- Email: [email protected]
Effect of conventional and rheocasting processes on microstructural characteristics of hypereutectic Al–Si–Cu–Mg alloy with variable Mg content
Journal of Materials Processing Technology 210 (2010) 767–775
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Effect of conventional and rheocasting processes on microstructural characteristics of hypereutectic Al–Si–Cu–Mg alloy with variable Mg content Alireza Hekmat-Ardakan, Frank Ajersch ∗ École Polytechnique de Montréal, Dép. de Génie Chimique, P.O. Box 6079, Centre-ville, Montreal, Quebec, Canada H3C 3A7
a r t i c l e
i n f o
Article history: Received 30 August 2009 Received in revised form 4 January 2010 Accepted 9 January 2010
Keywords: Aluminium alloys Rheocasting Eutectic solidification Magnesium Casting
a b s t r a c t Hypereutectic A390 alloy (Al–17%Si–4.5%Cu–0.5%Mg (wt%)) and similar alloys with the addition of 6 wt% and 10 wt% Mg were investigated using conventional and rheocast processes. Conventional castings were produced with two different cooling rates: −0.15 ± 0.05 ◦ C s−1 and −1.0 ± 0.2 ◦ C s−1 . The rheocast samples were subjected to a rotation speed of 260 rpm during solidification with a cooling rate of −0.15 ± 0.05 ◦ C s−1 . The microstructure of the conventional casting for the high cooling rate samples was found to be finer than for the low cooling rate samples. With high Mg content, the primary silicon transforms to primary Mg2 Si and the nature of the eutectic reaction is also changed from a binary (Al + Si) to a ternary (Al + Si + Mg2 Si) reaction. The morphology of eutectic silicon distributed in matrix microstructure significantly changes from large individual flakes to a fine, skeleton network with Chinese script morphology, similar to the eutectic Mg2 Si. This is caused by the decrease of the eutectic formation temperature resulting from the addition of Mg. The matrix microstructure of the rheocast samples was found to contain fragmented eutectic silicon together with globular alpha aluminium phase particles. The morphology of eutectic silicon in rheocast samples becomes finer with increasing Mg content. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Semisolid metal (SSM) processing can be widely used as a preferred fabrication route for near or net-shaped castings where partially solidified metal with a globular structure (SSM slurry) is injected into a mold, instead of liquid metal casting. SSM processing is generally identified as either rheoprocessing or thixoprocessing. The rheo-route involves preparation of SSM slurry from the liquid phase by stirring the alloy during solidification and transferring directly into a die or mold for component shaping. The thixo-route involves the preparation of a feedstock material with a fine equiaxed or globular structure. The feedstock is then reheated to a temperature between solidus and liquidus (mushy zone) in order to produce a semisolid material which is injected into the mold. A detailed review of SSM processing was described by Nafisi and Ghomashchi (2005). Alloys currently used for SSM processing are mainly conventional Al–Si casting alloys. Hypereutectic Al–Si alloys with more than 11.7 wt% Si offer the possibility of producing in situ metal matrix composites (MMCs) with excellent wear and corrosion resistance where the silicon acts as the reinforcing phase, as indi-
∗ Corresponding author. Tel.: +1 514 340 4711x4533; fax: +1 514 340 4468. E-mail address: [email protected] (F. Ajersch). 0924-0136/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2010.01.005
cated by Kapranos et al. (2003). These composites are a class of materials with an excellent combination of mechanical and physical properties for various applications in the automotive industry. Baiqing et al. (2003) indicated that the high strength and wear resistance of these alloys are attributed to the presence of both hard primary and eutectic Si particles. Zhang et al. (2000, Material Science Engineering A) have also shown that hypereutectic Al–Si alloys with high Mg content can also be considered as in situ Al–Mg2 Si MMCs with a high potential for wear resistance because of the excellent physical and mechanical properties of Mg2 Si intermetallic phase that is formed on solidification. The Mg2 Si phase has a high melting point (1085 ◦ C), very low density (1.99 × 103 kg/m−3 ), high hardness (4.5 × 109 Nm−2 ), a low thermal expansion coefficient (7.5 × 106 K−1 ) and a reasonably high elastic modulus of 120 GPa (Jiang et al., 2005). Consequently, both Al–Si and Al–Mg2 Si composite alloys can be used as wear resistant materials. In terms of properties and solidification behaviour, many similarities exist between the Mg2 Si and Si phase particles as compared by Zhang et al. (2000, Materials and Design). In the present work, the effects of conventional and rheocast processing on the microstructural characteristics, particularly on the morphology of eutectic silicon, of hypereutectic A390 alloy (Al–17%Si–4.5%Cu–0.5%Mg (wt%)) were investigated and compared to the microstructure of hypereutectic alloys with 6% and
768
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
Table 1 Chemical composition of test alloys and AZ91 Mg alloy (wt%). Test alloy
Al
Si
Cu
Mg
Fe
Mn
Zn
P
Ti
A390 6% Mg 10% Mg
Bal. Bal. Bal.
16.7 16.83 17.14
4.58 4.12 4.32
0.58 6.08 9.74
0.32 0.28 0.25
0.02 0.04 0.05
0.01 0.015 0.02
0.0003 0.0002 0.0002
0.02 0.02 0.019
AZ91
Mg
Al
Mn
Zn
Si
Ni
Cu
Bal.
9.3
0.12
0.62
0.02
0.0006
0.0007
10% Mg. The addition of Mg and the formation of both primary and eutectic Mg2 Si phases can also affect the morphology of eutectic silicon. Briefly, this work investigates the effects of the chemical composition, the casting process and the effect of cooling rate on the microstructure of alloy. 2. Experimental procedures The 6 wt% and 10 wt% Mg alloys were produced by adding AZ91 (Mg-base) alloy to A390 alloy. The chemical analysis of the test alloys and AZ91 Mg alloy is presented in Table 1. In order to account for the oxidation loss, an additional amount of 20% Mg (as AZ91) was added to the melt. The AZ91 alloy was wrapped in aluminium foil and added to the liquid A390 alloy at 750 ◦ C. Silicon and copper were also added to the melt in order to attain the chemical composition of these elements at the same value as for the A390 composition. The melt was degassed with argon and stirred for 2 min before pouring into a steel mould to form three ˚35 mm × 120 mm cylindrical samples. 2.1. Conventional casting After preparing the samples, specimens of 60 g portions were transferred into a graphite crucible with 28 mm inner diameter and 5 mm wall thickness and heated to 700 ◦ C using an electric resistance furnace, degassed with argon and stirred for 2 min before sampling. Samples were prepared with two cooling rates: high cooling rate (HCR) and low cooling rate (LCR). For the HCR samples,
Fe 0.0046
the graphite crucible and contents were transferred to the cooling station where a K-type thermocouple was quickly immersed into the melt at a position about 10 mm from the bottom at the center of the crucible. Temperatures were recorded using a data logging system at 15 per second intervals. For the HCR cooling tests, the cooling rates averaged −1.0 ± 0.2 ◦ C s−1 . For the low cooling rate (LCR) tests, the samples were cooled in the furnace at a controlled cooling rate of −0.15 ± 0.05 ◦ C s−1 . Metallographic specimens for the conventionally cast samples were prepared by sectioning transversely at a level 10 mm from the bottom of crucible.
