1. Introduction
Magnesium alloys have been applied as an alternative for plastic moldings in body frames for electronic equipment due to their light weight, high rigidity, good heat dissipation, high damping capacity, and good electromagnetic shielding effectiveness. The thixomolding process has been widely used to improve the performance requirement of magnesium alloys.1) In a previous study, we reported that fixing 0.1 mass% carbon nanoparticles on the AZ91D magnesium alloy chip surface during the thixomolding process drastically improved its mechanical property and fluidity.2,3) Currently, magnesium alloys are used in the frames of various electronic equipment such as note PC and tablets to reduce the size and weight of the equipment.4) The thixomolding process is suitable for molding thin magnesium sheets with complex shapes, and it has been applied to manufacturing body frames of electronic equipment.
Reducing the weight of a carbody is necessary to reduce the emission of carbon dioxide from cars to achieve a low-carbon society in the automobile industry. Magnesium alloys have been used in place of steel and iron to reduce the weight of cars. Consequently, the demand of magnesium alloys in the automobile industry has been increasing.5) To expand the future application of magnesium in automobiles, it is necessary to increase the wall thickness of the molding material. However, casting defects such as blow holes caused by the solidification contraction and void caused by the aerial entrainment limits the application of thick magnesium alloys. Therefore, it is necessary to improve the mechanical properties including the fatigue property of magnesium alloys produced by the thixomolding process, for applications in automobile parts, as strict durability is required in transportation equipment.
In this study, we prepared specimens of AM60B and AZ91D with different thicknesses using the thixomolding process. The effects of fixing carbon to the AM60B and AZ91D magnesium alloy chip surface on mechanical properties were examined. In addition, the void control mechanism by the carbon was discussed.
3. Results and Discussions
3.1 Effect of carbon fixed to the surface of magnesium alloy chips in the thixomolded specimens
Figure 2 shows the images of the carbon fixed and carbon free AM60B magnesium alloy chips. The carbon free chips had a shiny silver white color, but carbon fixed chips were slightly black and the luster had been lost. Although the amount of carbon nanoparticles added was as small as 0.1 mass% by weight, the carbon nanoparticles fixed to the surface significantly changed the appearance of the carbon fixed chips. The carbon fixed and carbon free chips of the AZ91D alloy were also similar in appearance to those of the AM60B alloy. The difference in the color tone of the carbon fixed and carbon free chips disappeared after thixomolding as the silver-white color with metallic luster of the chips reappeared as shown in Fig. 1. This result indicates that the addition of carbon nanoparticles had no effect on the appearance of the thixomolded specimens. In addition, no remarkable defects such as porosity were observed on the surface of the carbon fixed and carbon free specimens after lathing.
Figure 3 shows the secondary electron images of the surface of the AM60B and AZ19D specimen at the R part after lathing. In the carbon free and carbon fixed specimens, a tool mark based on lathing was observed, but remarkable casting defects such as voids were not observed. The carbon fixed and carbon free AZ91D specimens also exhibited similar surface conditions. Since the surface roughness affects the fatigue strength,6) the surface roughness of each specimen at the R part was measured. The surface roughness Ra of each specimen was in the range of 0.540 ± 0.030 µm, thus, confirming that there was no significant difference in surface roughness of the carbon fixed and carbon free specimens. It is expected that the residual stress due to the cutting process affects the fatigue strength, however, since all the specimens used in this experiment were cut under the same conditions, the fatigue strength was measured as a function of the effect of carbon fixing.
Figure 4 shows the cross-sectional macroscopic photographs of each specimen at the R part. Voids of various sizes were observed in each specimen, and it was presumed that these voids were generated by solidification shrinkage and entrainment of air due to the thixomolding process. In both the AM60B alloy and the AZ91D alloy, the number of voids in the thick part with a diameter of 10 mm (chucking position) was larger in number and size than that in the R part with a smaller diameter. In addition, the number and size of the voids in the R part with a diameter of 7 mm were larger than those in the R part with a diameter of 4 mm. Thus, confirming that voids tend to occur in the thick parts during the thixomolding process. Furthermore, these voids were located towards the center than at the outer periphery in contact with the mold. The distribution of these voids was consistent with the results of the visual observation and SEM observation (Fig. 3) at the R part.
To determine the effect of carbon fixing on void formation, the number and size of voids formed in the AM60B and AZ91D alloys were compared. The results show that fixing carbon to the chip surface in the thixomolding process reduced the number and size of voids generated in both alloys.3) The number of voids generated in the carbon-free AM60B alloy were compared to those generated in the carbon-free AZ91D alloy. The result showed that the number of voids in the AZ91D alloy was larger than those in the AM60B alloy. This result suggests that the aluminum present in the AZ91D alloy is closely involved in void formation.
