Abstract
In Al–3B master alloy, a higher fraction of borides is of the AlB2 variety, while in Al–8B alloy, the predominant species is the AlB12 phase. AlB12 is less stable than AlB2 and is engaged in exchange reactions, leading to the formation of transition metal diborides that subsequently settle at the bottom of the melt. Hence, AlB12 is involved more in precipitating transition elements than in refining the grain structure. With predominantly AlB12 particles, Al–8B master alloys are better suited for the removal of transition metal impurities in the manufacture of aluminium conductors. Al–3B, on the other hand, is a better grain refiner as the majority of its borides are of the AlB2 variety. Which of the two master alloys is used in grain refinement does not make a difference once the transition metal impurities have been precipitated. B is dedicated to grain refinement in an impurity free aluminium melt and produces exceptionally small equiaxed grains across the section of the samples.
Keywords
Introduction
Aluminium is the material of choice for power transmission lines owing to its good electrical conductivity, corrosion resistance and adequate mechanical properties in addition to light weight. 1 However, aluminium inevitably contains metal impurities that come from carbon anodes and bauxite used in the primary production of aluminium. 2 Transition metals such as Ti, Zr, V and Cr, in particular, are very detrimental for the electrical conductivity of aluminium. These impurities are taken care of during the manufacture of aluminium conductors with boron treatment that relies on Al–B master alloys. 3 When added into aluminium melts, B reacts with transition elements whose borides are more stable than those of aluminium 4 and are separated by gravity settling. 5
B additions were shown to be very effective also in grain refining aluminium foundry alloys.6–10 AlB2 dissolves in liquid aluminium at typical addition rates of 200 ppm B and then starts to crystallise from the melt. AlB2 has to be the primary phase, i.e. must crystallise before aluminium does, to offer grain refinement. This is made possible in Al–Si based alloys with at least 4 wt-Si. 11
While the performance of Al–B alloys in the grain refinement of aluminium foundry alloys has been well documented,6–14 the commercial Al–B master alloys are marketed only for the boron treatment of aluminium melts in the production of aluminium conductors. 15 They are produced with the well established halide salt process where KBF4 salt is added into molten aluminium. Commercial alloys have a number of drawbacks. 16 The borides are often agglomerated and heavily segregated. Commercial alloys are also contaminated with K–Al–F based spent salts. For all these reasons, it is of commercial interest to explore the potential of commercial Al–B master alloys in the grain refinement of aluminium foundry alloys. The present work was undertaken to evaluate the performance of the popular commercial Al–3B and Al–8B master alloys in this respect.
Experimental
Aluminium ingot with purity of 99·7 wt-Al, commercial purity silicon and magnesium were used to produce the Ti free AlSi7Mg0·3 alloy. Precalculated masses of the ingredients were melted in an electric resistance furnace and held at 800°C for 1 h to allow homogenisation and were finally cast into copper based permanent moulds for rapid solidification in order to avoid segregation.
The 1000 g of AlSi7Mg0·3 alloy thus obtained was melted in a resistance furnace, and the temperature of the melt was brought to 720°C for grain refinement performance tests. Master alloys, 6·55 g Al–3B and 2·45 g Al–8B, were then added to the melt. These are the exact amounts of the two grain refiners to bring the B concentration of 1 kg aluminium melt to 0·02 wt-B (200 ppm B). The melt was stirred with a graphite rod for 20 s right after master alloy addition. Samples were taken from the melt 5, 15, 30, 60 and 120 min after the addition and were solidified in small copper molds with a diameter of 25 mm and a height of 50 mm. Measures were taken to keep the temperature of the melt within 720±10°C during the entire process. A second set of grain refinement performance tests were performed to compensate for low B recoveries to ensure that the final B concentration of the melt was 200 ppm.
The commercial Al–3B and Al–8B alloys were analysed for their microstructural features. Samples sectioned from master alloy ingots were prepared with standard metallographic techniques. They were examined after etching with 0·5HF solution using an Olympus BX51M model optical microscope and a JEOL 6335F model field emission gun scanning electron microscope fitted with an Oxford INCA model energy dispersive X-ray analyser. The X-ray diffraction (XRD) patterns were recorded with a Shimadzu XRD 6000 diffractometer equipped with Cu Kα radiation at a scanning rate of 0·5 ° min−1. The chemical compositions of the cast samples thus produced were measured with an optical emission spectrometer. The samples were sectioned 20 mm from the bottom surface. They were etched with Poulton's reagent (12 mL HCl, 6 mL HNO3, 1 mL HF and 1 mL H2O) and then examined under a light microscope. The same series of samples were also anodised in Barker's solution, 5 mL HBF4 (48) in 200 mL water and then examined with an optical microscope under polarised light. Grain sizes were evaluated both visually and by the linear intercept method.
