Investigation of metallurgical properties of Al-Si-Mg casting alloys with integrated computational materials engineering for wheel production

3.1. Computational Materials Engineering (CME) Studies

Computational, experimental and industrial scale studies were carried out for AlSi5Mg0.3, AlSi5Mg0.5, AlSi5Mg0.7 alloys as an alternative to AlSi7Mg0.3 alloy to be used in the wheel production by the low pressure die casting method. The detailed computational and experimental elemental compositions of the alloys are comparatively given in Table 1. As the main alloying elements that change in the alloy compositions, the ratios of Si and Mg, are the ones that directly affect the microstructural and mechanical properties in aluminum casting. The amounts of other alloying elements are the same as the content of the base AlSi7Mg0.3 alloy. As it can be seen from Table 1, there is overlapping between computational and experimental composition data. It shows that the compositions developed by CME studies can be obtained under industrial scale production conditions.

In the CME prediction studies performed with JMatPro and ThermoCalc software, solidification characteristics, mechanical (hardness, yield and tensile strengths) and thermophysical (density, liquid viscosity, thermal conductivity) properties related to temperature, which are critical data related to alloys, especially in the casting process, were studied. Figure 1a shows the variation of liquid viscosity from the casting temperature (700 °C) to temperatures (~550 °C) at which the liquid phase is consumed.

Accordingly, the studied alloys have similar viscosity values towards the temperatures where the liquid phase is depleted while AlSi5 alloys have relatively lower viscosity values at high temperatures. It is a fact that the fluidity generally increases as the composition of the alloy approaches the eutectic. On the subject, initial considerations were that changes in metal viscosity near the eutectic caused this (Flemings et al., 1976Flemings, M.C., Riek, R.G., Young, K.P. (1976). Rheocasting. Mater. Sci. Eng. 25, 103-117. https://doi.org/10.1016/0025-5416(76)90057-4.). However, it is now known that this increase is due only to the solidification characteristic. When the concept of fluidity is evaluated from many aspects, it should not be represented as viscosity. The viscosity of a liquid is physically related to the movements that occur in the liquid (Campbell, 2003Campbell, J. (2003). Castings. 2^nd Edition, Londra: Butterworth-Heinemann.). Therefore, it has a definite value at a specified temperature. The fluidity of the metal depends on two basic factors, the metal properties and the variables of the test performed. Metal-related factors include viscosity, surface tension, surface films, gas content, residues, inclusion, solidification and crystallization. There is a belief that liquid metals take quite different values in terms of viscosity and that poor and slow-flowing metal has high viscosity. In addition, it is an insignificant factor in castings as the fluid viscosity changes little with temperature (Atasoy, 1990Atasoy, Ö.A. (1990). Eutectic Alloys: Solidification Mechanisms and Applications. Istanbul Technical University, Istanbul.). For this reason, as it is expected that the alloys used show close viscosity values, it can be said that the alloys will be an alternative to each other in terms of the casting process.

The graph in Fig. 1b shows the variation of the density of the alloys from the casting temperature (700 °C) to room temperature (25 °C). When metals melt, an increase in volume occurs. In this case, considering the density as the amount of mass per unit volume, it is possible to say that the density of the liquid is lower than that of the solid. In CME-JMatPro analysis, the density values of the alloys at 25 °C are 2.687, 2.685, 2.690, 2.680 g/cm³ for the alloys AlSi5Mg0.3, AlSi5Mg0.5, AlSi5Mg0.7, AlSi7Mg0.3 respectively. In this case, it can be said that there is a small decrease in the density values due to the Si amount in the alloys. The reason why such a small amount of change occurs is because the density is a physical quantity and the predominant element in all alloys is Al, over 91%.

