Obtaining a refined microstructure is the first step for using it in thixoforming processes. However, the chemical refining of the grain of aluminum alloys with high silicon content is impaired; therefore, grain refining via UST (Ultrasonic Melt Treatment) is a possible solution for these cases. This work aims to analyze the microstructural stability of the microstructure of the Al7Si2.5Cu alloy in the semisolid state. The material was prepared by conventional casting and submitted to processing via UST for 20 seconds with a steel horn. The material underwent reheating heat treatment for 0, 30, 60, 90, and 210 seconds under two solid fractions of 45% and 60%, respectively, 572 and 565 °C. The results show an average primary globule size of around 88 μm and a circularity form factor above 0.50. Regarding the eutectic morphology, it was quite fragmented when processed via UST; that is, the application of acoustic waves in the liquid metal causes it to mix effectively, and it is also capable of breaking the coarser Si particles. When comparing the microstructural evolution of the Al7Si2.5Cu alloy via UST with the ultra-grain refining technique, UST presents superior results in all analyzed scenarios.
Keywords: Thixoforming; semisolid material; ultrasonic melt treatment; Al7Si2.5Cu alloy
1. Introduction
In 1972, Spencer et al.1 observed the semisolid behavior of alloys at temperatures within the solidification range, i.e., between liquidus and solidus. Since then, much research has been done to understand this viscous behavior and processing routes to reach this state more easily. A consensus was established that it is necessary to achieve a mixture of tiny solids with spheroidal morphology immersed in liquid to achieve the best viscous behavior; several techniques were tested to achieve this situation of grain refining: chemical grain refining technique (GR), mechanical stirring (MS), electromagnetic stirring (EMS), ultrasonic melt treatment (UST), angular channel extrusion (ECAP), strain-induced melt activation (SIMA). The grain refining technique via ultrasonic melt treatment has been used since the 1970s and allows parts to be produced at a lower cost and with less complex production processes than other techniques. The use of ultrasonic melt treatment for the preparation of semisolid raw material is based on the application of high-power ultrasonic vibrations in the liquid metal to reduce the average grain size and better homogeneity of the material, among others, aiming for a better mechanical performance of products2-4.
Furthermore, there are several advantages in using the processing of materials in the semisolid state compared to conventional processes, namely: (a) energy efficiency, since there is no need to keep the material in the liquid state for a long time as in casting, (b) high productivity rates, due to the possibility of working with lower temperatures, consequently the solidification times are reduced, allowing a greater production for the same time interval, (c) smooth filling of the mold without trapping air, reaching improved quality on products, (d) increased die life due to the use of lower process temperatures gradients when compared to conventional processes, reducing wear on the dies, (e) obtaining refined and homogeneous microstructures, improving mechanical properties, (f) obtaining near net shape components, producing parts with dimensions close to the final product and with a reduced number of steps in its processing5. Therefore, the objective of this work is to analyze the microstructural stability during the reheating treatment for the times of 0, 30, 60, 90, and 210 seconds, under two conditions of solid fractions of 45 and 60%, respectively, 572 and 565 °C of the Al7Si2.5Cu alloy processed via UST; it should be borne in mind that optimal SSM processing takes place in this solid fraction range. This specific alloy is nothing more than the traditional 356 modified with 2.5 wt% Copper. The authors developed it to be used in semisolid state processing, and it has demonstrated excellent rheological behavior6. This work improves its manufacturing process with ultrasonic melt treatment, which implies generating the refined microstructure necessary for subsequent processing in the semisolid state. It is noted that adding Cu allows for better control of the solid-liquid transformation and allows the final product to undergo heat treatments for solubilization and aging, generating excellent mechanical properties7.
2. Materials and Methods
The alloy was produced using SiC crucibles (insulated with silico-aluminous coating QF180) in a resistive furnace at 750 °C through a mixture of base alloys, namely, alloy A356 and commercially pure copper, producing 0.5 kg of raw material at each test (1 ingot/condition). The chemical composition of the alloy, obtained via optical emission spectrometry equipment, can be seen in Table 1.
Chemical composition of Al7Si2.5Cu alloy (in wt%), obtained via optical emission spectrometry.
