EFFECTS OF Sr AND B INTERACTIONS IN HYPOEUTECTIC Al-Si FOUNDRY ALLOYS

Light Metals 6 Edited by Travis J. Galloway TMS (The Minerals, Metals & Materials Society), 6 EFFECTS OF Sr AND B INTERACTIONS IN HYPOEUTECTIC Al-Si FOUNDRY ALLOYS Liming Lu, Arne K Dahle CSIRO Minerals, P.O. Box 88, Kenmore, Qld. 469, Australia Division of Materials, School of Engineering, The University of Queensland, Brisbane, Qld. 47, Australia Keywords: Al-Si alloys, Eutectic modification, Grain refinement. Abstract Strontium is the most widely used and a very effective element for modifying the morphology of eutectic silicon, while boron is commonly present in commercial grain refiners used for Al-Si alloys. Recent work on the combined additions of Sr and Al-B master alloys and of Sr and Al5TiB master alloy has suggested that negative interactions occur between Sr and B added through the grain refiners. However, the effects of and mechanisms for such negative interactions are not fully understood. This paper documents the experimental work and results aimed at determining the effects of Sr and B interactions on the solidification of hypoeutectic Al-Si foundry alloys. The mechanisms responsible for such negative interactions are further discussed. Introduction Modification of eutectic Si is usually accomplished by adding certain modifying elements, or chemical modifiers [-5]. Alternatively, a refined silicon morphology can be achieved in casting processes where the cooling rate is high [, ]. Strontium is used commercially to treat hypoeutectic Al-Si foundry alloys in order to refine and modify the eutectic Si from a coarse acicular to a fine fibrous morphology. However, Sr addition has been reported to promote columnar growth of primary Al dendrites, which is deleterious to the mechanical properties of the alloys [6, 7]. Therefore, subsequent grain refinement of primary Al becomes a necessary step to further materialize the beneficial effects of eutectic modification on properties. Grain refinement is standard practice for commercial wrought aluminium alloys to achieve fine equiaxed aluminium grains. It can be achieved by different means: fast cooling, heterogeneous nucleation, solute addition, melt agitation, etc. Due to their simplicity and efficiency, solute addition and heterogeneous nucleation, usually in combination, have become common industrial practice for these alloys. The well-known Al5TiB master alloy was developed based on this mechanism, containing Al Ti particles which dissolve quickly to release strongly segregating Ti into the melt and also potent TiB nucleant particles. There is some evidence in the literature to suggest that an AlTi layer forms on the surface of TiB particles before nucleation of Al, perhaps being critical for grain refining efficiency. While this grain refiner is extremely effective for wrought alloys, it is not as effective in foundry alloys. Foundry alloys are already rich in solute and thus do not require the solute effect from dissolving Al Ti, and it has been proposed that the TiB particles are rendered less potent in Al-Si foundry alloys due to the formation of TiSi compounds on the surface of TiB particles in the presence of high Si in the melts [8]. A new type of grain refining master alloy with a Ti:B weight ratio close to or below., corresponding to the stoichiometric value of TiB, such as Al.5Ti.5B and Al.Ti.5B, were hence developed to eliminate Al Ti particles in the master alloys since extra solute is not required in foundry alloys. The excess B in the master alloy can react with Ti in the melt to form TiB, thus providing additional potential nucleant particles. Recent work on the combined additions of Sr and Al-B master alloys [7, 9, ] and of Sr and Al5TiB master alloy [6] has suggested that negative interactions occur between Sr and grain refiners added. The grain refiners reportedly reduce the strontium available for modification. However, the impact of such interactions on primary Al grain refinement was not reported. The mechanisms for such negative interactions are also not fully understood. This paper aims to determine the effects and mechanisms of Sr and B interactions in an Al-Si-.5Mg alloy. Melt Preparation Experimental An