2.2. Rheocasting Fig. 1a shows the schematic diagram of the apparatus used for rheocasting process. A spirally grooved cylindrical agitator was used for producing the rheocast samples. Each rheocast test consisted of 150 g of alloy contained in a graphite crucible with an inner diameter of 41 mm, 10 mm wall thickness and a height of 120 mm. Subsequently, the samples were heated to 700 ◦ C and then held at this isothermal temperature for a period of 5 min. After this period, the alloy was continuously cooled in the furnace at a rate of −0.15 ± 0.05 ◦ C s−1 and sheared at an average rate of ˙ ave = 52 s−1 (rotation speed = 260 rpm) as shown in Fig. 1a. Two K-type thermocouples located in the crucible wall were used to measure the temperature near the surface and near the bottom of the melt. Argon gas was introduced into the furnace chamber in order to prevent sample oxidation. When the upper limit of the viscometer torque was reached, the tests were stopped.
Fig. 1. (a) Schematic diagram of the apparatus used for the preparation of rheocast samples and (b) the location of the sampled regions.
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
769
Fig. 2. Cooling curves for A390, 6% Mg and 10% Mg samples for high cooling rate (HCR) measured by a thermocouple at the center of graphite crucible during conventional casting.
The rheocast specimens were prepared by sectioning the samples as shown in Fig. 1b. The specimens were polished with grinding paper up to 1200 followed by micro-polishing, using monocrystal diamond suspension (0.5 m). All samples were etched by using 0.5% HF solution agent. Optical and scanning electron microscopy (SEM) with EDX (energy dispersive X-ray) analysis were carried out to characterize the microstructure. The Clemex software was used to analyse the particle sizes observed in the microstructure. 3. Results 3.1. Conventional casting process Fig. 2 shows the cooling curves for A390, 6% Mg and 10% Mg alloy samples for the air cooled (HCR) condition. The temperature was measured by a thermocouple at the center of the melt during conventional casting. Fig. 3 compares the detail of the eutectic segment of the cooling curves for HCR and LCR cooling of conventionally cast A390, 6% Mg and 10% Mg alloys. The figure shows that for additions up to 6% Mg, the eutectic reaction temperature is reduced significantly and becomes almost constant for additions between 6% and 10% Mg. Fig. 3 also indicates that the values of the eutectic temperature are very similar for both LCR and HCR conditions. The values of the eutectic reaction temperature obtained experimentally by cooling curves for both cooling conditions are presented in (Table 2). Fig. 4 shows the BSE (back scattered electron) image of A390 alloy for the LCR conventional casting sample showing the form of the intermetallic compounds. The microstructure consists of primary polygonal silicon phase and the eutectic matrix consists mainly of binary Al + Si together with the intermetallic phases, CuAl2 (), Cu2 Mg8 Si6 Al5 (Q) and Al5 FeSi (). Other studies have also observed these microstructures with the presence of the Q phase associated with the phase (Li et al., 2006; Elmadagli and Alpas, 2006). However, the main intermetallic phase in these alloys is the phase. It should be noted that tested alloys also contain small amounts of Fe, Zn, etc., which also react to form of intermetallics. The Table 2 Comparison of the eutectic reaction temperature determined by thermodynamic prediction using FactSage and the measurement of cooling curves at LCR (−0.15 ± 0.05 ◦ C s−1 ), HCR (−1.0 ± 0.2 ◦ C s−1 ) for three alloys.
A390 6% Mg 10% Mg
Predicted (◦ C)
Measured (◦ C) LCR
Measured (◦ C) HCR
566.2 549.7 549.2
563 ± 1 545 ± 1 543 ± 1
564 ± 1 544 ± 1 542 ± 1
Fig. 3. The eutectic segment of the cooling curves at HCR and LCR for (a) A390, (b) 6% Mg and (c) 10% Mg alloys during conventional casting.
platelet Fe-containing phase, -Al5 FeSi, is the most common Fe intermetallic phase and contributes to the decrease of the mechanical properties of Al–Si cast alloys as reported by Sreeja Kumari et al. (2007), Taghiabadi et al. (2008), and Nafisi et al. (2006). With increasing Mg content, the Mg2 Si intermetallic phase first appears in the matrix in the form of a eutectic morphology and then as a primary phase. The microstructure of these three alloys for LCR and HCR conditions are compared in Fig. 5. For both cooling conditions, the morphology of grey primary silicon particles remained polygonal, whereas Mg2 Si crystals (dark particles) are either in the form of dendritic crystals with entrapped liquid as shown by Jiang et al.
770
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
Fig. 4. BSE (back scattered electron) image of A390 alloy for conventionally cast low cooling rate (LCR) sample. (a) Large faceted primary silicon (1). The matrix consists of binary Al + Si eutectic as well as the additional phases found in location A. (b) Magnified zone of location A, showing -CuAl2 (2), Al (3), Cu2 Mg8 Si6 Al5 (Q phase) (4) and -Al5 FeSi (5).
Fig. 5. Overall microstructure of A390 alloy (a and b), 6% Mg alloy (c and d) and 10% Mg alloy (e and f) for different cooling rates for conventionally cast samples. Images a, c and e correspond to LCR samples and b, d and f the HCR samples. The light grey phase is silicon, the dark phase is Mg2 Si and the white phase is ␣-Al.
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
(2005) and Qin et al. (2007) or in the form of compact polygonal shapes. The microstructure of the high cooling rate samples (Fig. 5b, d, and f) is finer than that of the low cooling rate samples (Fig. 5a, c, and e). Fig. 6 clearly shows that the eutectic morphology for both the HCR and LCR samples is significantly changed with the addition of Mg, particularly at 6% Mg when compared to A390 alloy. The eutectic silicon together with the dendritic ␣-Al shows a coarse flake structure for A390 alloy cooled at both low and high rates (Fig. 6a and b). However, the eutectic microstructure for 6% Mg alloys (Fig. 6c and d) and 10% Mg alloys (Fig. 6e and f) is similar and the eutectic silicon transforms into a fine, skeleton network with Chinese script morphology, similar to the form of eutectic Mg2 Si, as was also observed by Jiang et al. (2005). The eutectic matrices of A390 and of the 6% Mg alloy at the HCR condition are compared in Fig. 7 using SEM images. The morphology of the eutectic silicon for A390 and 6% Mg alloys for the HCR condition (shown in Fig. 7) is also compared in the SEM image of Fig. 8 where the surface was deep etched using 0.5% HF solution. These figures clearly show that the eutectic silicon in the 6% Mg alloy is much finer than in the A390 alloy.
