0 mm/s and 8.7 μm for the low casting speed of 1.0 mm/s, resulting from the longitudinal cross-sections. It is considered that the changes in SDAS are due to a different cooling rate, which lead to different solidification behavior.   As shown in Figure 8, the cooling rates (CR) of the casting samples were also estimated from the following equation [10]: CR = 2 × 10 4 × SDAS −2.67 (2) The cooling rates significantly increase with increasing casting speeds for OCC samples. Furthermore, the estimated cooling rates of OCC samples are higher than those of GC samples, regardless of the casting speed. It can be explained by that OCC samples have much smaller volumes with cooled directly with sprayed water.

Mechanical Properties
As shown in Figure 9, it is clearly observed that the hardness of OCC samples significantly increases with increasing casting speeds, showing similar behaviors with those of CR plotted in Figure 8. Grain boundaries generally act as barriers to dislocation motion. As mentioned in Section 3.1, the effects of SDAS and grain size on mechanical properties are similar for the Al-Si-Cu-Mg alloy. Consequently, it is considered that the refinement of SDAS, with increasing cooling rate, leads to increase in hardness. Moreover, the mean hardness of OCC samples cast at speeds over 3 mm/s exceeds that of GC sample.  In addition, Figure 10 shows the variations of ultimate tensile strength as functions of casting speeds. It can be observed that the mean ultimate tensile strength increases at higher casting speeds. These results from the tensile tests are similar to the hardness results shown in Figure 9. The mean tensile strength shows a maximum for the samples produced at the highest casting speed of 4.0 mm/s, which is about 70% higher than that for the samples produced at a casting speed of 1.0 mm/s. Moreover, the mean tensile strength of OCC samples cast at 4.0 mm/s is also much higher (by more than 2 times) than that of the GC samples. The variations of fracture strain as functions of casting speeds are also shown in Figure 11. No sharp difference is observed with a changing casing speed. Nevertheless, the fracture strains of OCC samples are slightly higher compared to that of GC samples, regardless of the casting speeds.
It is believed that such improvements of mechanical properties of OCC samples can be explained by the presence of unidirectional microstructures and refined dendritic array. Particularly, the improvements of the tensile strength of OCC samples are owing to the refinement of SDAS, which may be the obstacles to the movement of dislocations. SDAS is particularly a significant factor determining the mechanical properties of cast Al-Si-Cu-Mg alloys due to grain-boundary hardening, although the mechanical properties are also affected by other microstructural characteristics, such as eutectic and Si phases.  Figure 12a. On the other hand, from the microstructural observations, the fine casting defects (pores or inclusions) are detected in all of OCC samples. The casting defects could give rise to the stress concentration, leading to the crack formation during the tensile test [11]. It should be noted that the large pore, among the casting defects, is the most responsible for crack formation. Hydrogen (H), as the only gas capable of dissolving to significant quantities in an aluminum melt, is the main factor influencing gas porosity. It is considered that the large pore of GC sample is generated by H gas during the solidification in this study. This is because the morphology of the large pore on the fracture surface does not seem to be a shrinkage pore with an irregular shape, but a gas pore of a spherical shape [11]. It is considered that the large gas pore in the GC sample is generated by the slower cooling rate compared to the OCC samples. Furthermore, not only SDAS, but also the casting defects are smaller and better distributed at a higher casting speed. Consequently, the lower strength and ductility of GC alloys are assumed to result from not only enlarged α-Al and SDAS but also internal defects in this study.

Summary
In this study, the effects of casting speeds and cooling conditions on the cast microstructures and mechanical properties of Al-Si-Cu-Mg alloy were investigated. The primary findings of this study are the following: 1) The microstructures, particularly those of the sizes of α-Al grains, differ among the samples, which resulted from the different cooling rate. SDAS of α-Al dendrites in OCC samples significantly decreases with increasing casting speed. Moreover, SDAS of α-Al dendrites in OCC samples is smaller than that in GC samples, regardless of the casting speed.