Another parameter that plays an effective role in the casting process is the thermal conductivity coefficient. Figure 1c shows the variation of thermal conductivity of alloys from the casting temperature (700 °C) to room temperature (25 °C). One of the most important factors affecting the service life of the product in Al alloys is the accumulation of thermal stresses in the product. For this reason, ensuring that the thermal gradient in the product is homogeneous during casting is extremely important for the efficiency of the process and product properties. When the studies investigating the effects of Al casting alloy composition on the thermal conductivity were examined, it was concluded that this was directly related to the microstructural properties of the alloy. Accordingly, while the thermal conductivity is lower in alloys with thinner dendritic network structure in aluminum casting alloys, the increase in the ratio of secondary phases in the structure decreases the conductivity. In alloys, a critical solidification rate must be determined in order to achieve the desired thermal conductivity value. According to Matthiessen’s rule for thermal conductivity, increasing the concentration of elements in the solid solution increases the resistance of the overall alloy almost linearly (Yamashita and Hayakawa, 1976Yamashita, J., Hayakawa, H. (1976). Deviations from Matthiessen’s rule in electrical resistivity of nickel-based alloys. Prog. Theor. Phys. 56 (2), 361-374. https://doi.org/10.1143/PTP.56.361.). In this context, Mülazimoğlu et al. (1989) in their study on the effect of silicon and magnesium amount on the thermal conductivity of the alloy in aluminum casting alloys, determined that the electrical and thermal conductivity decrease linearly with the increase of Si or Mg concentrations (Mulazimoglu et al., 1989Mulazimoglu, M.H., Drew, R.A.L., Gruzleski, J.E. (1989). The electrical conductivity of cast Al−Si alloys in the range 2 to 12.6 wt pct silicon. Metall. Mater. Trans. A 20, 383-389. https://doi.org/10.1007/BF02653917.). When the microstructural properties of the alloys are examined, the lowest thermal conductivity is observed in the AlSi7Mg0.3 alloy, which has the thinnest aluminum dendrite network structure at the room temperature. As the amount of alloying element contained in the alloys decreases, the thermal conductivity values increase as expected. This trend shows parallelism at high and low temperatures.

Solidification analysis were obtained for AlSi5Mg0.3, AlSi5Mg0.5, AlSi5Mg0.7, AlSi7Mg0.3 alloys by defining all alloying elements to the Thermo-Calc system and by using the “one axis” command. The graphs showing the changes in the equilibrium phase fractions of the alloys with the temperature, obtained by cooling the alloy to the room temperature from the casting temperature are given in Fig. 2. When the temperature evolution of the different phases is examined, it can be said that the changes in the Si and Mg content of the alloys do not affect the eutectic transformation temperature values marked and it can be seen in Fig. 2. On the other hand, it is clearly seen that it significantly changes the solidification range, which is one of the parameters that are important in casting. The solidification range is the difference between the liquidus temperature at which the solidification process begins and the temperature at which solidification is completed. In alloys with a narrow solidification range, the tendency to form microsegregation in the cast structure is lower. Microsegregation is another consequence of the compositional differences occurring during the solidification of the alloys and is the change of concentration from the core outward in solidifying grains (Lumley, 2011Lumley. R. (2011). Fundamentals of aluminium metallurgy. Wood Publishing Limited, Oxford.). From the phase diagrams, the solidification process of the alloys is analyzed using to the solidus curve under equilibrium conditions. However, stable phase transformations do not occur since there are non-equilibrium cooling conditions during solidification in casting. Therefore, the problem of stratified solidification called microsegregation arises when the solidification range is large. For this reason, alloys with a narrow solidification range and close to eutectic composition are preferred in casting. The solidification ranges of the alloys were determined as 73.1, 58.4, 52.3, 66.0 °C, respectively. Accordingly, the lowest microsegregation tendency is seen in AlSi5Mg0.5 alloy.

3.2. Low Pressure Die Casting Simulation Studies

In materials, the microstructure and mechanical properties are closely linked. The possible phases in the microstructure affecting the mechanical properties of aluminum alloys were determined before casting with the Magmasoft software. The results of the phase distributions of α-Al, eutectic-Si, Mg₂Si and AlFeSi intermetallics in different regions of wheels, depending on alloy composition, are given in Fig. 3.

The numerical calculations of the colored data in Fig. 3 obtained from the Magmasoft casting simulation are given in Table 2 in detail. Accordingly, it is possible to say that as the amount of Si increases, the amount of α-Al phase decreases in both the hub, spoke and flange regions of the alloys when the distribution of α-Al phase in the wheel is examined. This is an expected result as the eutectic composition is approached with increasing Si content and the solidification range is narrowed. On the other hand, the ratio of the eutectic phase increases with the increase of the Si content, inversely to the α-Al phase. Since the alloys are not heat-treated, it is an expected result that the Mg₂Si phase is close to zero in the hub, spoke or flange regions of the wheel. The amount of Mg₂Si phase is between 0 and 0.7%. Iron (Fe) is an undesirable element in Al alloys (Fortini et al., 2016Fortini. A., Merlin. M., Fabbri. E., Pirletti. S., Garagnani. G.L. (2016). On the influence of Mn and Mg additions on tensile properties. microstructure and quality index of the A356 aluminum foundry alloy. Procedia Struct. Integr. 2, 2238-2245. https://doi.org/10.1016/j.prostr.2016.06.280.). The presence of iron compounds in the microstructure and their formation in needle-like morphologies adversely affect the mechanical properties (Narayanan et al., 1994Narayanan, L.A., Samuel, F.H., Gruzleski, J.E. (1994). Crystallization behavior of iron containing intermetallics compounds in 319 aluminum alloy. Metall. Mater. Trans. A 25, 1761-1773. https://doi.org/10.1007/BF02668540.; Ji et al., 2013Ji, S., Yang, W., Gao, F., Watson, D., Fan, Z. (2013). Effect of iron on the microstructure and mechanical property of Al-Mg-Si-Mn and Al-Mg-Si diecast alloys. Mater. Sci. Eng. A. 564, 130-139. https://doi.org/10.1016/J.MSEA.2012.11.095.). According to the casting simulation results, the amount of undesirable compounds formed by Fe with Al and Si in the microstructure is insignificant. In all regions of the wheel, the amount of AlFeSi intermetallic phases varies between 0 and 0.4%. These values are acceptable values to produce Al-Si based alloy wheels by the LPDC method.