771
3.2. Rheocasting process As described in the experimental procedure, the cooling condition for the rheocast samples is similar to the LCR condition since the samples are stirred during continuous cooling solidification. The tests were stopped at the upper limit of the viscometer torque. This maximum torque was attained at different temperatures for each of the alloys. For A390 alloy this occurred at 541 ◦ C (fliquid = 25.1%) and for the 6% and 10% Mg alloys the temperatures were found to be 533 ◦ C (fliquid = 26.3%) and 530 ◦ C (fliquid = 23.3%), respectively. Therefore, the solidified structure of the remaining liquid should be similar to the conventional cooling structure without shearing effect at a solid fraction of about 25%. A significant segregation of primary phases was detected in the microstructure of the rheocast samples due to the shear force imposed on the slurry during solidification. As a result, the microstructure is different for two distinct regions. The region with a radial distance between 2 mm and 3 mm from the crucible wall (near crucible zone shown in Fig. 1b) is a region where the most of the primary phase particles are concentrated, as shown in left side images of Fig. 9 for A390, and the 6% and 10% Mg alloys. Wang
Fig. 6. The eutectic microstructure of A390 alloy (a and b), 6% Mg alloy (c and d) and 10% Mg alloy (e and f) alloys for different cooling rates for conventionally cast samples. Images a, c and e correspond to LCR samples and b, d and f the HCR samples. Light grey phase is silicon, dark phase is Mg2 Si and the white phase is ␣-Al grains.
772
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
Fig. 7. Comparison of eutectic silicon phase (arrows) and microstructures of conventionally solidified alloys at HCR condition for (a) A390 and (b) 6% Mg, at the same magnification.
et al. (1994) have shown that during the solidification of hypereutectic Al–Si/SiC composite, the SiC particles in the composite melt serve as the substrates for primary Si phase nucleation and all the primary Si particles form heterogeneously on the SiC surface. In the melt-stirred condition, all the Si particles become agglomerated and are attached by the SiC particles. Sundarrajan et al. (1998) have also explained this phenomenon based on the large undercooling required for the growth of polygonal Si crystals along low-energy, atomically smooth and slow-growing (1 1 1) planes. It was shown that ceramic composites such as aluminium oxide, silicon carbide and graphite can eliminate such large undercooling and result in preferential nucleation and growth of Si on these materials. In this case undercooling is eliminated as a result of capillary (interfacial) forces which bend the liquid–solid interface near the contact line of (1 1 1) facets with the composite surface. This phenomenon can explain the segregation of Si and Mg2 Si particles resulting for interfacial forces which are greater than the centrifugal forces even though the density of the liquid phase is higher than that of the particles. The region next to the agitator (near agitator zone shown in Fig. 1b) with a radial distance between 5 mm and 6 mm from agitator, shows no primary phases. The globular ␣-Al grains observed in this region form aggregates as shown in right side images of Fig. 9. When the primary phases in the region near the crucible wall are compared with non-sheared LCR samples in Fig. 5, it is clear that the segregation of primary phases caused by the shear forces is much higher in the rheocast (sheared) samples. The primary silicon of sheared A390 alloy (Fig. 9a) also agglomerates. Fig. 10 indicates that all three sheared samples show an increase of the size of the silicon particles when compared to the non-sheared samples as measured by image analysis. This implies that the shear promotes the coarsening of silicon particles. Lee et al. (1995) have also
reported that significant fragmentation and changes of primary Si morphology occur for an Al–15.5%Si alloy during the isothermal stirring but little change was observed when a sample was continuously cooled at a shear rate of ˙ = 200 s−1 and at a cooling rate of −0.03 ◦ C s−1 . This shear rate however is able to fragment the dendritic morphology of the Mg2 Si primary phase crystals with entrapped liquid as shown in Fig. 5c and e, forming smaller, compact polygonal shapes as shown in Fig. 9c and e. Fig. 11 compares the average particle size of the primary Mg2 Si particles for the 6% and 10% Mg alloys for conventionally and rheocast samples. The reduction in particle size is primarily due to the shear forces resulting in the fragmentation of the larger eutectic structure of the conventionally solidified samples. In the region near the agitator, the agglomerated ␣-Al grains are present only in the matrix of the eutectic phase as shown in Fig. 12a, c and e for each of these alloys. Fig. 12b, d and f represents the regions between the agglomerations consisting mostly of the eutectic region. For the A390 alloys (Fig. 12a and b), there is no significant difference in the microstructure of ␣-Al grains in matrix. However, for the high Mg-containing alloys of 6% and 10% Mg, respectively shown in Fig. 12c and d, and e and f, the microstructural difference is more pronounced. The ␣-Al grains are generally surrounded by intermetallic phases, mainly by the -phase. It can be clearly observed that the eutectic silicon for the sheared A390 sample (Fig. 12b) becomes fragmented when compared to the non-sheared sample shown in Fig. 6a. For the high Mg-containing alloys, the skeleton network of eutectic silicon observed for the LCR condition, is completely destroyed and fragmented, as shown in Fig. 12d (6% Mg) and Fig. 12f (10% Mg). For the rheocast samples, the morphology of eutectic silicon becomes finer with increasing the Mg content.
Fig. 8. Deep etched images showing the morphology of the eutectic Si in (a) A390 alloy and (b) 6% Mg alloy for conventionally HCR solidified cast samples at the same magnification.
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
773
Fig. 9. The rheocast microstructure of A390 alloy (a and b), 6% Mg alloy (c and d) and 10% Mg alloy (e and f). Images a, c and e represent the area near the crucible wall and images b, d and f represent the area near the agitator.
Fig. 10. Average particle size of primary silicon for conventional and rheocasting processes for three alloys.
Fig. 11. Average particle size of primary Mg2 Si for conventional and rheocasting processes for alloys with 6% and 10% Mg alloys.
774
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
Fig. 12. The microstructures near to the agitator for stir cast samples. A390 alloy (a and b), 6% Mg alloy (c and d) and 10% Mg alloy (e and f). Images a, c and e show agglomeration of ␣-Al grains and images b, d and f illustrate the eutectic microstructure between the agglomeration of ␣-Al grains.
4. Discussion The results of this study have shown that Mg additions to hypereutectic Al–Si alloy have a significant effect on the solidification structure of conventionally cast and rheocast alloys. As shown in Fig. 7, the change in morphology of eutectic silicon is due to the decrease of the eutectic temperature with the addition of Mg. Nafisi and Ghomashchi (2006) also investigated the effect of the reduction of eutectic temperature on the modification of the eutectic silicon morphology of A356 Al–Si alloy when Sr is added as modifier. In fact, they attributed this temperature reduction to the eutectic nucleation temperature which has a direct effect on the number of potential nuclei found in the melt. This produces fewer barriers for nucleation with increasing nucleation temperature and generates a number of isolated eutectic Si particles, which results in the flake morphology that was observed for A390 alloy. Fig. 13 compares the liquid fraction vs. temperature curves for the three alloys according to Scheil solidification (no diffusion) using the FACTSAGE software (Bale et al., 2002). The starting point of the eutectic reaction, which occurs after the solidification of pri-
Fig. 13. The liquid fraction vs. temperature curves of A390, 6% Mg and 10 wt% Mg alloys calculated according to the thermodynamic predictions using the FACTSAGE software for Scheil (no diffusion) solidification.
