Abstract
The size of defects in castings affects mechanical properties such as fatigue strength. These properties are the most important properties used for designing aluminum alloy casting components, and usually increase with decreasing size of internal defects. Large internal defects mainly form due to the volumetric shrinkage of melt at the last solidified position. In this study, we therefore examined the possibility of generating microbubbles in aluminum alloy melts for dispersing internal defects in castings.
Ultrasonic treatment was used for generating nitrogen or argon microbubbles in the melt and was found to be suitable for forming fine gas porosities dispersed in castings. The fine gas remained in the melt for 3600 s, and the generated argon gas was detected from the porosities in the castings. The ultrasonic treatment could also remove inclusions and hydrogen in the melt. Moreover, the fine gas porosities in castings did not decreases the tensile strength of the castings.
This Paper was Originally Published in Japanese in J. JFS 93 (2021) 394–399.
1. Introduction
Aluminum alloy, because of its low specific gravity, is being used to make engine parts and structural body components to reduce the weight of automobiles. To manufacture these parts by casting, it is necessary to prevent casting defects, which degrade the material strength; thus, the development of defect reducing methods is being promoted. For example, regarding gas defects in die castings, the melt pressure transmission behavior required for reducing this kind of defect has been investigated,1–3) and vacuum die casting has also been developed to decrease the occurrence of this kind of defect. The research and development in this area principally focuses on preventing casting defects. However, it has been reported that internal defects smaller than ϕ180 µm have no adverse effect on the fatigue strength, the specified property in design standards4) for castings. This means that if the size of internal defects can be controlled to be smaller than ϕ180 µm, the fatigue strength of castings can be maintained. Coarse defects larger than ϕ180 µm are mainly caused by the consolidation of individual shrinkage cavities. Therefore, if the shrinkage cavities are dispersed, castings without strength decrease are expected to be obtained.
The shrinkage defect occurs due to the cessation of melt feeding into the shrinkage cavity and it coarsens when the final solidification regions concentrate in one place. The concentrated shrinkage cavities can be dispersed by generating numerous micro-pores with a melt containing a high H2 gas content.5) Therefore, if fine inert gas bubbles are introduced into the melt properly, the shrinkage cavities will be dispersed and form smaller cavities. Fine gas bubbles introduced into the melt can remain there for a long time due to their low ascension velocity.6) The method for introducing fine gas bubbles into a liquid is generally classified into the following three types: first, injecting gas into a liquid through narrow holes; second, generating microbubbles by changing the pressure and temperature;7) and third, entrapping gas bubbles in a liquid by stirring the gas-liquid interface violently. Among these methods, the third one is currently considered to be the most suitable method for introducing fine gas bubbles into molten aluminum alloys. Specifically, in this method, the gas is dispersed into fine bubbles by shear force8) or discharging a liquid containing gas through narrow slits.9) The ultrasonic pressure vibration has been proposed as a means for stirring the gas-liquid interface.10,11)
In the present study, we investigated the possibility of introducing fine gas bubbles into molten aluminum alloy by ultrasonic vibration, which has been used for the microstructure improvement of castings.12) Then, we evaluated the effect of the fine gas bubbles on the shrinkage behavior of aluminum alloy.
2. Experimental Method
2.1 Casting method
The molten metal of the Japan Industrial Standard ADC12.1 alloy was used, and its composition is shown in Table 1. The alloy was melted at 953 K, and the ultrasonic treatment described below was conducted at 943 K after the slag was removed. Then the melt was air-cooled to a predetermined temperature and poured into a shell mold shown in Fig. 1(a). The aluminum alloy (gross weight of 1.6 kg) was melted in a crucible with an outer diameter of 140 mm and height of 191 mm (made by Nippon Crucible Co., Ltd., size No. 6), which was placed in a cylindrical heater (with an inner diameter of 180 mm and a depth of 240 mm) and covered with heat-insulating material. The ultrasonic treatment was performed by inserting a horn into the melt from above the crucible. The copper Lansley mold illustrated in Fig. 1(b) was used without preheating for making the castings for analyzing the amount of dissolved H2 in the melt and assessing the strength of castings.
Table 1 Composition of ADC12.1 alloy used in this study.
The macrostructures of the castings were observed by cutting the ingots through the center and polishing the cut section with #800 emery paper prior to etching with a 10% hydrofluoric acid aqueous solution. The state of the internal shrinkage in some castings was also confirmed using an X-ray CT device (Tesco Co. Ltd.).