The casting temperature was estimated via Thermo-Calc® software simulation using a pouring temperature of 50 °C above the liquidus temperature; this increase in temperature is due to the drop in temperature associated with the moment of removal of the molten raw material from the furnace until the moment of its pouring at the copper refrigerated mold. For the working temperatures determined via Thermo-Calc® software simulation, a calculation routine was employed, evaluating the solidification conditions in and out of equilibrium based on Scheil models; the main objective is to have a condition in which the transition from the solid to the liquid can be easily controlled; using the TTAL5 database. Therefore, the temperatures obtained are solidus temperature of 525 °C, liquidus temperature of 606 °C, and working temperatures for the solid fractions of 45 and 60%, respectively, 572 and 565 °C6. One of the criteria proposed by Liu et al.8 for the raw material to be considered thixoformed is that the complete solidification range (ΔT) is less than 130 °C and the alloy under study has this criterion, obtaining a solidification range of 81 °C.
After complete alloy melting, the UST processing technique was carried out for 20s using a steel horn (AISI 4340 steel). For this purpose, a horn was immersed in the liquid metal (3 mm below its surface), while the melt temperature was measured by a thermocouple type K Cromel-Alumel. It should be noted that the choice of treatment time and the horn type is due to conclusions of previous works9,10. Processing via UST was performed using Sonitron equipment (20 kHz, 2.8 kW); then, the alloy was poured into a cooled copper mold to produce ingots with lengths and diameters of 250 mm and 30 mm, respectively. This technology is believed to produce high-intensity vibrations and thoroughly agitate the liquid metal, improving the raw material's grain refinement and homogenization effects. The complete production cycle of the raw material and the UST equipment schematics can be seen in Figure 1.
Schematization of the raw material production complete cycle (a) and the equipment illustration used to refine the alloy through UST (b) processing.
Reheating heat treatments were performed on samples of 25 mm in height and 30 mm in diameter for approximately 8 minutes in a Norax induction furnace (25 kW, 8 kHz) with a heating rate between 80 and 100 °C/min. When the samples reached the temperatures referring to the solid fractions of 45 and 60%, they were kept for 0, 30, 60, 90, and 210 seconds and then cooled in water. The samples were characterized by optical microscopy to assess their morphological stability in the semisolid state using conventional metallography and polarized metallography (color). For microstructural characterization via conventional metallography, the samples were sanded with 220 up to 1500 sandpapers polished with 6 μm and 1 μm diamond paste and subsequently attacked with HF reagent (1ml and 99 ml of H2O) for 10s. For the microstructural characterization via color metallography, the same samples used in the previous characterization were subjected to an electrolytic attack with deposition of HBF4 (fluoroboric acid) in a 2% solution and voltage of 25 V for approximately 6 minutes under moderate and constant agitation6,7. A Leica DM ILM optical microscope was used to acquire images. For the acquisition of color images, polarizing filters were used, causing grains with the same crystalline orientation to present similar coloration, which makes identification and characterization easier. To measure the average size of primary globules and grains, the Heyn Intercept Method was used, governed by the ASTM E112 standard11. The imaging software ImageJ 1.40g was used to characterize the circularity form factor. With information on the average size of primary globules (GLS), grains size (GS), and circularity shape factor (SF), it is possible to calculate the rheocast quality index (RQI), which correlates these microstructural parameters according to Equation 1. RQI close to 1 indicates that each solid particle in the semisolid state is a grain, without many interconnections, such as a particle with a less complex shape. Furthermore, the greater the circularity of this particle, the greater the ease of movement of the semisolid slurry and the lower the viscosity of the paste.
Chemical composition analysis was also carried out using energy dispersive X-ray spectroscopy (EDS) using a ZEISS EVO/MA15 microscope to identify the elements present in the alloy (map with the chemical distribution and specific elements of the matrix and constituents).
Figure 2 shows the expected phase diagram for the Al7Si2.5Cu alloy obtained via Thermo-Calc® software with its respective microstructures, considering the amount of 2.5% by weight of Cu. Analyzing the microstructure of the as-casting condition, a typically dendritic microstructure is observed, formed by the α phase, surrounded and permeated by the eutectic phase. The microstructures under the solid fractions of 45 and 60% will present a practically globular morphology and an increase in their degree of sphericity, mainly for the solid fraction of 45%.
Expected phase diagram for the Al7Si2.5Cu alloy obtained via Thermo-Calc® software with its corresponding microstructures.