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
mary phase, is indicated by arrows on the curves. This point is called “knee” point. For A390, the knee point represents the start of binary reaction at 566.2 ◦ C (liquid → Al + Si), whereas for 6% and 10% Mg alloys, the knee point represents the start of ternary reaction at 549.7 ◦ C and at 549.2 ◦ C, respectively, according to the following reaction: liquid → Al + Si + Mg2 Si The calculations show that the knee point temperature decreases significantly for the 6% Mg alloy when compared to the value for A390. This decrease is almost constant up to 10% Mg. The FACTSAGE calculation for Scheil solidification of A390 alloy also shows that 68% of matrix microstructure is solid at 549 ◦ C, whereas for 6% and 10% Mg alloys, this is the starting point for the solidification of eutectic matrix. Table 2 shows the value of eutectic reaction temperature calculated by thermodynamic prediction compared to the values obtained from the cooling curve tests for each of the three alloys. This table also indicates that the values of the eutectic temperatures are very similar for both LCR and HCR conditions. However, for both cases, the value of eutectic reaction temperature decreases significantly with the addition of 6% Mg from 566.2 ◦ C to 549.7 ◦ C. Increasing the Mg content to 10% decreases the knee point temperature only slightly below the value for 6% Mg, to 549.2 ◦ C. For conventionally cast samples, the microstructural evolution caused by the addition of Mg, particularly of the eutectic silicon in the matrix (Fig. 8), should contribute to the improvement of wear behaviour of Al–Si–Cu–Mg alloys. This supposition needs more comprehensive tests because the presence of a finer structure of eutectic phase generally results in higher wear resistance of these alloys. 5. Conclusions • For conventionally cast samples, large and individual flake eutectic silicon phase particles were observed in A390 alloy for both high cooling rate and low cooling rate conditions. The microstructure of the high cooling rate samples is finer than for low cooling rate samples. • With the addition of Mg, conventionally cast samples have shown that the eutectic silicon particles form a fine, skeleton network with Chinese script morphology, similar to the eutectic Mg2 Si. This is caused by the decrease of the eutectic formation temperature resulting from the addition of Mg. • For the rheocast samples (rotation speed = 260 rpm, ˙ ave = 52 s−1 ), significant segregation was observed in the three samples due to the shear forces imposed during the solidification. The microstructure is different for two distinct regions. Almost all primary phases are concentrated in the region near the crucible wall (outer layer) whereas the second region (inner layer) next to the agitator shows no primary phases. This attributed to capillary forces between graphite crucible and the particles which are greater than centrifugal forces. Therefore, in terms of wear resistance, the outer layer of samples becomes harder than the inner layer (near to the agitator). • Agglomerated ␣-Al grains with a globular morphology are present in the eutectic phase matrix of the rheocast samples.
775
An increase of the size of the silicon particles was observed for sheared samples confirming that shear promotes the coarsening of silicon particles. However, this shear force is also able to fragment the dendrite morphology of Mg2 Si primary phase crystals and allows them to form of polygonal shapes. • Fragmentation of the eutectic silicon was also observed for the sheared A390 sample when compared to the conventionally solidified sample. With increasing Mg content, the skeleton network of eutectic silicon was also completely destroyed and fragmented. Acknowledgements The authors gratefully acknowledge the financial support from the Fonds Quebecois de Recherche sur la Nature et les Technologies (FQRNT) and Natural Sciences and Engineering Research Council (NSERC) of Canada. References Baiqing, X., Yongan, Z., Qiang, W., Likai, S., Changan, X., Chengjia, S., Xinlai, H., 2003. The study of primary Si phase in spray forming hypereutectic Al–Si alloy. J. Mater. Process. Technol. 137, 183–186. Bale, C.W., Chartrand, P., Degterov, S.A., 2002. FactSage thermochemical software and databases. Calphad 26, 189–228. Elmadagli, M., Alpas, A.T., 2006. Progression of wear in the mild wear regime of an Al–18.5% Si (A390) alloy. Wear 261, 367–381. Jiang, Q.C., Wang, H.Y., Wang, Y., Ma, B.X., Wang, J.G., 2005. Modification of Mg2 Si in Mg–Si alloys with yttrium. Mater. Sci. Eng. A 392, 130–135. Kapranos, P., Kirkwood, D.H., Atkinson, H.V., Rheinlander, J.T., Bentzen, J.J., Toft, P.T., Debel, C.P., Laslaz, G., Maenner, L., Blais, S., Rodriguez-Ibabe, J.M., Lasa, L., Giordano, P., Chiarmetta, G., Giese, A., 2003. Thixoforming of an automotive part in A390 hypereutectic Al–Si alloy. J. Mater. Process. Technol. 135, 271– 277. Lee, J.I., Lee, H.I., Kim, M.I., 1995. Formation of spherical primary silicon crystals during semi-solid processing of hypereutectic Al–15.5 wt%Si alloy. Scripta Metall. Mater. 32, 1945–1949. Li, J., Elmadagli, M., Gertsman, V.Y., Lo, J., Alpas, A.T., 2006. FIB and TEM characterization of subsurfaces of an Al–Si alloy (A390) subjected to sliding wear. Mater. Sci. Eng. A 421, 317–327. Nafisi, S., Emadi, D., Shehata, M.T., Ghomashchi, R., 2006. Effects of electromagnetic stirring and superheat on the microstructural characteristics of Al–Si–Fe alloy. Mater. Sci. Eng. A 432, 71–83. Nafisi, S., Ghomashchi, R., 2006. Effects of modification during conventional and semi-solid metal processing of A356 Al–Si alloy. Mater. Sci. Eng. A 415, 273–285. Nafisi, S., Ghomashchi, R., 2005. Semi-solid metal processing routes: an overview. Can. Metall. Quart. 44, 289–304. Qin, Q.D., Zhao, Y.G., Zhou, W., Cong, P.J., 2007. Effect of phosphorus on microstructure and growth manner of primary Mg2 Si crystal in Mg2 Si/Al composite. Mater. Sci. Eng. A 447, 186–191. Sreeja Kumari, S.S., Pillai, R.M., Rajan, T.P.D., Pai, B.C., 2007. Effects of individual and combined additions of Be, Mn, Ca and Sr on the solidification behaviour, structure and mechanical properties of Al–7Si–0.3Mg–0.8Fe alloy. Mater. Sci. Eng. A 460–461, 561–573. Sundarrajan, A., Mortensen, A., Kattamis, T.Z., Flemings, M.C., 1998. Kinetic undercooling in solidification of a hypereutectic Al–Si alloy: effect of solidifying within a ceramic perform composite. Acta Mater. 46, 91–99. Taghiabadi, R., Ghasemi, H.M., Shabestari, S.G., 2008. Effect of iron-rich intermetallics on the sliding wear behavior of Al–Si alloys. Mater. Sci. Eng. A 490, 162–170. Wang, W., Ajersch, F., Löfvander, J.P.A., 1994. Si phase nucleation on SiC particulate reinforcement in hypereutectic Al–Si alloy matrix. Mater. Sci. Eng. A 187, 65–75. Zhang, J., Fan, Z., Wang, Y., Zhou, B.L., 2000a. Hypereutectic aluminium alloy tubes with graded distribution of Mg2 Si particles prepared by centrifugal casting. Mater. Des. 21, 149–153. Zhang, J., Fan, Z., Wang, Y.Q., Zhou, B.L., 2000b. Microstructural development of Al–15 wt%Mg2 Si in situ composite with mischmetal addition. Mater. Sci. Eng. A 281, 104–122.