2.2 Method for introducing fine bubbles into the molten metal
A conical horn with a tip diameter of ϕ8 mm and a gas injection hole with an interior diameter of ϕ0.9 mm was attached to an ultrasonic oscillator (made by Seidensha Electronics Co., Ltd., type 625D, frequency of 28.5 kHz), as shown in Fig. 2. The tip of the horn was submerged in the melt to a tip depth of 100 mm and vibrated with an amplitude of 45 µm. At the same time, the gas was discharged into the melt through the ϕ0.9 mm hole at the tip of the horn. Because the ultrasonic horn was designed for vibration at room temperature, it could only vibrate for a few tens of seconds in the melt before its temperature increased excessively.
Therefore, the tip of the horn was changed several times to obtain a total duration of 60 to 150 s for ultrasonic vibration and gas introduction. The tip of the horn was designed to be detachable and the tip exchange was carried out within 20 s. Fine bubbles were introduced into the melt by discharging N2 gas from the gas introduction hole during ultrasonic vibration at a flow rate of 500 cc/min (pressure of 0.5 MPa) (hereafter referred to as ultrasonic treatment).
It is assumed that fine bubbles stayed in the melt, but large bubbles floated up during the ultrasonic treatment. Therefore, the holding time (at 933 K) of the melt between the end of the ultrasonic treatment and pouring the melt into the Ransley mold at 923 K was changed for several levels up to 3600 s to examine the presence of the gas in the castings by gas analysis.
2.3 Gas analysis and assessment of mechanical properties
To confirm the introduction of gas into the melt, ultrasonic treatment (vibration time of 150 s) using Ar gas, which is almost nonexistent in the atmosphere, was also performed to introduce a marker gas. Test pieces cut out from the sampling position in Fig. 1(b) were used for gas analysis by both the Ransley method and the gas mass spectrometry method during cutting. In the Ransley method, the total gas amount was calculated from the pressure increment caused by melting the test piece in vacuum, and the amount of gas except H2 was calculated from the residual pressure after removing H2 with a palladium filter. In the mass spectrometry method, the total gas amount was calculated from the pressure change due to the released gas by boring a hole 3 mm in diameter and 10 mm in depth in the test piece in a vacuum (1 × 10−5 Torr) at room temperature and the composition of the gas was determined by mass spectrometric analysis. The gas pressure (1 × 10−5 to 1 × 103 Torr) was measured by a Baratron pressure gauge (MKS Co., Ltd., USA, model 390H), which measures the absolute value of the pressure, and mass spectrometric analysis was conducted with a quadrupole mass spectrometer (Anerva Co., Ltd., type M-101QA-TDM), which has a mass measurement range of 1 to 100 amu.
The melt ultrasonically treated by introducing N2 gas at 943 K for 140 s was held for 900 s and poured at 923 K into the mold to make the castings shown in Fig. 1(b) for assessing the mechanical properties. For the purpose of comparison, castings were also made by pouring a melt with the same initial temperature, holding time, and casting temperature but without ultrasonic treatment. Specimens for tensile tests (total length of 65 mm, parallel part length of 27 mm, and parallel part diameter of 4.5 mm) were cut from the sampling position shown in Fig. 1(b). The tensile tests were performed at a speed of 0.5 mm/min (the corresponding strain rate was calculated as 3.1 × 10−4/s) (Shimadzu Co., Ltd., test machine model AG-50kNLC). The fracture strain was measured by butting the two fractured pieces after the test. The tests were all carried out with N = 3.
2.4 Solidification simulation method
The solidification and pore formation processes (the shrinkage amount of each element) after pouring the ADC12 melt into the shell mold were analyzed by dividing the shell mold into a 1-mm cubic mesh. In the analysis, the contraction factor in the liquid state and the value of the critical fraction of solid at which the melt flow stops were set to 0.025%/K and 0.9,13) respectively. The final porosity (shrinkage amount) of each element was calculated by assuming that the contraction volume of an element is compensated by the melt flowing from the adjacent elements having higher temperatures during each time increment until the critical fraction of solid is reached. The above analyses were performed with TOPCAST (Toyota Communication System Co., Ltd.), which has all the required analysis functions. The thermophysical properties shown in Table 2 were used for the ADC12 alloy. The initial temperatures of the melt and the mold were taken as 923 K and 298 K (room temperature), respectively. The heat transfer coefficient between the melt and the shell mold was taken as 800 W/m2·K, at which the simulated and the measured solidification times showed almost the same value.
Table 2 Thermo physical properties of aluminum alloy used for simulation.