The distributions and amounts of macro and micro porosities on the wheel were determined with the LPDC module of Magmasoft software. Also, the variation of the secondary dendrite arm spacing values along the wheel section, depending on the chemical composition, was analyzed as given in Fig. 4. In addition to the microstructural properties, the variation of the mechanical properties according to the regions of the wheel in alloys with different compositions was analyzed by the casting simulation. The visual and numerical results are given in Fig. 5 and Table 3, respectively. Depending on the directional solidification, the minimum and maximum values of the microstructural (macro, micro-porosity, SDAS values) and mechanical (tensile, yield strengths and elongation) properties are given in the hub, spoke and flange regions of the wheel. Accordingly, no macroporosity formation was observed in the hub and spoke region of the wheels in the alloys. On the other hand, the probability of macroporosity in the flange areas of the wheel is 0-7%. When the distribution of macroporosities in the flange region is examined (blue colored regions in Fig. 4), it is seen that it is similar to the AlSi7Mg0.3 alloy. The simulation results of the distribution and probabilities of the microporosities (regions marked in red) on the wheel are similar to the macroporosity results. In this case, there is no possibility of microporosity in the hub areas of the wheels in AlSi5 and AlSi7 alloys. For AlSi5Mg0.3, AlSi5Mg0.5, AlSi5Mg0.7 alloys, this probability is minimum 2.6% and maximum 3.1% in the spoke regions, On the other hand, there is no possibility of microporosity in the spoke of AlSi7Mg0.3 alloy according to the LPDC simulation results. For all alloys there is a possibility of microporosity in the flange regions, but it is very low and can be neglected.

Another factor affecting the microstructure and thus the properties in Al casting alloys is the secondary dendrite arm spacing (SDAS) values. As this value decreases, it is expected that the mechanical properties of the alloys will improve. When the analysis results given in Fig. 4 are examined, the SDAS value decreases from the hub to the flange for the alloys. The section thickness of the wheel decreases from the hub to the flange. Accordingly, the cooling rate increases and a microstructure with fine grains emerges in the flange region.

In the automotive industry, tensile tests are carried out in 3 different regions, namely the spoke, inner and outer flanges. According to the automotive standards, the minimum yield strength value required for a wheel that has not been heat treated in the spoke, inner and outer flange is 80 MPa. According to the Magmasoft analysis results (Table 3), it is clearly seen that this value is exceeded in all parts of the wheel, except for AlSi5Mg0.3 alloy. In AlSi5Mg0.3 alloy, this yield strength value is ~75 MPa in the hub, ~76 MPa in the spoke and ~78 MPa in the flange. The casting process is a difficult process to keep under control and it is possible to say that this alloy has achieved the standard value. The yield strength values after the heat treatment should also be considered since the heat treatment has not been applied yet.

When the alloys are evaluated according to the increasing Mg amount, the improving effect of the Mg element on the mechanical properties is clearly noticeable. For example, while the yield strength value obtained in the spoke of the AlSi5Mg0.3 alloy is ~76 MPa, this value is ~110 MPa in the same region for the AlSi5Mg0.7 alloy. Tensile strength values are similar to the tendency of yield strength ones. As it is seen in the virtual simulation results, the highest yield strength was recorded in the flange of the AlSi5Mg0.7 alloy. Another important mechanical property tested on wheels in the automotive industry is the elongation. From the computational elongation data, it can be said that increasing the Mg amount affects the elongation in alloys negatively. In addition, the elongation closest to the AlSi7Mg0.3 alloy was observed in the AlSi5Mg0.3 alloy in both hub, spoke and flange.