Results and discussion
The microstructures of the Al–3B and Al–8B master alloys are shown in Fig. 1a, c and b, d respectively. There appears to be two types of Al–B compound particles with two distinct morphologies: dark coloured discrete platelets embedded inside the aluminium matrix (Fig. 1c) and prismatic particle clusters that stick out of the aluminium matrix after electropolishing (Fig. 1d). There are occasionally K and F rich salt layers believed to originate from the poor decanting practice during production that has failed to remove the spent salt from the master alloy melt. The platelets were identified with energy dispersive X-ray analyser to be AlB2 compound particles, while the prismatic particles of the clusters were much richer in B and were thus claimed to be of the AlB12 variety (Fig. 2). This result is in agreement with that reported in (Ref. 17). Both AlB2 and AlB12 particles are heavily segregated, and the latter invariably agglomerated.
a, b optical and c, d scanning electron micrographs of a, c Al–3B and b, d Al–8B master alloys
Energy dispersive X-ray analyser analysis of a AlB2 platelets and b AlB12 clusters
There is a marked difference between the Al–3B and Al–8B master alloys regarding the microstructural features (Fig. 1). A higher fraction of Al–B compound particles are of the AlB2 variety in the Al–3B master alloy (Fig. 1a and c), while the predominant species in the Al–8B alloy is the AlB12 phase (Fig. 1b and d). These microstructural features are further evidenced by the XRD analysis of the respective master alloys (Fig. 3). The XRD analysis of the Al–3B alloy reveals predominantly AlB2 reflections, while the XRD spectrum of the Al–8B alloy presents, in addition to those of the AlB2 compound, prominent reflections of the AlB12 phase. The AlB12 compound is normally not expected to be present in either of these alloys as it takes as much as 45 wt-B to form this binary compound according to the Al–B binary phase diagram (Fig. 4). Segregation of B during the commercial halide salt process apparently leads to local enrichment of B and thus to the formation of B rich AlB12 compound.
X-ray diffraction spectra of a Al–3B and b Al–8B master alloys
Al–B binary phase diagram 18
It has been shown very recently that the aluminium foundry alloys must be Ti free for B addition to offer a remarkable grain refinement effect. 19 Ti is the standard alloying addition in most aluminium foundry alloys for grain size control but works against grain refinement when these alloys are to be grain refined with B. The grain refinement performances of these two master alloys were thus investigated with the Ti free version of the AlSi7Mg0·3 alloy at a B addition rate of 200 ppm. The results are shown for Al–3B and Al–8B master alloys in Fig. 5a and b respectively. There is a considerable grain refinement effect with the Al–3B master alloy 5 min after inoculation (Fig. 5a). The grain size estimated to be ∼1500 μm before B addition is reduced to ∼440±120 μm 5 min after the addition. The grains are fine equiaxed and uniform across the section of the sample. This effect is more or less retained for 30 min but starts to fade upon further holding. Finally, the grain refining efficiency is almost completely lost 120 min after inoculation. This sequence occurs in nearly the same fashion in the case of inoculation with Al–8B (Fig. 5b). However, the grains of the inoculated samples are relatively bigger and the fade starts relatively earlier with the Al–8B master alloy, suggesting that Al–8B is inferior as a grain refiner than the Al–3B master alloy.
Grain refinement performance of a Al–3B and b Al–8B master alloys at intended addition rate of 200 ppm B. Note that final B concentrations in cast samples were lower
It should be noted, however, that the B concentration in the AlSi7Mg0·3 alloy after the addition of Al–3B and Al–8B master alloys was 145 and 100 ppm respectively, considerably less than the intended addition rate of 200 ppm B. The reduced B recovery is claimed to be responsible for the poor grain refinement performance of the two alloys. The increasingly bigger grains in spite of a constant level of B suggest that an increasing fraction of the B measured in the inoculated samples is not involved in grain refinement. The relatively inferior performance of the Al–8B master alloy is consistent with the lower B recovery in the inoculated alloy. It is clear from the foregoing that 100 to 150 ppm B fails to offer any improvement over the current industrial grain refining practice that relies on approximately 1000 ppm Ti coming from the AlSi7Mg0·3 alloy ingot and up to 200 ppm Ti added in the form of Al–5Ti–1B. 10 It has been shown recently that the B addition rate in the grain refinement of aluminium foundry alloys with Al–B master alloys must be at least 200 ppm to achieve an average grain size smaller than 200 μm. 11
The lower than expected B recovery in the present case implies the boride settlement phenomenon typical of the boron treatment. 20 The low B recovery is indeed linked with the transition metals present in the AlSi7Mg0·3 melt as inferred from the concentration versus holding time curves shown in Fig. 6a. The marked decrease in the concentration of V and Ti upon B addition is evident (Fig. 6a). The Zr and Cr levels have also dropped once the Al–B master alloys were added into the melt. The underlying issue with the curves in Fig. 6 is that the decrease in transition metal concentrations is relatively higher with the Al–8B master alloy than with the Al–3B master alloy, in spite of the fact that both master alloys were added at the same addition rate of 200 ppm B. The higher rate of removal of the transition elements with the Al–8B master alloy explains why the B recovery was lower and the grain refining efficiency worse with the Al–8B alloy than with the Al–3B alloy.
a Change in concentration of B and Ti, V, Zr and Cr in AlSi7Mg0·3 alloy with time after addition of Al–3B and Al–8B master alloys and b change in removal efficiency of Ti, V, Zr and Cr in AlSi7Mg0·3 alloy with time after addition of Al–3B and Al–8B master alloys
Of the two commercial Al–B master alloys, Al–8B performs relatively poorly as a grain refiner, but is more efficient in removing transition metals. This behaviour of the Al–8B master alloy is attributed to its microstructure with predominantly AlB12 particles. AlB12 is less stable than AlB2 in aluminium melts (Fig. 7). Hence, it is expected to be engaged in exchange reactions, leading to the formation of transition metal diborides that subsequently precipitate at the bottom of the crucible owing to their much higher densities than that of molten aluminium. Hence, AlB12 is involved in precipitating transition elements rather than in refining the grain structure of the AlSi7Mg0·3 alloy.