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Effect of conventional and rheocasting processes on microstructural characteristics of hypereutectic Al–Si–Cu–Mg alloy with variable Mg content Alireza Hekmat-Ardakan, Frank Ajersch ∗ École Polytechnique de Montréal, Dép. de Génie Chimique, P.O. Box 6079, Centre-ville, Montreal, Quebec, Canada H3C 3A7
a r t i c l e
i n f o
Article history: Received 30 August 2009 Received in revised form 4 January 2010 Accepted 9 January 2010
Keywords: Aluminium alloys Rheocasting Eutectic solidification Magnesium Casting
a b s t r a c t Hypereutectic A390 alloy (Al–17%Si–4.5%Cu–0.5%Mg (wt%)) and similar alloys with the addition of 6 wt% and 10 wt% Mg were investigated using conventional and rheocast processes. Conventional castings were produced with two different cooling rates: −0.15 ± 0.05 ◦ C s−1 and −1.0 ± 0.2 ◦ C s−1 . The rheocast samples were subjected to a rotation speed of 260 rpm during solidification with a cooling rate of −0.15 ± 0.05 ◦ C s−1 . The microstructure of the conventional casting for the high cooling rate samples was found to be finer than for the low cooling rate samples. With high Mg content, the primary silicon transforms to primary Mg2 Si and the nature of the eutectic reaction is also changed from a binary (Al + Si) to a ternary (Al + Si + Mg2 Si) reaction. The morphology of eutectic silicon distributed in matrix microstructure significantly changes from large individual flakes to a fine, skeleton network with Chinese script morphology, similar to the eutectic Mg2 Si. This is caused by the decrease of the eutectic formation temperature resulting from the addition of Mg. The matrix microstructure of the rheocast samples was found to contain fragmented eutectic silicon together with globular alpha aluminium phase particles. The morphology of eutectic silicon in rheocast samples becomes finer with increasing Mg content. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Semisolid metal (SSM) processing can be widely used as a preferred fabrication route for near or net-shaped castings where partially solidified metal with a globular structure (SSM slurry) is injected into a mold, instead of liquid metal casting. SSM processing is generally identified as either rheoprocessing or thixoprocessing. The rheo-route involves preparation of SSM slurry from the liquid phase by stirring the alloy during solidification and transferring directly into a die or mold for component shaping. The thixo-route involves the preparation of a feedstock material with a fine equiaxed or globular structure. The feedstock is then reheated to a temperature between solidus and liquidus (mushy zone) in order to produce a semisolid material which is injected into the mold. A detailed review of SSM processing was described by Nafisi and Ghomashchi (2005). Alloys currently used for SSM processing are mainly conventional Al–Si casting alloys. Hypereutectic Al–Si alloys with more than 11.7 wt% Si offer the possibility of producing in situ metal matrix composites (MMCs) with excellent wear and corrosion resistance where the silicon acts as the reinforcing phase, as indi-
∗ Corresponding author. Tel.: +1 514 340 4711x4533; fax: +1 514 340 4468. E-mail address: [email protected] (F. Ajersch). 0924-0136/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2010.01.005
cated by Kapranos et al. (2003). These composites are a class of materials with an excellent combination of mechanical and physical properties for various applications in the automotive industry. Baiqing et al. (2003) indicated that the high strength and wear resistance of these alloys are attributed to the presence of both hard primary and eutectic Si particles. Zhang et al. (2000, Material Science Engineering A) have also shown that hypereutectic Al–Si alloys with high Mg content can also be considered as in situ Al–Mg2 Si MMCs with a high potential for wear resistance because of the excellent physical and mechanical properties of Mg2 Si intermetallic phase that is formed on solidification. The Mg2 Si phase has a high melting point (1085 ◦ C), very low density (1.99 × 103 kg/m−3 ), high hardness (4.5 × 109 Nm−2 ), a low thermal expansion coefficient (7.5 × 106 K−1 ) and a reasonably high elastic modulus of 120 GPa (Jiang et al., 2005). Consequently, both Al–Si and Al–Mg2 Si composite alloys can be used as wear resistant materials. In terms of properties and solidification behaviour, many similarities exist between the Mg2 Si and Si phase particles as compared by Zhang et al. (2000, Materials and Design). In the present work, the effects of conventional and rheocast processing on the microstructural characteristics, particularly on the morphology of eutectic silicon, of hypereutectic A390 alloy (Al–17%Si–4.5%Cu–0.5%Mg (wt%)) were investigated and compared to the microstructure of hypereutectic alloys with 6% and
768
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
Table 1 Chemical composition of test alloys and AZ91 Mg alloy (wt%). Test alloy
Al
Si
Cu
Mg
Fe
Mn
Zn
P
Ti
A390 6% Mg 10% Mg
Bal. Bal. Bal.
16.7 16.83 17.14
4.58 4.12 4.32
0.58 6.08 9.74
0.32 0.28 0.25
0.02 0.04 0.05
0.01 0.015 0.02
0.0003 0.0002 0.0002
0.02 0.02 0.019
AZ91
Mg
Al
Mn
Zn
Si
Ni
Cu
Bal.
9.3
0.12
0.62
0.02
0.0006
0.0007
10% Mg. The addition of Mg and the formation of both primary and eutectic Mg2 Si phases can also affect the morphology of eutectic silicon. Briefly, this work investigates the effects of the chemical composition, the casting process and the effect of cooling rate on the microstructure of alloy. 2. Experimental procedures The 6 wt% and 10 wt% Mg alloys were produced by adding AZ91 (Mg-base) alloy to A390 alloy. The chemical analysis of the test alloys and AZ91 Mg alloy is presented in Table 1. In order to account for the oxidation loss, an additional amount of 20% Mg (as AZ91) was added to the melt. The AZ91 alloy was wrapped in aluminium foil and added to the liquid A390 alloy at 750 ◦ C. Silicon and copper were also added to the melt in order to attain the chemical composition of these elements at the same value as for the A390 composition. The melt was degassed with argon and stirred for 2 min before pouring into a steel mould to form three ˚35 mm × 120 mm cylindrical samples. 2.1. Conventional casting After preparing the samples, specimens of 60 g portions were transferred into a graphite crucible with 28 mm inner diameter and 5 mm wall thickness and heated to 700 ◦ C using an electric resistance furnace, degassed with argon and stirred for 2 min before sampling. Samples were prepared with two cooling rates: high cooling rate (HCR) and low cooling rate (LCR). For the HCR samples,
Fe 0.0046
the graphite crucible and contents were transferred to the cooling station where a K-type thermocouple was quickly immersed into the melt at a position about 10 mm from the bottom at the center of the crucible. Temperatures were recorded using a data logging system at 15 per second intervals. For the HCR cooling tests, the cooling rates averaged −1.0 ± 0.2 ◦ C s−1 . For the low cooling rate (LCR) tests, the samples were cooled in the furnace at a controlled cooling rate of −0.15 ± 0.05 ◦ C s−1 . Metallographic specimens for the conventionally cast samples were prepared by sectioning transversely at a level 10 mm from the bottom of crucible.