Since it is not possible to obtain color visual results related to the hardness analysis from the Magmasoft casting simulation, the hardness values in different parts of the wheel were calculated using the “UTS (MPa)=3.45xHB” relationship between the ultimate tensile strength (UTS) and Brinell hardness (HB). In the calculations, a single value was obtained instead of the minimum and maximum values, since the computationally reached maximum tensile strength value in the relevant region was used. When the hardness data obtained by the ICME studies were examined, an increase was observed in hardness values from AlSi5Mg0.3 alloy to AlSi5Mg0.7 alloy, depending on the increasing amount of Mg. For example, while the AlSi5Mg0.3 alloy obtained a hardness value of ~43 HB in the spoke, a hardness of ~53 HB was achieved in the same region in AlSi5Mg0.7 alloy.

3.3. Test and Characterization Studies

In this section, the material data obtained from ICME and the data obtained from laboratory-scale testing and characterization studies are given comparatively. For macrostructural examinations, images were taken from the hub of the sample with a stereo microscope for each alloy composition. Figure 6 shows stereo microscope images of the studied alloys. In Magmasoft casting analysis carried out before the industrial-scale studies, the probability of macro porosity was 0% in the hub and spoke, while this probability was predicted as 7% in the flange. It is possible to say that the results at the laboratory scale macrostructural studies are in parallel with Magmasoft analysis. Accordingly, while casting defects such as micro-level and acceptable amounts of porosity were observed in the samples of the alloys in macrostructure examinations, no macrostructural defects were found.

After the macrostructural evaluations, detailed microstructural studies were carried out with the image analysis systems in an optical microscope. Microstructure images, given in Fig. 7, were recorded from the hub, spoke and flange regions of each sample. In addition, phase analyzes were performed from the hub and the results are given in Fig. 8. From the optical microscope images, it is clear that the microstructure of the alloys consists of two main phases, α-Al and eutectic represented in light and dark gray, respectively. The eutectic phase is located between the primary α-Al dendrites. This is the microstructure of hypoeutectic (Si˂ 12.6 wt.%) Al-Si alloys. For the alloys, the microstructure is refined from the hub to the flange and the size of the main phase, α-Al dendrites, is reduced. In addition, as the Si amount of the alloys increases, the amount of the eutectic phase increases, while the amount of the α-Al phase decreases. Moreover, it is concluded that the grain refinement and modification processes applied respectively to the alloys before the casting process by using AlTi5B1 and AlSr10 master alloys are successful when the microstructures are examined. With the modification process, the morphology of the eutectic phase was transformed from a needle-like structure to a fibrous structure (Timpel et al., 2012Timpel, M., Wanderka, N., Schlesiger, R., Yamamoto, T., Lazarev, N., Isheim, D., Schmitz, G., Matsumura, S., Banhart, J. (2012). The role of strontium in modifying aluminium-silicon alloys. Acta Mater. 60 (9), 3920-3928. https://doi.org/10.1016/j.actamat.2012.03.031.).

The phase analysis given in Fig. 8 are based on the representation of different phases with different colors in the optical microscope and the determination of the phase percentages in the relevant area. In the analysis, green color was chosen for the α-Al and dark blue color was selected for the eutectic phase and measurements were taken 5 times from the hub for each alloy. The data in Fig. 8 are the average values obtained from 5 measurements. Accordingly, α-Al phase was determined as 93.6%, 94.8%, 92.7% and 91.3% in AlSi5Mg0.3, AlSi5Mg0.5, AlSi5Mg0.7, AlSi7Mg0.3 alloys, respectively. The eutectic phase content was recorded as 6.4%, 5.2%, 7.3%, 8.7% in AlSi5Mg0.3, AlSi5Mg0.5, AlSi5Mg0.7, AlSi7Mg0.3 alloys, respectively. The phase percentages do not differ according to the hub, spoke or flange on a wheel. It is only the percent of the phases that varies by region. In these analyzes, only the areal percentage of the phases represented by different colors in the relevant image area is calculated. For this reason, the samples were analyzed only from the hub. When the results obtained from the Magmasoft casting simulation and phase analysis using optical microscope image analysis techniques are compared with the experimental data, it is seen that the trends are similar. In both computational and industrial scale evaluations, the amount of eutectic phase increases and the amount of α-Al phase decreases from AlSi5Mg0.3 to AlSi7Mg0.3 alloy.