3. Results and Discussion
Firstly, it is necessary to emphasize that this refined raw material is ideal for processing via thixoforming since the refined structure will undergo Ostwald ripening and coalescence phenomena that generate the desired globalized solid fraction. The microstructural characterizations comprise the characterization of the casting structure (processed via UST) and of the thermally treated ones at temperatures of 45 and 60% of solid fraction and treatment times of 0, 30, 60, 90, and 210 seconds, making up the size count average of primary globules and grains, determination of roundness form factor and RQI calculations. It must be remembered that industrial processing uses time, not temperature, as a control parameter. Therefore, it is necessary to evaluate the stability of this material in different treatment times, and the interval between 0 and 210 seconds is more than enough for the heated billet to reach the target temperature. If there is a significant variation in the size and morphology of the primary phase, there will be variation in rheological terms, making the process unstable. If there is microstructural stability, rheological stability is expected and, therefore, excellent process control.
Delving deeper into the research, the alloy's microstructure under the casting condition is analyzed through conventional metallography. A typically dendritic microstructure is observed, formed by the α phase, and surrounded and permeated by the eutectic phase, as seen in Figure 3. As can be seen, the difference between the microstructure treated with ultrasound (Figure 3b and 3d) and the microstructure of the untreated material (Figure 3a and 3c) presents a subtle difference. No grains of small size are observed in the untreated structure; this fragmentation and these small grains originate from the fragmentation caused by the UTS. This perception is not possible in conventional metallography; hence, RQI that correlates both the GS and GLS is used to differentiate the structures6,7.
Microstructure of the Al7Si2.5Cu alloy under different conditions: as-cast (a), as-cast and processed via UST for 20 seconds (b) via color metallography; as-cast (c), as-cast and processed via UST for 20 seconds (d) via conventional metallography.
The average values of primary globule size and average grain size provide a comprehensive understanding of the material's structure. The cavitation mechanism, a key aspect of the research, results in dendrites' nucleation, growth, and fragmentation during solidification, producing high-pressure and low-pressure sites within the liquid metal2-4. This understanding of the cavitation mechanism is crucial for further research and development.
Figure 4 shows the matrix's chemical composition via RX-EDS map and the eutectic and intermetallic phases for the molten condition (processed via UST) for 20 seconds. The elements aluminum (Al), silicon (Si), copper (Cu), and magnesium (Mg) were responsible for the highest concentrations; a similar result was also obtained in the point-type chemical composition analysis. Specifically, in Figure 4b, it is possible to verify that the chemical composition of the matrix was practically composed of aluminum (Alα). In contrast, in Figure 4c, it was evident that the eutectic composition was composed of a high percentage of silicon (Si). The intermetallics comprised a high percentage of Cu, as seen in Figure 4d. The elements Mg and Fe were distributed in small amounts throughout the material, as shown in Figure 4e-f.
RX-EDS map of the chemical composition present in the Al7Si2.5Cu alloy processed via UST for 20 seconds (a), Analyzed elements (b) Al, (c) Si, (d) Cu, (e) Mg and (f) Fe.
Figure 5 identifies and presents the analysis points of the punctual chemical composition with the respective phases and compositions (% in weight) listed in Table 2. The chemical composition of the matrix, measured in point 1, is composed essentially of aluminum (Al) with 93.88%, containing low levels of silicon (Si) and copper (Cu). The average composition of the eutectic, measured at point 2, is composed of a large amount of silicon (Si) of 55.68%, but with the presence of the elements aluminum (Al) and copper (Cu). At point 3, one has the Al2Cu intermetallic with a high copper (Cu) value above 18%.
SEM micrograph of the Al7Si2.5Cu alloy processed via UST for 20 seconds, showing the points used for the punctual chemical analysis: Alα matrix (point 1), Al-Si eutectic phase (point 2), and the Al2Cu intermetallic (point 3).
Figures 6 and 7 show the microstructural characterization via polarized metallography (color) and conventional metallography of the heat-treated alloy at temperatures corresponding to the solid fractions of 45 and 60% and treatment times of 0, 30, 60, 90, and 210 seconds.
Color metallography of the Al7Si2.5Cu alloy processed via ultrasonic melt treatment and heat treated in the solid fraction of 45% (a - e) and solid fraction of 60% (f - j) and treatment times of 0, 30, 60, 90 and 210 seconds, showing the grains` microstructure.
Conventional metallography of Al7Si2.5Cu alloy processed via ultrasonic melt treatment and heat treated in solid fraction of 45% (a - e) and solid fraction of 60% (f - j) and treatment times of 0, 30, 60, 90 and 210 seconds, showing the microstructure of the primary globules.