AlSi.5Mg ternary alloy was selected as a base alloy in order to achieve a microstructure with a high volume fraction of eutectic. The base alloy melt was prepared from commercial purity aluminium, silicon and magnesium in an induction furnace and then transferred to an electric resistance furnace, which was held at 7 C. Upon thermal equilibrium, the melt was further alloyed using AlSr and AlTiB (Al5TiB or Al.5Ti.5B) master alloys. A weighed AlSr rod was added into the molten base alloy to ensure that the Sr level in the melt reached about 5 ppm. After being held for minutes for homogenization, the Srmodified melt was further treated by adding different amounts of AlTiB master alloy. It should be emphasized that some of the addition levels of grain refiners used in this research were much higher than those used industrially. Nevertheless, this will help to assess and understand the interactions between Sr and the grain refiners. Thermal Analysis In order to determine the solidification behaviour of both base alloy and treated alloys, thermal analysis was performed approximately min after each addition of AlTiB master alloys, or at a predetermined interval while the treated melt was held. For comparison, thermal analysis was also performed on the base alloy and the Sr-treated melt. Thermal analysis was performed using a preheated graphite crucible with a centrally located Type N thermocouple. The facility was carefully calibrated just prior to testing using high-purity aluminium to achieve a standard deviation below.5 C for the characteristic temperatures. Since 87

only the temperature differences, i.e. T, are used for discussion in most cases in this paper, an even better accuracy is expected. The cooling rate in the liquid just prior to nucleation of primary aluminium was about o C/s. Based on the thermal analysis, the characteristic temperatures, T N, T Min and T G, for both the primary Al and the eutectic reaction are readily determined. Sample Characterization Samples for chemical analysis were collected immediately after thermal analysis and analyzed using an optical emission spectrometer. Fully solidified TA samples were sectioned horizontally at the level of the thermocouple. These samples were mounted in resin and prepared using a standard procedure with a final polishing stage of.5 m colloidal silica suspension. The micrographs were taken in the median region of the section, mm from the bottom of the samples, in the unetched condition. Samples from the bottom of the melts at the end of holding were also collected and examined chemically and microscopically to determine any possible formation of heavy dense compounds between Sr and the AlTiB grain refiners during melt holding. While all polished samples were evaluated in an optical microscope, selected samples were also examined by scanning electron microscopy (SEM) and electron probe microanalysis (EPMA). Results and Discussion Microstructures and Fading Behavior of AlTiB Master Alloys Microstructures. In Figure, two other phases in addition to the Al matrix can be observed in the AlTiB master alloys; one is dark grey and very coarse and the other is light grey and extremely fine. According to the X-ray diffraction spectra in Figure, the coarse dark grey particles are believed to be Al Ti and the fine light grey particles are boride particles, mainly TiB. While Al5TiB was found to contain both TiB and Al Ti particles, only fine boride particles were observed in the Al.5Ti.5B master alloy, as shown in Figures and. Since there is an excess amount of B in the latter, it is believed to contain mixed borides, ie. (Alx, Ti(-x))B particles. This is in agreement with the results reported in the literature []. expected that these two master alloys will behave differently on addition to the Al-Si melts. Intensity, CPS 8 6 4 Al5TiB Al.5Ti.5B : Al (4-787) : TiB (755) : AlTi (659) 4 5 6 7 8 9 Theta Figure X-ray diffraction spectra of the AlTiB master alloys used. Fading Behavior. Figure presents the Ti and B concentrations as a function of time after the AlTiB master alloy was added into the unmodified melts. Fading of both elements was observed to some extent for both master alloys under the present conditions. The