Another important result derived from the microstructure evaluations is related to Mg element. Since the alloys are still in as-cast condition, the Mg element is dissolved in the microstructure. When heat treatment is applied to the alloys, Mg and Si elements will react at the appropriate temperatures and conditions to form the Mg₂Si phase in the microstructure which contributes to the increase in strength. Therefore, Mg₂Si phase was not found in the microstructure, as seen in the computational and experimental results.

The yield strength values obtained from the tensile test and ICME studies results are given in Fig. 9. The tensile test was carried out on the spoke, outer and inner flange specimens in accordance with automotive industry standards. The minimum yield strength values required for the base alloy AlSi7Mg0.3 in the spoke and flanges is 80 MPa. In the tests and analyzes, the yield strength of both the spoke and flange for AlSi5Mg0.3 alloy remained below 80 MPa. On the other hand, other alloy compositions showed a yield strength above the limit value of 80 MPa in both the spoke and flange regions, and exhibited a character that could be an alternative to AlSi7Mg0.3 alloy in terms of strength. Among the studied alloys, the highest yield strength was recorded in AlSi5Mg0.7. When the results are interpreted numerically, the yield strength value in the spoke is 94.9 MPa in the tensile tests for the AlSi5Mg0.5 alloy, while this value is 92.3 MPa in the virtual data obtained with the Magmasoft casting simulation. In the case of a traditional and dynamic method such as casting, a very low difference of 2.8% was recorded between the numerical and experimental yield strength values. Furthermore, there is no outer and inner flange distinction for these regions in terms of yield strength in a wheel. Since these regions are very close to each other in terms of cross-section, the values recorded separately for the outer and inner flange regions from the tensile tests were averaged and compared with the simulation values. Similarly, when the yield strength values of the flange region are examined, the yield strength value obtained as a result of the tensile test at the flange for AlSi5Mg0.7 alloy is ~100 MPa, while the virtual result is ~111 MPa. In this case, the difference between experimental and computational data is again extremely low.

Another important result of the tensile test is the elongation. The elongation values obtained for the spoke, inner and outer flange regions of an unheated wheel produced from AlSi7Mg0.3 alloy in the automotive sector were determined as 5.5%, 14.0% and 12.0%, respectively. Figure 10 shows the elongation values of the related alloys. Since a large elongation was recorded in the different regions of the rim with the simulation studies, the average of these values recorded for the hub, spoke and flange regions was taken as the basis to compare with experimental data. According to this, the results above the required value of 5.5% were obtained in AlSi5Mg0.3, AlSi5Mg0.5, AlSi5Mg0.7 alloys, 6.4%, 7.1% and 5.9%, respectively, on the spoke. In AlSi5Mg0.3, AlSi5Mg0.5, AlSi5Mg0.7 alloys, results were obtained above the required value of 5.5%, 6.4%, 7.1% and 5.9%, respectively, on the spoke. On the other hand, higher values were recorded for the outer and inner flanges, with only 15.2% and 13.1% for the AlSi5Mg0.3 alloy, respectively, compared to the base alloy AlSi7Mg0.3 alloy. However, the desired elongation value could not be reached in the outer and inner flange regions of other alloy compositions.

When the tensile test results calculated with the ICME are examined, the values obtained in the simulated environment vary between 3% and 7%, regardless of the wheel region, while the actual data varies between 4% and 15%. In this case, it is not correct to compare virtual and real data in terms of elongation. The reason for this is that the Magmasoft casting simulation software, which can calculate the elongation, is not yet efficient enough for low pressure die casting process design.

After performing the tensile tests of the alloys, the results of the hardness measurements carried out to evaluate the mechanical properties are given in Fig. 11. As it was stated in the “Materials and Method” section, the hardness values of the visualized regional of the wheel cannot be reached with the Magmasoft casting simulation software. For this reason, the equation of ultimate tensile strength and Brinell hardness “UTS (MPa)=3.45xHB” in metallic materials was used. As it is seen in Fig. 11, the results of computational and physical comparison data with very low standard deviation value show the usability and efficiency of ICME approaches in terms of low pressure die casting process efficiency in aluminum alloys. For example, while the virtual calculations for the base alloy AlSi7Mg0.3 has a hardness of ~50 HB in the flange region, this value is ~49 HB at the same point according to the experimental hardness test results. To give another example, while the hardness value of 41.5 HB was reached in the hub of AlSi5Mg0.3 alloy with the computer aided calculations, a value of 43.3 HB was obtained from the experimental hardness measurements. The difference between the computational and experimental values is 4%, which is extremely low and negligible for the casting process. It is predicted that the alloys can be an alternative to AlSi7Mg0.3 alloy by utilizing ICME approaches in terms of hardness values obtained in all the regions of a wheel.