3. Results and Discussion
3.1 Gas amount in castings made of melt with ultrasonic treatment
Figure 3 shows the relationship between the gas volume in the castings and the holding time of the melt before pouring after ultrasonic treatment for 61 to 77 s by N2 gas. In the figure, the ■ mark is the total gas volume, the 
mark is the H2 gas volume, and the gas volume in the castings made of the melt without ultrasonic treatment is shown at 0 s. Because the castings made without ultrasonic treatment were not degassed, their total gas volume is as much as 0.29 cc/100 g Al and the H2 gas volume is as much as 0.28 cc/100 g Al. The ultrasonic treatment reduced the total gas volume and total H2 volume in the castings (holding time of 150 s after ultrasonic treatment) to 0.22 cc/100 g Al and 0.20 cc/100 g Al, respectively. The degassing effect of the ultrasonic treatment attributed to flotation of the coarse bubbles generated by ultrasonic treatment. In the floating process, H2, and other gases in the melt were expected to be captured and floated to the surface of the melt by the coarse bubbles. After that, both the total gas volume and the H2 gas volume increased with increased holding time. As the holding time reached 3600 s, the total gas volume and the H2 gas volume increased to 0.33 cc/100 g Al and 0.31 cc/100 g Al, respectively. However, only the H2 gas volume increased, while the difference between the total gas volume and the H2 gas volume (0.02 cc) remained constant in the melt holding process. Therefore, it is considered that the increase of the gas volume was caused by the increase of the H2 gas that was generated by the reaction between moisture in the atmosphere and the melt and then partly dissolved in the melt and absorbed by the fine bubbles. The increment of the difference between the total gas volume and the H2 gas volume in the castings with ultrasonic treatment and that without ultrasonic treatment is the volume of N2 gas introduced by the ultrasonic treatment. The accuracy of the above measurement was about 0.01 cc/100 g Al.
For each melt holding time, the difference between the total gas volume and the H2 gas volume in the castings with ultrasonic treatment is more than 0.02 cc/100 g Al, while that in the castings without ultrasonic treatment is only 0.01 cc/100 g Al. Therefore, it is inferred that gas was introduced into the melt by the ultrasonic treatment. The existing state of the N2 gas introduced into the melt was not investigated, but it is presumed to have been contained within the fine bubbles and/or to attached to the oxides. With the density being ignored and the shape being assumed as spherical, the ascension velocity of the gas bubbles in the melt can be obtained from the Stokes theory as vs = 2/9 × (r2 × ρ × g)/η, where r is the radius of the gas bubble, ρ is the density of the melt, g is the acceleration of gravity, and η is the viscosity of the melt. If the density, the viscosity, and the height of the Al melt in the crucible are 2700 kg/m3, 0.0013 kg/m·s, and 0.15 m, respectively, the time for a bubble with a radius as large as 1 mm to float up to the surface of the melt is calculated as 3.3 s, which is a very high ascension velocity. The ascension time of the bubble increases drastically from 330 to 3700 and 33,000 s with the decrease of its radius from 0.1 to 0.03 and 0.01 mm, respectively. In this study, the radius of the gas bubble introduced by the ultrasonic treatment is considered smaller than 0.03 mm because the amount of the gases in the melt decreased very little after 3600 s.
3.2 Effect of ultrasonic treatment on shrinkage behavior
The simulation results of the solidification time distribution (pouring temperature of 923 K) on the cross section and the state of shrinkage cavities of the shell-mold castings are shown in Fig. 4(a) and Fig. 4(b). The simulation results for these castings show that the solidification of the whole casting was completed in 357 s, and the center region from the tip of the protective tube to the upper surface of the casting solidified slowest. Because the shrinkage cavity in the center region was compensated by the melt in the upper region of the casting, the shrinkage cavity concentrated in the upper region for the castings made from a melt without fine bubbles, as shown in Fig. 4(b).
The cross-sectional macrostructures of castings made from the melt without ultrasonic treatment and the castings made from the melt with ultrasonic treatment are shown in Fig. 5. The melt was air-cooled to 923 or 873 K before pouring after the ultrasonic treatment. For the castings made without ultrasonic treatment, external shrinkage occurred in the upper part of the castings, as described for the simulation result. This external shrinkage can become smaller as the pouring temperature decreases to 873 K. This is due to the decrease in the shrinkage volume of the liquid phase before solidification with a reduced pouring temperature.