results also show that Ti and B in the melts fade in a similar manner for both master alloys. Since borides are common phases in both master alloys, it is believed that TiB particles (theoretical density: 4.495 g/cm []) introduced by the master alloys tend to settle down, consequently leading to a decrease in Ti and B contents in the melt. This is further confirmed by the samples withdrawn from the bottom layer of the AlTiB treated melts at the end of holding. The EDS spectra of the settled particles shown in Figure 4 confirm that the particles are TiB. Again the borides in Al5TiB are relatively coarse and separate, while they form clusters in Al.5Ti.5B. The TiB containing grain refiner has been reported to help remove inclusions by settling in molten aluminum alloys []. Ti in the melt, ppm 4 8 6 4 Al.5Ti.5B Al5TiB 5 5 4 Minutes after ATiB master alloy was added Figure Micrographs of the AlTiB master alloys Al5TiB and Al.5Ti.5B. The micrographs in Figure also suggest that the boride particles in both master alloys are evenly dispersed in the aluminium matrix. While the borides in Al5TiB are relatively coarse and separate, in Al.5Ti.5B they tend to form clusters. Therefore it is B in the melt, ppm 8 6 4 Al.5Ti.5B Al5TiB 5 5 Minutes after AlTiB master alloy was added Figure Ti and B concentrations as a function of time after the AlTiB master alloy was added into the unmodified melts. 88

Furthermore, while the B concentration in the Al5TiB-treated melt almost completely disappears within 5 minutes after the master alloy has been added, there is still considerable B remaining in the Al.5Ti.5B-treated melt (See Figure b). This is because, unlike Al5TiB, Al.5Ti.5B contains mixed borides, and some borides, i.e. AlB, are extremely fine [4] and less dense (theoretical density:.55 g/cm []), and they therefore remain dispersed in the melt longer. Comparing the changes in melt Ti level between these two master alloys over time indicates that the Al Ti particles present in Al5TiB do not settle down and instead are fully dissolved in the melt while the melt is held. This in turn contributes to a much higher recovery of Ti in the Al5TiB treated melt. T=TG-TN - - a.4 kg/tonne.8 kg/tonne.5 wt% Ti 5 5 75 5 5 Time after Al5TiB was added, min b - T=TG-TN.4 kg/tonne.8 kg/tonne.5 wt% Ti - 5 5 75 5 5 Time after Al.5Ti.5B was added, min Figure 4 BSE images and EDS spectra of samples withdrawn from the bottom layer of the Al5TiB and Al.5Ti.5B treated melts at the end of holding, indicating the presence of TiB in both samples. Figure 5 Difference between primary aluminium growth temperature and nucleation temperature of melts treated with combined additions of Sr and Al5TiB and Sr and Al.5Ti.5B as a function of time after combined additions. Effects of Combined Additions of Sr and AlTiB Grain Refiners Primary Al Solidification. From the thermal analysis, the characteristic temperatures of the primary Al reaction could be determined. The difference ( T) between the growth (T G ) and the nucleation temperature (T N ) is sometimes used to assess the grain refinement efficiency. T is plotted as a function of time after addition of the AlTiB grain refiners in Figure 5. It is clear that both melts display a similar grain refining potential over time at a certain addition level. Furthermore, the grain refining potential becomes stronger for both AlTiB master alloys as the addition level is increased. Figure 6 shows the effect of combined additions of Sr and AlTiB grain refiners on the microstructures of anodized samples solidified at min after combined additions. Since these samples have been anodized, the grain structure of the aluminium dendrites can be observed clearly under cross-polarised light. As expected from the thermal analysis, both melts contain a fine microstructure at all addition levels even after being held for min. The grain size decreases as the addition level increases for both AlTiB master alloys. While not predicted by the thermal analysis, both melts showed a slight coarsening of the aluminium grains during holding, which is a likely result of settling of nucelant particles over time as discussed earlier..4 kg/tonne.8 kg/tonne (c).5 wt% Ti Figure 6 Effect of combined additions of Sr and AlTiB master alloys, Al5TiB (left) and Al.5Ti.5B (right) respectively, on the grain structure of samples solidified min after addition. All images were taken at the same magnification. 89