In Figure 6, through the characterization via color metallography, the rosette-shaped grains are observed from the heat treatment time of 0 s, with a slight increase in the average grain size during the morphological evolution until the treatment time of 210 s. The rheocast quality index (RQI) is the quality index of the recast material given by the quotient between the grain size and the primary globule size, that is, the ratio between the microstructure (color metallography) and the microstructure (conventional metallography) of the material and multiplied by the circularity form factor, it presented satisfactory values, being established that values superior to 0.20 the morphology presents reasonable expectation of formability in the semisolid state.
Through the characterization via conventional metallography, the evolution towards a practically globular morphology can be observed in Figure 7, even for the heat treatment time of 0 s, with a slight increase in the average size of primary globules during the maintenance of the material in the semisolid range for the treatment time of 210 s, as well as a significant increase in its size and degree of sphericity during the morphological evolution for both solid fractions. The value of the circularity form factor indicates how much the structure is spherical (circular), with the vast majority having values above 0.50; it can be noted that the best results occurred for the treatment time of the 60 s, showing that the phenomena of coalescence and Ostwald ripening, both dependent on the time the alloy remains at the heat treatment temperature, acted simultaneously. Benati and Zoqui12, Benati13, working with AlSi2.5Cu alloys (variations from Si) and using the chemical grain ultra-refining technique, observed that with the increase in silicon content, the alloys started to present more pronounced dendritic structures, mainly the Al7Si2.5Cu alloy. Analyzing the same Al7Si2.5Cu alloy, it is observed that the alloy structure becomes almost globular using UST processing, even for the shortest treatment time. The microstructures showed homogeneity concerning the primary globules and the grains. There were no significant differences between the results obtained for the alloy treated at 45 and 60% of solid fraction and treatment times. Table 3 summarizes the quantitative data obtained after reheating the alloy to semisolid temperatures. The 60% solid fraction provided lower values for the circularity form factor and RQI due to coalescence phenomena, which are highly favored by larger solid fractions since there is greater contact between the solid particles; it is also observed that the treatment time of 60 seconds presents the best results for both solid fractions, indicating that it is the ideal time for processing this alloy.
Primary globule size (GLS), circularity shape factor (SF), grain size (GS), and the corresponding rheocast quality index (RQI) for all analyzed conditions.
Since there is no significant difference between the 45 and 60% solid fraction condition, as well as to compare with previous paper12, Figure 8 presents a comparison of the microstructural evolution of the Al7Si2.5Cu alloy using only the solid fraction of 45% and treatment times of 0, 30, 90, and 210 seconds for processing via ultrasonic melt treatment versus grain refine technique12,13 analyzing the values of the average size of primary globules and grain, circularity, and RQI. The grain refine technique introduces nucleating agents to the liquid metal for its refinement; thus, a nucleating agent is a substance intentionally added to liquid metal to act as a nucleation catalyst; a large number of inoculating agents are used in the refining of grains of commercial alloys, for example, inoculants of the Al-Ti-B family and Al-Nb-B, being an efficient technique in obtaining refined microstructures and consequently contributing to the improvement of the mechanical properties of the material14,15.
Processing the material via ultrasonic melt treatment generally presents superior results in all analyzed scenarios. Figure 8a shows an average difference of 26 µm for the size of primary globules, but when the treatment time of 210 s is observed alone, this value increases to 38 µm. Figure 8b presents the roundness form factor, which has values above 0.50, with a difference of 34% for the same material processed via grain ultra-refining technique. Figure 8c, which deals with the average grain size, shows that for the treatment condition of 0 s, there was the smallest drop of 61 µm. However, when the treatment condition of 210 s is observed, there is a drop of 192 µm. Figure 8d shows an increase of approximately 50% in the rheocast quality index (RQI) values, demonstrating that processing via UST becomes an alternative for the microstructural refinement of the material.
It can be noted that in both conditions, production via grain refining and UST, the microstructure is excellently stable, with very low variation in the size of the primary phase, circularity, and grain size within the evaluated time range. UTS, although, has a strong advantage, resulting in a more refined structure.