In contrast, for the castings made with ultrasonic treatment, fine pores formed throughout the castings and the shrinkage cavities did not concentrate in the upper region of the castings. This dispersion behavior of pores occurred similarly even if the pouring temperature was changed. The above result is considered to have been caused by the expansion of the fine bubbles during solidification, which were introduced into the melt by the ultrasonic treatment and dispersed in the melt. Because the fine bubbles have a high interior pressure,14) the solidification shrinkage of the melt is considered to have been compensated by the expansion of the fine bubbles. In addition, the H2 dissolved in the melt in an atomic state may have become supersaturated during solidification and precipitated into the fine bubbles. As a result, the shrinkage cavities concentrated in the upper region of the castings made of the melt without ultrasonic treatment are considered to have been dispersed due to the scatter of the final solidification regions for the castings made of the melt with the ultrasonic treatment.
To confirm the existence of fine bubbles, the lower portion of the castings with a pouring temperature of 873 K, as shown in Fig. 5, was cut and the cross-sectional microscopic structure was observed, as shown in Fig. 6. Large shrinkage cavities were observed in the castings made from the melt without ultrasonic treatment. On the contrary, many fine shrinkage cavities were scattered in the castings made from the melt with ultrasonic treatment. The sizes of the shrinkage cavities measured from these photographs of the microscopic structure are shown in Fig. 7. In this study, the size of the shrinkage cavities was considered to be the same as the size of fine bubbles, although the fine bubbles might expand slightly due to the shrinkage of the castings during solidification. The size of the shrinkage cavities in the castings made without ultrasonic treatment is 50 µm on average with a range from 20 to 80 µm. However, the size of the shrinkage cavities in the castings made with ultrasonic treatment is about 15 µm on average, and ultrasonic treatment also reduce the range, to 8 to 30 µm.
The defect distribution in the cross section of the shell castings made of the melt with the ultrasonic treatment by Ar gas was also observed by optical microscopy and X-ray CT, as shown in Fig. 8. It was confirmed that shrinkage cavities (white spots) also occurred throughout the castings, even when Ar gas was used for ultrasonic treatment. The dispersed shrinkage cavities can also be observed in the X-ray CT image (black spots in Fig. 8). The microstructure of the Ransley castings made of the same melt is shown in Fig. 9. The existence of fine bubbles can also be confirmed as pores in the Ransley castings. Table 3 shows the semi-quantitative analysis result of the gas generated during heating and melting of the sample in the Ransley castings at 943 K. Here, the content of each gas component is shown as a percentage of the total gas by volume, which was 0.26 cc/100 g Al. It is found that the H2 gas content is as high as 90% in the melt. In the table, N2 and CO cannot be distinguished because they have the same mass number of 28, but the remaining gas is estimated to be N2 mainly by the spectral pattern. Although most of the remaining gas is N2, Ar is also present at 1%. Therefore, the Ar gas can be considered to have been introduced into the melt by the ultrasonic treatment because it is rare in the atmosphere.
Table 3 Semi-quantitative analysis of gas components in castings.
Next, the effect of the introduced fine bubbles on the shrinkage of the shell castings was further investigated. Figure 10 shows the macrostructures in the cross sections of the castings made of the melt with holding times of 150, 900, 1800, and 3600 s after the ultrasonic treatment by N2 gas (vibration time of 61 s) prior to pouring into a shell mold. The castings with a holding time of 150 s were made by pouring the melt at 923 K after air cooling but without holding in the furnace after the ultrasonic treatment (elapsed time of the process was 150 s). In each of these castings, a large external shrinkage, similar to the one shown in Fig. 6, did not occur. However, the concentration of the shrinkage cavities on the upper side of the castings increased with the extension of the melt holding time. Some concentration of shrinkage cavities on the upper side of the castings was observed for the melt holding time of 3600 s.
3.3 Effect of ultrasonic treatment on mechanical properties
Figure 11 shows the tensile strengths and the elongation of the test pieces. The tensile strengths of the castings made from the melt with ultrasonic treatment are 269 to 285 MPa, which are almost similar to the 290 MPa of the castings (B-2) made of the melt without ultrasonic treatment. The elongations of the castings made with ultrasonic treatment are 3.2 to 3.8%, which are almost similar to the maximum value (B-2), 3.9% of the castings made of the melt without ultrasonic treatment, and the variation range is also smaller. Thus, it was confirmed that the presence of fine defects does not affect the strength of the castings. Figure 12 shows the macrostructures of the fracture surfaces after the tensile test. As shown by images A-1 and A-2, there is almost no coarse oxide on the fracture surface of the castings made from the melt with ultrasonic treatment. In contrast, coarse oxides can be observed on all of the fracture surfaces of the castings made without ultrasonic treatment. These results can be explained by the oxides in the melt being removed during the flotation of large bubbles, which is also effective for dehydrogenation, introduced by the ultrasonic treatment, or that the oxides were crushed by the ultrasonic treatment.