Eutectic Solidification. The characteristic temperatures of the Al+Si eutectic reaction were also determined from the thermal analysis. In Figure 7, the degree of depression of the eutectic thermal arrest due to the Sr treatment, TG TG, t TG,, is plotted as a function of time after combined additions of Sr and AlTiB grain refiner. It is clear in Figure 7 that both melts show good modification efficiency when the addition level of AlTiB grain refiner is low at.4 and.8 kg/tonne. A well-modified structure was obtained with both grain refiners. However, at the highest addition level of.5%ti, the Al.5Ti.5B treated melt almost completely loses its modification 5 min after the grain refiner was added, while the Al5TiB treated melt shows a negligible change in eutectic thermal arrest over a time span of min after combined addition. The microstructural observations of solidified samples further support the observations from the thermal analysis. Microstructures of samples cast at 5 and min after addition of the AlTiB grain refiners (.5%Ti), corresponding to Points and in Figure 7, respectively, are shown in Figure 8. It is not difficult to conclude that, while a slight coarsening of eutectic Si occurs over time for the modified melt treated with the Al5TiB grain refiner, the Al.5Ti.5B grain refiner has a strong poisoning effect on the Sr modification under the present conditions. The mechanism for this poisoning effect is discussed later in this paper. eutectic solidification than on the primary Al reaction. As a result, the melt treated with combined additions of Sr and Al.5Ti.5B still shows good grain refinement efficiency even after completely losing its eutectic modification. TG, C -.4 kg/tonne.8 kg/tonne.5 wt% Ti a 5min min Figure 8 Microstructures of samples solidified from the melts at 5 min (left) and min (right) after combined additions (.5%Ti) of Sr and Al5TiB, and Sr and Al.5Ti.5B. All images were taken at the same magnification. TG, C - - - 5 5 75 5 5 Time after Al5TiB was added, min.4 kg/tonne.8 kg/tonne.5 wt% Ti 5 5 75 5 5 Time after Al.5Ti.5B was added, min Figure 7 Depression of eutectic Al-Si thermal arrest of modified melts treated with Al5TiB and Al.5Ti.5B as a function of time after combined additions. The results clearly demonstrate negative interactions between Sr and Al.5Ti.5B, and these become more pronounced as the addition level of Al.5Ti.5B increases. It was also found that the negative interactions have a much more profound effect on the b Mechanisms Responsible for the Negative Interactions between Sr and Al.5Ti.5B Figure 9 shows the Sr concentration in the melts as a function of time for the highest addition level of AlTiB grain refiners used in this work (.5% Ti). The concentration at zero time in Figure 9 is the Sr level in the melts prior to addition of either of the AlTiB grain refiners. While Sr fading was observed for both melts, the results clearly demonstrate that the melt treated with Al.5Ti.5B loses its Sr much more quickly, particularly in the initial stage after addition, compared with the melt treated with Al5TiB. This explains the quick loss of eutectic modification in the Al.5Ti.5B treated melt, i.e. there is insufficient free Sr in the melt to refine all the eutectic silicon. It is well known that molten Al-Si alloys can lose their Sr through surface oxidation and/or vaporization. However, whichever mechanism is operating, the Sr concentration in the melt is expected to decrease gradually over time. Results from the experiment where the melt was treated only with Sr are also included in Figure 9 (dotted line). It is readily apparent that both the Sr-only treated melt and the melt treated with (Sr+Al5TiB) behave similarly, showing a slow and almost linear loss of Sr over time. Surface oxidation and vaporization are believed to operate in these two melts. Addition of the Al5TiB grain refiner does not significantly accelerate the loss of Sr in the melt. However, Figure 8