Figure 5 shows the EBSD analysis results near the outer circumference in the cross-section at the R part of the carbon free and carbon fixed AM60B and AZ91Z alloy. In both the carbon free AM60B alloy and the AZ91D alloy, the grain sizes in the R part with a diameter of 4 mm was finer than those of the 7 mm diameter. This is attributed to the fact that the cooling rate of the R part with a diameter of 4 mm is faster than that of R part with a diameter of 7 mm. In the carbon fixed AM60B and AZ91D alloys the grain sizes were refined. However, the AZ91D alloy had a finer grain size than the AM60B alloy. In particular, the grain size of the carbon fixed AZ91D alloy in the R part with a diameter of 7 mm was the finest in this study. To explain the mechanism of grain refinement by carbon fixing to magnesium alloy, it has been reported that the fixed carbon acts as a precipitation nucleus to refine the crystal grains.7) Our results confirm this mechanism. However, the result that the type of alloy influences the crystal grain refining suggests that the fixed carbon does not directly act as a precipitation nucleus. Therefore, it can be suggested that the reaction between carbon and the components in the alloy acts as precipitation nuclei. In the future, more detailed examination on this point will be discussed.
Table 2 shows the 0.2% proof stress, tensile strength, and elongation of each alloy obtained from the tensile testing. The result shows that carbon fixing improved the 0.2% proof stress, tensile strength, and elongation of each alloy, indicating the importance of carbon for the tensile properties. However, the effect of carbon fixing on the mechanical properties of AZ91D alloy was greater than that of the AM60B alloy, and it depended on the chemical composition of the alloy. The effect of carbon fixing on suppressing voids (as shown in Fig. 4) and refining crystal grains (as shown in Fig. 5) was superior in AZ91D alloy than in AM60B alloy. These results were consistent with the tensile properties.
Table 2 Tensile properties of various specimens.
Figure 6 and Fig. 7 shows an S–N curve obtained by the rotating bending fatigue test for the AM60B alloy and AZ91D alloy, respectively. In the low load range, the specimen did not break even when the rotation cycle reached 107, the test was terminated at that point, and the evaluation was performed using the fatigue strength (arrows in the figure) at that time. For both alloys, the test results for the R part with a diameter of 4 mm fluctuated significantly compared to those for the R part with a diameter of 7 mm. This is considered to be due to the effect of voids on the R part with a diameter of 4 mm as shown as Fig. 4.
It was difficult to determine the fatigue strength value of the AM60B alloy, due to the large variation in the test results for the R part with a diameter of 4 mm; however, the carbon fixing slightly improved the fatigue strength. On the other hand, the fluctuation was suppressed in the test result for the R part with a diameter of 7 mm, and carbon fixing improved the fatigue strength by approximately 6%. This value was equivalent to 7% of the 0.2% proof stress. In addition, since the fatigue strength of the R part with a diameter of 7 mm was lower than that of the R part with a diameter of 4 mm, a dimensional effect8) in which the fatigue strength decreases with an increase in wall thickness was observed. However, the grain size significantly reduced the fatigue strength. From the results of the EBSD analysis shown in Fig. 5, it was observed that regardless of carbon addition, the coarsening of the crystal grains reduced the fatigue strength, as the grain sizes of the R part with a diameter at 7 mm was larger than those of the R part with a diameter of 4 mm.
The test result of the AZ91D alloy was similar to that of the AM60B alloy. The test result of the R part with a diameter of 4 mm exhibited large fluctuations, making it difficult to determine the fatigue strength; however, the fatigue strength was significantly improved by carbon addition. The variation was suppressed in the test result of the R part with a diameter of 7 mm, and with carbon addition, the fatigue strength was increased to 97 MPa, which is approximately 20% higher than that of the carbon free specimen with a fatigue strength of 80 MPa. This value exceeded the fatigue strength of the R part with a diameter of 7 mm of the carbon-free AM60B alloy. In the tensile properties shown in Table 2, the tensile strength and elongation of the AM60B alloy were significantly higher than those of the AZ91D alloy. The elongation, which is closely related to the fatigue strength, was approximately ten times higher in the AM60B alloy. Based on the results of the tensile properties, it was expected that the fatigue strength of the AM60B alloy will be higher than that of the AZ91D alloy. In the test results, for the carbon fixed R part with a diameter of 4 mm, the fatigue strength of the AM60B alloy exceeded that of the AZ91D alloy, reflecting the tensile properties. In contrast, for the R part with a diameter of 7 mm, the fatigue strength of the AZ91D alloy was higher than that of the AM60B alloy. This is due to the reduced grain size of the AZ19D alloy as shown in Fig. 5. This result confirms that the grain sizes of the carbon fixed AZ91D alloy in the R part with a diameter of 7 mm was significantly finer than those of the AM60B alloy having the same diameter. This result confirmed that the fine grain size of the AZ91D alloy contributed to the improved fatigue strength.