2.2. Rheocasting Fig. 1a shows the schematic diagram of the apparatus used for rheocasting process. A spirally grooved cylindrical agitator was used for producing the rheocast samples. Each rheocast test consisted of 150 g of alloy contained in a graphite crucible with an inner diameter of 41 mm, 10 mm wall thickness and a height of 120 mm. Subsequently, the samples were heated to 700 ◦ C and then held at this isothermal temperature for a period of 5 min. After this period, the alloy was continuously cooled in the furnace at a rate of −0.15 ± 0.05 ◦ C s−1 and sheared at an average rate of ˙ ave = 52 s−1 (rotation speed = 260 rpm) as shown in Fig. 1a. Two K-type thermocouples located in the crucible wall were used to measure the temperature near the surface and near the bottom of the melt. Argon gas was introduced into the furnace chamber in order to prevent sample oxidation. When the upper limit of the viscometer torque was reached, the tests were stopped.
Fig. 1. (a) Schematic diagram of the apparatus used for the preparation of rheocast samples and (b) the location of the sampled regions.
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
769
Fig. 2. Cooling curves for A390, 6% Mg and 10% Mg samples for high cooling rate (HCR) measured by a thermocouple at the center of graphite crucible during conventional casting.
The rheocast specimens were prepared by sectioning the samples as shown in Fig. 1b. The specimens were polished with grinding paper up to 1200 followed by micro-polishing, using monocrystal diamond suspension (0.5 m). All samples were etched by using 0.5% HF solution agent. Optical and scanning electron microscopy (SEM) with EDX (energy dispersive X-ray) analysis were carried out to characterize the microstructure. The Clemex software was used to analyse the particle sizes observed in the microstructure. 3. Results 3.1. Conventional casting process Fig. 2 shows the cooling curves for A390, 6% Mg and 10% Mg alloy samples for the air cooled (HCR) condition. The temperature was measured by a thermocouple at the center of the melt during conventional casting. Fig. 3 compares the detail of the eutectic segment of the cooling curves for HCR and LCR cooling of conventionally cast A390, 6% Mg and 10% Mg alloys. The figure shows that for additions up to 6% Mg, the eutectic reaction temperature is reduced significantly and becomes almost constant for additions between 6% and 10% Mg. Fig. 3 also indicates that the values of the eutectic temperature are very similar for both LCR and HCR conditions. The values of the eutectic reaction temperature obtained experimentally by cooling curves for both cooling conditions are presented in (Table 2). Fig. 4 shows the BSE (back scattered electron) image of A390 alloy for the LCR conventional casting sample showing the form of the intermetallic compounds. The microstructure consists of primary polygonal silicon phase and the eutectic matrix consists mainly of binary Al + Si together with the intermetallic phases, CuAl2 (), Cu2 Mg8 Si6 Al5 (Q) and Al5 FeSi (). Other studies have also observed these microstructures with the presence of the Q phase associated with the phase (Li et al., 2006; Elmadagli and Alpas, 2006). However, the main intermetallic phase in these alloys is the phase. It should be noted that tested alloys also contain small amounts of Fe, Zn, etc., which also react to form of intermetallics. The Table 2 Comparison of the eutectic reaction temperature determined by thermodynamic prediction using FactSage and the measurement of cooling curves at LCR (−0.15 ± 0.05 ◦ C s−1 ), HCR (−1.0 ± 0.2 ◦ C s−1 ) for three alloys.
A390 6% Mg 10% Mg
Predicted (◦ C)
Measured (◦ C) LCR
Measured (◦ C) HCR
566.2 549.7 549.2
563 ± 1 545 ± 1 543 ± 1
564 ± 1 544 ± 1 542 ± 1
Fig. 3. The eutectic segment of the cooling curves at HCR and LCR for (a) A390, (b) 6% Mg and (c) 10% Mg alloys during conventional casting.
platelet Fe-containing phase, -Al5 FeSi, is the most common Fe intermetallic phase and contributes to the decrease of the mechanical properties of Al–Si cast alloys as reported by Sreeja Kumari et al. (2007), Taghiabadi et al. (2008), and Nafisi et al. (2006). With increasing Mg content, the Mg2 Si intermetallic phase first appears in the matrix in the form of a eutectic morphology and then as a primary phase. The microstructure of these three alloys for LCR and HCR conditions are compared in Fig. 5. For both cooling conditions, the morphology of grey primary silicon particles remained polygonal, whereas Mg2 Si crystals (dark particles) are either in the form of dendritic crystals with entrapped liquid as shown by Jiang et al.
770
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
Fig. 4. BSE (back scattered electron) image of A390 alloy for conventionally cast low cooling rate (LCR) sample. (a) Large faceted primary silicon (1). The matrix consists of binary Al + Si eutectic as well as the additional phases found in location A. (b) Magnified zone of location A, showing -CuAl2 (2), Al (3), Cu2 Mg8 Si6 Al5 (Q phase) (4) and -Al5 FeSi (5).
Fig. 5. Overall microstructure of A390 alloy (a and b), 6% Mg alloy (c and d) and 10% Mg alloy (e and f) for different cooling rates for conventionally cast samples. Images a, c and e correspond to LCR samples and b, d and f the HCR samples. The light grey phase is silicon, the dark phase is Mg2 Si and the white phase is ␣-Al.
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
(2005) and Qin et al. (2007) or in the form of compact polygonal shapes. The microstructure of the high cooling rate samples (Fig. 5b, d, and f) is finer than that of the low cooling rate samples (Fig. 5a, c, and e). Fig. 6 clearly shows that the eutectic morphology for both the HCR and LCR samples is significantly changed with the addition of Mg, particularly at 6% Mg when compared to A390 alloy. The eutectic silicon together with the dendritic ␣-Al shows a coarse flake structure for A390 alloy cooled at both low and high rates (Fig. 6a and b). However, the eutectic microstructure for 6% Mg alloys (Fig. 6c and d) and 10% Mg alloys (Fig. 6e and f) is similar and the eutectic silicon transforms into a fine, skeleton network with Chinese script morphology, similar to the form of eutectic Mg2 Si, as was also observed by Jiang et al. (2005). The eutectic matrices of A390 and of the 6% Mg alloy at the HCR condition are compared in Fig. 7 using SEM images. The morphology of the eutectic silicon for A390 and 6% Mg alloys for the HCR condition (shown in Fig. 7) is also compared in the SEM image of Fig. 8 where the surface was deep etched using 0.5% HF solution. These figures clearly show that the eutectic silicon in the 6% Mg alloy is much finer than in the A390 alloy.