9 shows that the melt treated with Al.5Ti.5B behaves differently, losing its Sr rapidly in the initial stages after addition of Al.5Ti.5B. It is believed that other mechanisms in addition to surface oxidation and vaporization are responsible for this initial rapid loss of Sr. 45 4 5 Sr+Al.5Ti.5B Sr+Al5TiB Sr only addition Formation of such dense SrB 6 particles explains the poisoning effect observed between Sr and the Al.5Ti.5B grain refiner. It is assumed that a certain component in the grain refiner will react with Sr while the treated melt is held. Since no discernable changes in settled TiB particles was observed between Figure 4 and Figure, it is believed that there is no reaction between TiB and Sr, otherwise such a poisoning effect should also be observed in the melt treated with combined addition of Sr and Al5TiB. Sr in the melt, ppm 5 5 5 4 6 8 4 Time after ATiB is added, min Figure 9 Sr concentrations as a function of time after AlTiB grain refiners were added to achieve.5%ti in the melts (solid lines). The dotted line refers to the case of Sr modification without any grain refiner. In order to understand the mechanisms involved in the quick fade of Sr in the Al.5Ti.5B treated melt, chemical analysis was conducted on samples extracted from the bottom layer of the remaining melt at the end of holding. A higher than expected concentration of Sr was obtained. Since Sr is completely miscible in the base alloy under the present conditions, it would not be expected to settle to the bottom of the melt unless it combines with other elements to form dense compounds. Microstructural observations of the fully solidified sample from the bottom layer of the melt treated with Sr and a high addition level (.5%Ti) of the Al.5Ti.5B grain refiner confirmed the presence of such dense phases. Figure shows a typical BSE image of the sample from the bottom layer of the melt treated with combined addition of Sr and Al.5Ti.5B at the end of holding. While clusters of extremely fine particles, similar to those observed in Figure 4, believed to be settled TiB particles, were observed in Figure a, the melt also contains a number of bright, almost spherical particles with a diameter of approximately 8- m. Further characterization of these particles was conducted using EDS and EPMA in order to determine their chemical composition. Since the Si K line is at around.79 kev, which is very close to the Sr L (.86 kev), a further scan was conducted at kv to reveal the Sr K line. The EDS spectra in Figures b and c suggest that the bright spherical particles in the melt treated with Al.5Ti.5B predominantly contain B, Sr and Si. The carbon peak in the spectra is due to the carbon coating of the samples to improve the electrical conductivity. EPMA results in Table further suggest that the composition of the bright spherical particles in the Al.5Ti.5B treated melt is very close to the stoichiometric value of SrB 6. (c) Figure BSE image of the sample from the bottom layer of the melt treated by combined addition of Sr and Al.5Ti.5B and EDS spectra measured from the bright spherical particles in the BSE image at 8 and kv (c), respectively. In addition to TiB, Al.5Ti.5B is expected to introduce excess B into the melt. This excess B, most likely in the form of AlB, is expected to react with Sr to form SrB 6 in the following reaction: [Sr] + AlB (S) SrB 6 (S) + [Al] () These freshly formed SrB 6 particles are relatively high density (theoretical density:.4 g/cm []) and are therefore expected to settle down in the melt. The SrB 6 particles continue to grow while settling. As a result, a Sr and B rich intermetallic compound is expected in the bottom layer of the melt and the concentration 8