3.3 Void suppression effect by carbon fixing
In the preceding paragraph, it was confirmed that voids generated inside the thixomolded specimen and crystal grain sizes affected the mechanical properties. Here, the cause of void generation and the effect of suppressing voids by carbon fixing are discussed. As shown in Fig. 4, the voids generated in the carbon free AZ91D alloy were larger and more numerous than those of the carbon free AM60B alloy. In addition, it was confirmed that no visually observable voids are generated in a thixomolded product using an Mg–Zn alloy chip containing no aluminum regardless of the presence of carbon. These results indicate that the presence of aluminum in the alloy is closely related to void formation.
Therefore, gas analysis was performed for the AZ91D alloy containing a large amount of aluminum. The AZ91D alloy ingot and the chuck portion where more voids were formed after thixomolding were analyzed, and the results are shown in Table 3. For the ingots, the concentration of H2 was 10.09 cc/Mg100 g, which was the most frequently detected gas, and the concentration of other gases was less than 1.00 cc/Mg100 g each. In the carbon free specimen, the concentration of H2, CO/N, and CO2 increased, with H2 increasing remarkably to 364.02 cc/Mg100 g. The concentration of gases in the carbon fixed specimen was lower than that of the carbon free specimen, with the concentration of H2 reduced to less than a quarter. Since CO/N and CO2 were gases caused by air, it was presumed that air was entrained in the material during the thixomolding process. Therefore, the content of CO/N and CO2 in the carbon free specimen with many voids was higher than that of the carbon free specimen with little void. Furthermore, the presence of H2 originates from the material, as it is not present in large amounts in the atmosphere. The H2 content in the ingot was 10.09 cc/Mg100 g, but after chipping, the H2 content of the carbon free thixomolded specimen increased to 364.02 cc/Mg100 g. As described above, H2 was generated during the thixomolding process, and it was presumed that the generated H2 was involved in the formation of voids. The H2 content of the carbon fixed thixomolded specimen was 79.27 cc/Mg100 g, which was significantly lower than that of the carbon free specimen, suggesting that voids were formed by H2.
Table 3 Results of gas analysis. (cc/Mg 100 g)
The factors responsible for the generation of H2 gas in the thixomolding process are discussed below. In the gas analysis, the hydrogen amount of the chip was found to be much larger than that of the ingot. This result suggests that water vapor in the atmosphere is adsorbed on the chip surface before thixomolding, and that magnesium hydroxide is generated on the surface. During thixomolding, the chips are heated in a cylinder and sent to the injection side while being mechanically stirred by a screw. Magnesium hydroxide is decomposed into magnesium oxide and water when it is heated to approximately 600 K or more.9) Since magnesium chips are heated to approximately 873 K in a cylinder, magnesium hydroxide on the chip surface is decomposed into magnesium oxide and water. It has been reported that a large amount of hydrogen is generated by the reaction between aluminum and water.10) Therefore, it is probable that the water generated by the decomposition of magnesium hydroxide reacted with aluminum in the AZ91D alloy and generated a large amount of hydrogen. The result explains the generation of high number of voids in the AZ91D alloy with more aluminum content compared to that of AM60B alloy with less aluminum content.
On the other hand, the voids generated in the thixomolded specimen were suppressed by fixing carbon nanoparticles to the chip surface. If the void is based on the above-mentioned hydrogen generation due to the reaction between aluminum and water, it is conceivable that carbon nanoparticles on the chip surface suppress the reaction between aluminum and water. Carbon has water absorbency. In particular, as the size becomes smaller, the water absorbability greatly increases.11) The average particle size of the carbon nanoparticles used in this study was extremely fine at 20 nm. Therefore, the carbon nanoparticles efficiently absorbed the water generated by the decomposition of magnesium hydroxide, thus suppressing the reaction between aluminum and water, and consequently, reducing the generation of hydrogen.
Figure 8 shows the appearance of a thixomolded specimen using a spiral mold with a molded part thickness of 0.5 mm under the same conditions (injection temperature: 853 K, injection speed: 2.5 m/s). The flow length of the carbon fixed AZ91D alloy was significantly longer than that of the carbon free AZ91D alloy, and the flowability of the molten metal was improved by carbon fixing. These results indicated that the molten metal flowed smoothly into the mold during injection molding, and this result confirmed the reduction of H2 in the molten metal due to carbon fixing.