771
3.2. Rheocasting process As described in the experimental procedure, the cooling condition for the rheocast samples is similar to the LCR condition since the samples are stirred during continuous cooling solidification. The tests were stopped at the upper limit of the viscometer torque. This maximum torque was attained at different temperatures for each of the alloys. For A390 alloy this occurred at 541 ◦ C (fliquid = 25.1%) and for the 6% and 10% Mg alloys the temperatures were found to be 533 ◦ C (fliquid = 26.3%) and 530 ◦ C (fliquid = 23.3%), respectively. Therefore, the solidified structure of the remaining liquid should be similar to the conventional cooling structure without shearing effect at a solid fraction of about 25%. A significant segregation of primary phases was detected in the microstructure of the rheocast samples due to the shear force imposed on the slurry during solidification. As a result, the microstructure is different for two distinct regions. The region with a radial distance between 2 mm and 3 mm from the crucible wall (near crucible zone shown in Fig. 1b) is a region where the most of the primary phase particles are concentrated, as shown in left side images of Fig. 9 for A390, and the 6% and 10% Mg alloys. Wang
Fig. 6. The eutectic microstructure of A390 alloy (a and b), 6% Mg alloy (c and d) and 10% Mg alloy (e and f) alloys for different cooling rates for conventionally cast samples. Images a, c and e correspond to LCR samples and b, d and f the HCR samples. Light grey phase is silicon, dark phase is Mg2 Si and the white phase is ␣-Al grains.
772
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
Fig. 7. Comparison of eutectic silicon phase (arrows) and microstructures of conventionally solidified alloys at HCR condition for (a) A390 and (b) 6% Mg, at the same magnification.
et al. (1994) have shown that during the solidification of hypereutectic Al–Si/SiC composite, the SiC particles in the composite melt serve as the substrates for primary Si phase nucleation and all the primary Si particles form heterogeneously on the SiC surface. In the melt-stirred condition, all the Si particles become agglomerated and are attached by the SiC particles. Sundarrajan et al. (1998) have also explained this phenomenon based on the large undercooling required for the growth of polygonal Si crystals along low-energy, atomically smooth and slow-growing (1 1 1) planes. It was shown that ceramic composites such as aluminium oxide, silicon carbide and graphite can eliminate such large undercooling and result in preferential nucleation and growth of Si on these materials. In this case undercooling is eliminated as a result of capillary (interfacial) forces which bend the liquid–solid interface near the contact line of (1 1 1) facets with the composite surface. This phenomenon can explain the segregation of Si and Mg2 Si particles resulting for interfacial forces which are greater than the centrifugal forces even though the density of the liquid phase is higher than that of the particles. The region next to the agitator (near agitator zone shown in Fig. 1b) with a radial distance between 5 mm and 6 mm from agitator, shows no primary phases. The globular ␣-Al grains observed in this region form aggregates as shown in right side images of Fig. 9. When the primary phases in the region near the crucible wall are compared with non-sheared LCR samples in Fig. 5, it is clear that the segregation of primary phases caused by the shear forces is much higher in the rheocast (sheared) samples. The primary silicon of sheared A390 alloy (Fig. 9a) also agglomerates. Fig. 10 indicates that all three sheared samples show an increase of the size of the silicon particles when compared to the non-sheared samples as measured by image analysis. This implies that the shear promotes the coarsening of silicon particles. Lee et al. (1995) have also
reported that significant fragmentation and changes of primary Si morphology occur for an Al–15.5%Si alloy during the isothermal stirring but little change was observed when a sample was continuously cooled at a shear rate of ˙ = 200 s−1 and at a cooling rate of −0.03 ◦ C s−1 . This shear rate however is able to fragment the dendritic morphology of the Mg2 Si primary phase crystals with entrapped liquid as shown in Fig. 5c and e, forming smaller, compact polygonal shapes as shown in Fig. 9c and e. Fig. 11 compares the average particle size of the primary Mg2 Si particles for the 6% and 10% Mg alloys for conventionally and rheocast samples. The reduction in particle size is primarily due to the shear forces resulting in the fragmentation of the larger eutectic structure of the conventionally solidified samples. In the region near the agitator, the agglomerated ␣-Al grains are present only in the matrix of the eutectic phase as shown in Fig. 12a, c and e for each of these alloys. Fig. 12b, d and f represents the regions between the agglomerations consisting mostly of the eutectic region. For the A390 alloys (Fig. 12a and b), there is no significant difference in the microstructure of ␣-Al grains in matrix. However, for the high Mg-containing alloys of 6% and 10% Mg, respectively shown in Fig. 12c and d, and e and f, the microstructural difference is more pronounced. The ␣-Al grains are generally surrounded by intermetallic phases, mainly by the -phase. It can be clearly observed that the eutectic silicon for the sheared A390 sample (Fig. 12b) becomes fragmented when compared to the non-sheared sample shown in Fig. 6a. For the high Mg-containing alloys, the skeleton network of eutectic silicon observed for the LCR condition, is completely destroyed and fragmented, as shown in Fig. 12d (6% Mg) and Fig. 12f (10% Mg). For the rheocast samples, the morphology of eutectic silicon becomes finer with increasing the Mg content.
Fig. 8. Deep etched images showing the morphology of the eutectic Si in (a) A390 alloy and (b) 6% Mg alloy for conventionally HCR solidified cast samples at the same magnification.
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
773
Fig. 9. The rheocast microstructure of A390 alloy (a and b), 6% Mg alloy (c and d) and 10% Mg alloy (e and f). Images a, c and e represent the area near the crucible wall and images b, d and f represent the area near the agitator.
Fig. 10. Average particle size of primary silicon for conventional and rheocasting processes for three alloys.
Fig. 11. Average particle size of primary Mg2 Si for conventional and rheocasting processes for alloys with 6% and 10% Mg alloys.
774
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
Fig. 12. The microstructures near to the agitator for stir cast samples. A390 alloy (a and b), 6% Mg alloy (c and d) and 10% Mg alloy (e and f). Images a, c and e show agglomeration of ␣-Al grains and images b, d and f illustrate the eutectic microstructure between the agglomeration of ␣-Al grains.
4. Discussion The results of this study have shown that Mg additions to hypereutectic Al–Si alloy have a significant effect on the solidification structure of conventionally cast and rheocast alloys. As shown in Fig. 7, the change in morphology of eutectic silicon is due to the decrease of the eutectic temperature with the addition of Mg. Nafisi and Ghomashchi (2006) also investigated the effect of the reduction of eutectic temperature on the modification of the eutectic silicon morphology of A356 Al–Si alloy when Sr is added as modifier. In fact, they attributed this temperature reduction to the eutectic nucleation temperature which has a direct effect on the number of potential nuclei found in the melt. This produces fewer barriers for nucleation with increasing nucleation temperature and generates a number of isolated eutectic Si particles, which results in the flake morphology that was observed for A390 alloy. Fig. 13 compares the liquid fraction vs. temperature curves for the three alloys according to Scheil solidification (no diffusion) using the FACTSAGE software (Bale et al., 2002). The starting point of the eutectic reaction, which occurs after the solidification of pri-
Fig. 13. The liquid fraction vs. temperature curves of A390, 6% Mg and 10 wt% Mg alloys calculated according to the thermodynamic predictions using the FACTSAGE software for Scheil (no diffusion) solidification.