Furthermore, fine shrinkage cavities were also observed near the surface of the shell mold castings, as shown in Fig. 8. The effect of these fine shrinkage cavities near the surface on the mechanical properties, especially the fatigue strength, of the castings needs to be examined separately.
4. Summary
Through investigating the possibility of introducing fine bubbles into molten aluminum alloy by ultrasonic treatment and the effect of the fine bubbles in the melt on the shrinkage behavior of castings, the following results were obtained:
-
(1)
The shrinkage cavities that are normally concentrated in the upper region of the shell mold castings were dispersed throughout the castings by introducing fine bubbles by ultrasonic treatment.
-
(2)
More fine shrinkage cavities were observed in the castings made from the melt with ultrasonic treatment using Ar as a marker gas than in the castings made from the melt without ultrasonic treatment. The gas in the castings was mainly H2, but Ar, which is rare in the atmosphere, was also detected, confirming that fine bubbles were introduced into the melt by ultrasonic treatment.
-
(3)
During ultrasonic treatment, coarse bubbles are also generated together with the fine bubbles. The total gas volume in the castings (and melt) was largely reduced by the degassing effect of the coarse bubbles, but the gas volume increased in the castings when the melt holding time after ultrasonic treatment was increased. The increase of the gas is considered to have been caused by H2 gas being generated from the reaction between moisture in the atmosphere and the melt and entering the fine bubbles or dissolving into the melt. However, the volume of other gases, such as N2, hardly changed during the melt holding; therefore, it is considered that the fine bubbles can stay in the melt for a long time.
-
(4)
The shrinkage cavities changed from the dispersed state to a concentrated state with an increase in the melt holding time after ultrasonic treatment. However, this change from a dispersed state to a concentrated state did not happen unless the melt holding time was as long as 3600 s after the ultrasonic treatment. This may be due to the low ascension velocity of the fine bubbles.
-
(5)
The introduction of the fine bubbles into the melt had almost no adverse effect on the mechanical properties of the castings. Stable mechanical properties of the castings can be obtained by the ultrasonic treatment due to its effect of removing or crushing oxide inclusions.
REFERENCES
- 1) Y. Iwata, S. Dong, Y. Sugiyama and H. Iwahori: Mater. Trans. 53 (2012) 483–488. doi:10.2320/matertrans.F-M2011858
- 2) Y. Iwata, Y. Sugiyama and H. Iwahori: J. JFS 72 (2000) 263–267. doi:10.11279/jfes.72.263
- 3) Y. Iwata, S. Dong and H. Iwahori: J. JFS 83 (2011) 421–426. doi:10.11279/jfes.83.421
- 4) R. Izawa: Doctoral thesis of Yokohama National University (2014) 18.
- 5) Y. Awano: J. JILM 56 (2006) 335–342. doi:10.2464/jilm.56.335
- 6) R. Kitagaki: J. Japan Soc. Precis. Eng. 83 (2017) 647–650. doi:10.2493/jjspe.83.647
- 7) Y. Tamakawa and T. Makuta: J. Jpn. Soc. Exp. Mech. 12 (2012) 17.
- 8) H. Taisei: IPC Patent No. 3397154 (2003).
- 9) M. Takahashi, H. Makuta and H. Takemura: New Technology for Fine Bubbles, (NTS Publishing, Tokyo, 2006) p. 17.
- 10) T. Makuta, F. Takamura and E. Hihara: J. Fluid Mech. 548 (2006) 113–131. doi:10.1017/S0022112005007470
- 11) T. Makuta, R. Suzuki and T. Nakao: Ultrasonics 53 (2013) 196–202. doi:10.1016/j.ultras.2012.05.010
- 12) For example, K. Tsunekawa: J. JFS 81 (2009) 469–474. doi:10.11279/jfes.81.469
- 13) Y. Iwata, S. Dong, Y. Sugiyama and J. Yaokawa: J. JFS 91 (2019) 529–533. doi:10.11279/jfes.91.529
- 14) H. Tsuge: Bull. Soc. Sea Water Sci. Jpn. 64 (2010) 4–10. doi:10.11457/swsj.64.4