of solute Sr in the melt decreases drastically over time for the Al.5Ti.5B treated melt, as was observed in these experiments. Table Average composition (wt%) of the coarse particles in the sample from the bottom layer of the melt treated with combined addition of Sr and Al.5Ti.5B Sr Si Al Ti B (Sr + Al.5Ti.5B) treated melt 46.5.5.. 48.6 Conclusions The microstructures and fading behaviour of AlTiB grain refiners were studied using optical/electron microscopy and X-ray diffraction. While the Al5TiB master alloy was found to contain TiB and Al Ti particles, only clusters of fine boride particles were observed in the Al.5Ti.5B master alloy. The boride particles in Al.5Ti.5B are believed to be mixed TiB and AlB particles. TiB particles in both grain refiners are dense and tend to settle down on addition to the melt, consequently leading to a decrease in Ti and B in the melt during holding. The effects of combined addition of Sr and AlTiB grain refiners on the Al+Si eutectic and primary Al solidification were also studied. While slight coarsening of both eutectic Si and primary aluminum grains occurs during holding, no obvious interactions are observed between Sr and the grain refiners when the addition level of the grain refiners is low. A well modified and grain refined microstructure was obtained with both grain refiners at low addition levels. No poisoning effect was observed between Sr and Al5TiB even at higher addition levels of the grain refiner. A slight coarsening of eutectic Si over time is believed to be due to the slow and continuous loss of Sr as a result of surface oxidation and vaporization. The interactions between Sr and Al.5Ti.5B became stronger as the addition level of the grain refiner increased. It was found that these interactions have a much more profound effect on the eutectic than the primary Al solidification. As a result, the melt treated with combined addition of Sr and Al.5Ti.5B still showed good grain refinement efficiency even after losing its modification completely. The mechanism responsible for the negative interactions between Sr and the Al.5Ti.5B grain refiner is also proposed and further confirmed by microanalysis using SEM and EPMA. It involves the reaction between AlB and Sr resulting in settling of SrB 6 particles, thus reducing the amount of solute Sr in the melt available for modification of eutectic silicon. References. G.K. Sigworth, "Theoretical and practical aspects of the modification of Al-Si alloys," AFS Trans., 66(98), 7-6.. L.M. Hogan and M. Shamsuzzoha, "Crystallography of the flake-fiber transition in Al-Si eutectic," Materials- Forum, (987), 7-77.. S.Z. Lu and A. Hellawell, "The mechanism of silicon modification in aluminium-silicon alloys: Impurity Induced Twinning," Metallurgical and Materials Transactions A, 8A(987), 7-7. 4. J.E. Gruzleski, "The art and science of modification," AFS Transactions, 64(99), 678. 5. J.E. Gruzleski and B.M. Closset, The treatment of liquid aluminium-silicon alloys (Des Plaines, Illinois, USA: American Foundrymen's Society, 999), -56. 6. H. Liao, Y. Sun, and G. Sun, "Effect if Al-5Ti-B on the microstructure of near-eutectic Al-.%Si alloys modified with Sr.," Journal of Materials Science, 7(), 48995. 7. H. Liao and G. Sun, "Mutual poisoning effect between Sr and B in Al-Si casting alloys," Scripta Materialia, 48(), 5-9. 8. Y.C. Lee, A.K. Dahle, D.H. StJohn, and J.E.C. Hutt, "The effect of grain refinement and silicon content on grain formation in hypoeutectic Al Si alloys," Materials Science and Engineering: A, 59()(999), 4-5. 9. K. Nogita, S.D. McDonald, and A.K. Dahle, "Effects of boron-strontium interactions on eutectic modification in Al-mass%Si alloys," Materials Transactions, 44(), 6995.. K. Nogita and A.K. Dahle, "Effects of boron on eutectic modification of hypoeutectic Al-Si alloys," Scripta Materialia, 48()(), 7-.. P. Cooper, A. Hardman, D. Boot, and E. Burhop, "Characterisation of a new generation of grain refiners for the foundry industry" (Paper presented at Light Metals, San Diego, USA, ).. ICDD, JCPDS - International centre for diffraction data, PCPDFWIN Version., 999.. T. Gudmundsson, T.I. Sigfusson, D.G. McCartney et al., "Efficiency of holding time in a casting furnace" (Paper presented at Light Metals Conference, USA, 995), 85558. 4. W. Schneider and P. Cooper, "Influence of AlTiB master alloy type and casting conditions on grain refinement of aluminum alloys" (Paper presented at Eighth Australasian Conference on Aluminium Cast House Technology, Brisbane, ), 67. Acknowledgements The authors would like to acknowledge the financial support from the Cooperative Research Centre for Cast Metals Manufacturing (CAST). CAST was established and is supported by the Australian Government s Cooperative Research Centres Program. 8