A. Hekmat-Ardakan, F. Ajersch / Journal of Materials Processing Technology 210 (2010) 767–775
mary phase, is indicated by arrows on the curves. This point is called “knee” point. For A390, the knee point represents the start of binary reaction at 566.2 ◦ C (liquid → Al + Si), whereas for 6% and 10% Mg alloys, the knee point represents the start of ternary reaction at 549.7 ◦ C and at 549.2 ◦ C, respectively, according to the following reaction: liquid → Al + Si + Mg2 Si The calculations show that the knee point temperature decreases significantly for the 6% Mg alloy when compared to the value for A390. This decrease is almost constant up to 10% Mg. The FACTSAGE calculation for Scheil solidification of A390 alloy also shows that 68% of matrix microstructure is solid at 549 ◦ C, whereas for 6% and 10% Mg alloys, this is the starting point for the solidification of eutectic matrix. Table 2 shows the value of eutectic reaction temperature calculated by thermodynamic prediction compared to the values obtained from the cooling curve tests for each of the three alloys. This table also indicates that the values of the eutectic temperatures are very similar for both LCR and HCR conditions. However, for both cases, the value of eutectic reaction temperature decreases significantly with the addition of 6% Mg from 566.2 ◦ C to 549.7 ◦ C. Increasing the Mg content to 10% decreases the knee point temperature only slightly below the value for 6% Mg, to 549.2 ◦ C. For conventionally cast samples, the microstructural evolution caused by the addition of Mg, particularly of the eutectic silicon in the matrix (Fig. 8), should contribute to the improvement of wear behaviour of Al–Si–Cu–Mg alloys. This supposition needs more comprehensive tests because the presence of a finer structure of eutectic phase generally results in higher wear resistance of these alloys. 5. Conclusions • For conventionally cast samples, large and individual flake eutectic silicon phase particles were observed in A390 alloy for both high cooling rate and low cooling rate conditions. The microstructure of the high cooling rate samples is finer than for low cooling rate samples. • With the addition of Mg, conventionally cast samples have shown that the eutectic silicon particles form a fine, skeleton network with Chinese script morphology, similar to the eutectic Mg2 Si. This is caused by the decrease of the eutectic formation temperature resulting from the addition of Mg. • For the rheocast samples (rotation speed = 260 rpm, ˙ ave = 52 s−1 ), significant segregation was observed in the three samples due to the shear forces imposed during the solidification. The microstructure is different for two distinct regions. Almost all primary phases are concentrated in the region near the crucible wall (outer layer) whereas the second region (inner layer) next to the agitator shows no primary phases. This attributed to capillary forces between graphite crucible and the particles which are greater than centrifugal forces. Therefore, in terms of wear resistance, the outer layer of samples becomes harder than the inner layer (near to the agitator). • Agglomerated ␣-Al grains with a globular morphology are present in the eutectic phase matrix of the rheocast samples.
775
An increase of the size of the silicon particles was observed for sheared samples confirming that shear promotes the coarsening of silicon particles. However, this shear force is also able to fragment the dendrite morphology of Mg2 Si primary phase crystals and allows them to form of polygonal shapes. • Fragmentation of the eutectic silicon was also observed for the sheared A390 sample when compared to the conventionally solidified sample. With increasing Mg content, the skeleton network of eutectic silicon was also completely destroyed and fragmented. Acknowledgements The authors gratefully acknowledge the financial support from the Fonds Quebecois de Recherche sur la Nature et les Technologies (FQRNT) and Natural Sciences and Engineering Research Council (NSERC) of Canada. References Baiqing, X., Yongan, Z., Qiang, W., Likai, S., Changan, X., Chengjia, S., Xinlai, H., 2003. The study of primary Si phase in spray forming hypereutectic Al–Si alloy. J. Mater. Process. Technol. 137, 183–186. Bale, C.W., Chartrand, P., Degterov, S.A., 2002. FactSage thermochemical software and databases. Calphad 26, 189–228. Elmadagli, M., Alpas, A.T., 2006. Progression of wear in the mild wear regime of an Al–18.5% Si (A390) alloy. Wear 261, 367–381. Jiang, Q.C., Wang, H.Y., Wang, Y., Ma, B.X., Wang, J.G., 2005. Modification of Mg2 Si in Mg–Si alloys with yttrium. Mater. Sci. Eng. A 392, 130–135. Kapranos, P., Kirkwood, D.H., Atkinson, H.V., Rheinlander, J.T., Bentzen, J.J., Toft, P.T., Debel, C.P., Laslaz, G., Maenner, L., Blais, S., Rodriguez-Ibabe, J.M., Lasa, L., Giordano, P., Chiarmetta, G., Giese, A., 2003. Thixoforming of an automotive part in A390 hypereutectic Al–Si alloy. J. Mater. Process. Technol. 135, 271– 277. Lee, J.I., Lee, H.I., Kim, M.I., 1995. Formation of spherical primary silicon crystals during semi-solid processing of hypereutectic Al–15.5 wt%Si alloy. Scripta Metall. Mater. 32, 1945–1949. Li, J., Elmadagli, M., Gertsman, V.Y., Lo, J., Alpas, A.T., 2006. FIB and TEM characterization of subsurfaces of an Al–Si alloy (A390) subjected to sliding wear. Mater. Sci. Eng. A 421, 317–327. Nafisi, S., Emadi, D., Shehata, M.T., Ghomashchi, R., 2006. Effects of electromagnetic stirring and superheat on the microstructural characteristics of Al–Si–Fe alloy. Mater. Sci. Eng. A 432, 71–83. Nafisi, S., Ghomashchi, R., 2006. Effects of modification during conventional and semi-solid metal processing of A356 Al–Si alloy. Mater. Sci. Eng. A 415, 273–285. Nafisi, S., Ghomashchi, R., 2005. Semi-solid metal processing routes: an overview. Can. Metall. Quart. 44, 289–304. Qin, Q.D., Zhao, Y.G., Zhou, W., Cong, P.J., 2007. Effect of phosphorus on microstructure and growth manner of primary Mg2 Si crystal in Mg2 Si/Al composite. Mater. Sci. Eng. A 447, 186–191. Sreeja Kumari, S.S., Pillai, R.M., Rajan, T.P.D., Pai, B.C., 2007. Effects of individual and combined additions of Be, Mn, Ca and Sr on the solidification behaviour, structure and mechanical properties of Al–7Si–0.3Mg–0.8Fe alloy. Mater. Sci. Eng. A 460–461, 561–573. Sundarrajan, A., Mortensen, A., Kattamis, T.Z., Flemings, M.C., 1998. Kinetic undercooling in solidification of a hypereutectic Al–Si alloy: effect of solidifying within a ceramic perform composite. Acta Mater. 46, 91–99. Taghiabadi, R., Ghasemi, H.M., Shabestari, S.G., 2008. Effect of iron-rich intermetallics on the sliding wear behavior of Al–Si alloys. Mater. Sci. Eng. A 490, 162–170. Wang, W., Ajersch, F., Löfvander, J.P.A., 1994. Si phase nucleation on SiC particulate reinforcement in hypereutectic Al–Si alloy matrix. Mater. Sci. Eng. A 187, 65–75. Zhang, J., Fan, Z., Wang, Y., Zhou, B.L., 2000a. Hypereutectic aluminium alloy tubes with graded distribution of Mg2 Si particles prepared by centrifugal casting. Mater. Des. 21, 149–153. Zhang, J., Fan, Z., Wang, Y.Q., Zhou, B.L., 2000b. Microstructural development of Al–15 wt%Mg2 Si in situ composite with mischmetal addition. Mater. Sci. Eng. A 281, 104–122.