Figure 9 shows the chemical composition, via RX-EDS map, of the matrix and the eutectic and intermetallic phases for the condition processed via UST for 20 seconds and thermally treated in the solid fraction of 60% and time of 60 s. The chemical composition is very similar to the condition only processed via UST previously presented, where the elements aluminum (Al), silicon (Si), copper (Cu), and magnesium (Mg) were responsible for the highest concentrations, a similar result also obtained in point-type chemical composition analysis. Figure 10 identifies and presents the points of analysis of the punctual chemical composition with the respective phases and compositions listed in Table 4. The chemical composition of the matrix, measured in point 1, is composed essentially of aluminum (Al) with 94.76wt%, containing low levels of silicon (Si) and copper (Cu). The average composition of the eutectic, measured at point 2, is composed of a large amount of silicon (Si) of 24.52wt%, but with the presence of the elements aluminum (Al) and copper (Cu). At point 3 we have the Al2Cu intermetallic with a high copper (Cu) value of 20.55wt%.
RX-EDS map of the chemical composition present in the alloy Al7Si2.5Cu processed via UST for a time of 20 seconds and thermally treated in the solid fraction of 60% and time of 60 s (a) - elements analyzed (b) Al, (c) Si, (d) Cu, (e) Mg, and (f) Fe.
SEM micrograph of the Al7Si2.5Cu alloy processed via UST for 20 seconds and thermally treated in the solid fraction of 60% and time of 60 s, showing the points used for the punctual chemical analysis: Alα matrix (point 1), eutectic phase Al-Si (point 2) and the Al2Cu intermetallic (point 3).
EDS points analysis of the Al7Si2.5Cu alloy processed via UST for 20 seconds and thermally treated at a solid fraction of 60% and a time of 60 s.
Finally, Figure 11 compares the eutectic morphology for three material conditions: conventional casting, casting processed via UST for 20 seconds, and thermally treated at a solid fraction of 60% and a time of 60 seconds. It is observed that the eutectic of the molten material Figure 11a presents itself in a fiber-like format in the same way as in the reheated material Figure 11c, but this one is much more refined. In turn, the eutectic of the material processed via ultrasonic melt treatment appears fragmented Figure 11b; this difference is due to cavitation, that is, the application of acoustic waves to the liquid metal, causing it to mix in a more homogeneous manner as well as being able to break down the coarsest Si particles. To quantify the size of the eutectic, the image software ImageJ 1.40g was used, obtaining the respective average values of the eutectic area for each analyzed condition, being 29 μm2 for the eutectic of the conventional casting material, 19 μm2 for the eutectic of the processed via ultrasonic melt treatment and 10 μm2 for the eutectic of the reheated material.
Eutectic morphology in the material: conventional casting (a), casting with processing via UST for a time of 20 seconds (b) and thermally treated in the solid projection of 60% and time of 60 seconds (c).
For processing via ultrasonic melt treatment to be effective in refining the microstructure, some parameters stand out, namely the initial temperature of the horn, the types of flows or cavitation as a function of the submersion depth of the horn, and the position of the horn in relative to the center of the crucible during ultrasonic melt treatment. The initial temperature of the horn during the first moments of processing via ultrasound causes the formation of small crystals on its surface that will be dispersed in the molten metal when the horn is vibrated, thus increasing the nucleation sites. The types of flows or cavitation as a function of the submersion depth of the horn also significantly influence the degree of refinement of the material treated by ultrasound, with a depth of 3 mm presenting the largest active area treated (region directly below the horn). Finally, the positioning of the horn is another configuration that influences the processing via ultrasound because when the horn obeys the stipulated configuration (central part of the crucible), the secondary flow is practically eliminated, the dispersion becomes more violent, and practically only the region prevails more intense (active) under the tip of the horn16.
4. Conclusions
The results demonstrate that the Al7Si2.5Cu alloy processed via UST and subjected to reheating heat treatments at times of 0, 30, 60, 90, and 210 seconds and under two conditions of solid fractions of 45% and 60% presents good stability in the semisolid state, with the ideal time for reheating heat treatment being 60 seconds; the following conclusions are given below:
-
The primary globules present a near globular morphology from a heat treatment time of 0s for both solid fractions;
-
The circularity form factor presented values above 0.50 in most cases during the morphological evolution for both solid fractions, indicating that the structure is globular.
-
Comparing the microstructural evolution of the Al7Si2.5Cu alloy processed by UST versus chemical ultra-grain refining techniques of previous papers, the processing of the material via UST presents superior results in all analyzed scenarios. That is, it presents a drop of 26µm for the average size of primary globules; regarding the circularity form factor, all values are above 0.50, with a difference of 34% for the same material processed via ultra-grain refining technique for the average grain size under the 0s treatment condition there is a drop of 61 µm, however, under the 210 s treatment condition there is an even more significant drop of 192 µm; for the rheocast quality index (RQI) values there is an increase of approximately 50% in their values.