Free energy of formation of borides in temperature range of 650–900°C 20
It is possible to estimate from Fig. 6b the efficiency of boron treatment in the removal of different transition elements from aluminium melts. ZrB2 is inferred from the Gibbs free energy for formation of the pure diboride compounds to be the most stable among the transition metal diborides (Fig. 7). The results from the present work show, however, that Ti and V are removed more readily than Zr, implying the order of the stability of transition metal diborides in molten aluminium to be TiB2, VB2, ZrB2 and CrB2, slightly different from that reported in Ref. 20. This may be attributed to the initial concentrations of transition metals in the aluminium melt. It is relatively easier for B atoms to find a Ti or V atom than a Zr atom since the concentration of the latter is only 11 ppm, much less than either Ti or V. The change in Cr concentration with time, on the other hand, is consistent with the literature. With a lower stability with respect to AlB2, CrB2 was reported to be difficult to remove with the boron treatment.20,21
An extra amount of Al–3B and Al–8B master alloys was added to the melt 2 h after the first time additions to compensate for the low B recovery. The average grain size in the inoculated AlSi7Mg0·3 alloy was reduced to 123±13 and 127±16 μm with the Al–3B and Al–8B master alloys respectively, once the B content of the molten AlSi7Mg0·3 alloy was raised to ∼200 ppm B (Fig. 8). This is a remarkable grain size for Al–Si foundry alloys and is retained for up to 2 h without any fading. It is fair to conclude that once the transition metals are precipitated, B is dedicated to grain refinement and produces exceptionally small equiaxed grains across the section of the samples. This exceptional grain refinement efficiency can be achieved by making B additions at once, instead of in two steps over time, provided that the addition rate is higher than 200 ppm to account for the transition elements present in aluminium foundry alloys.
Grain refinement performance tests after B concentration of molten aluminium alloy was raised to 200 ppm B to compensate for low B recovery in Fig. 5
It is concluded from the foregoing that for an efficient grain refinement, B in the Al–B master alloy must be in the form of AlB2 particles. AlB12 particles, on the other hand, are more effective in the removal of the transition elements as they are less stable than AlB2. The Al–B master alloys to be used in the grain refinement of aluminium foundry alloys must be B lean, i.e. Al−XB, where X<3, in order to ensure that the borides are of the AlB2 variety. Those master alloys used to remove transition elements in the manufacture of high conductivity aluminium, on the other hand, must be dominated with AlB12 particles and thus B rich. While it is possible to compensate for the low B recovery through extra Al–B master alloy addition, it should be kept in mind that the extra B is tied up in the transition metal borides and settle with time at the bottom of the crucible. This leads to the sludge formation that has to be taken care of during foundry operations. It is better if the aluminium foundry alloys to be grain refined with B additions are free of not only Ti but also of V, Zr and Cr to avoid the formation of boride sludge as well as to maximise B recovery.
Conclusions
A higher fraction of the borides is of the AlB2 variety in the Al–3B master alloy, while the predominant species in the Al–8B alloy is the AlB12 phase. Matrix grains are smaller and more resistant to fading when the AlSi7Mg0·3 alloy is inoculated with the Al–3B master alloy, while the grains of the cast samples are relatively bigger and the fade starts relatively earlier with the Al–8B master alloy. However, neither of the commercial Al–B master alloys offers any improvement over the commercial practice owing to a low B recovery linked with the transition metal impurities in the AlSi7Mg0·3 melt.
Al–8B is very efficient, however, in removing transition metals thanks to the predominance of AlB12 particles. AlB12 is less stable than AlB2 and is engaged in exchange reactions, leading to the formation of transition metal diborides that subsequently settle at the bottom of the melt. Hence, AlB12 is involved more in precipitating transition elements than in refining the grain structure. B in the Al–B master alloy must be in the form of AlB2 particles for an efficient grain refinement of aluminium foundry alloys if they contain transition metal impurities.
Nevertheless, further addition of Al–3B and Al–8B master alloys to raise the B concentration of the melt to 200 ppm produces an average grain size as small as 120 μm. Which of the two master alloys is used in grain refinement does not make a difference once the transition metal impurities have been precipitated. B is dedicated to grain refinement in an impurity free aluminium melt and produces exceptionally small equiaxed grains across the section of the samples.
Footnotes
Acknowledgement
It is a great pleasure to thank Mr F. Alageyik for his help with the experiments.