Therefore, processing via UST becomes an excellent alternative for refining the material's microstructural structure and preparing the raw material for the thixoforming processes.
5. Acknowledgments
The authors would like to thank the Brazilian research funding agencies FAPESP (São Paulo Research Foundation – projects 2018/11802-4 and 2022/05050-5) CNPq (National Council for Scientific and Technological Development – project 303299/2021-5) and CAPES (Federal Agency for the Support and Improvement of Higher Education) for providing financial support for this study. The authors are also indebted to the Faculty of Mechanical Engineering at the University of Campinas for the generous practical support and to the Federal Institute of Education, Science, and Technology of São Paulo - IFSP, Bragança Paulista campus.
6. References
- 1 Spencer DB, Mehrabian R, Flemings MC. Rheological behavior of Sn-15%Pb in the crystallization range. Metall Trans. 1972;3:1925-32.
- 2 Abramov V, Abramov O, Bulgakov V, Sommer F. Solidification of aluminium alloys under ultrasonic irradiation using water-cooled resonator. Mater Lett. 1998;37(1-2):27-34.
- 3 Zhang ZQ, Le QC, Cui JZ. Microstructures and mechanical properties of AZ80 alloy treated by pulsed ultrasonic vibration. Trans Nonferrous Met Soc China. 2008;18(1):113-6.
- 4 Luo SJ, Keung WC, Kang YL. Theory and application research development of semisolid forming in China. Trans Nonferrous Met Soc China. 2010;20(9):1805-14.
- 5 Atkinson HV. Modeling the semisolid processing of metallic alloys. Prog Mater Sci. 2005;50:341-412.
- 6 Torres LV, Proni CTW, Zoqui EJ. Comparison of morphological evolution of Al-7wt%Si-2.5wt%Cu alloy produced by direct chill casting/electromagnetic stirring and ECAP. Diffus Defect Data Solid State Data Pt B Solid State Phenom. 2016;256:17-24.
-
7 Torres L.V., Zoqui E.J. Mechanical properties of thixoformed Al–Si–Cu alloys. Int J Met Cast. 2024;18(1):769-88. http://doi.org/10.1007/s40962-023-01086-7
» http://doi.org/10.1007/s40962-023-01086-7 - 8 Liu D, Atkinson HV, Jones H. Thermodynamic prediction of thixoformability in alloys based on the Al-Si-Cu and Al-Si-Cu-Mg systems. Acta Mater. 2005;53:3807-19.
- 9 De paula LC, Tokita S, Kadoi K, Inoue H, Zoqui EJ. Analysis of the 355 aluminium alloy microstructure for application in thixoforming. Diffus Defect Data Solid State Data Pt B Solid State Phenom. 2019;285:277-82.
- 10 De paula LC, Zoqui EJ. Microstructural refinement of 355 aluminium alloys by ultrasonic melt treatment (UST). SN Applied Sciences. 2019;1:1-16.
- 11 ASTM International. E112-13: standard test methods for determining average grain size. West Conshohocken, PA; ASTM International; 2013.
- 12 Benati DM, Zoqui EJ. Effect of silicon on the thixoformability of Al-Si-Cu alloys. J Mater Eng Perform. 2014;23(9):3165-79.
- 13 Benati DM. Evaluation of thixoformability of Al-Xwt%Si-2.5wt%Cu-0.5wt%Mg alloys [thesis]. Campinas: State University of Campinas; 2008.
- 14 Chen TJ, Wang RQ, Ma Y, Hao Y. Semisolid microstructure evolution of AZ91D magnesium alloy refined by Al-Ti-B. Mater Res. 2011;14(4):532-40.
- 15 Narducci C Jr, Antunes AS, Abdalla AJ. Effect of the heterogeneous nucleation of the primary α-Al grain via the Al-4Nb-0.5B master alloy in Al-Si alloys with high Fe contents. Mater Res. 2021;24(4):e20200544.
- 16 Zhang L, Eskin DG, Katgerman L. Influence of ultrasonic melt treatment on the formation of primary intermetallics and related grain refinement in aluminum alloys. J Mater Sci. 2011;46(15):5252-9.