A new nucleation mechanism of primary Si by like-peritectic coupling of AlP and Al 4 C 3 in near eutectic Al Si alloy

Journal of Alloys and Compounds 429 (2007) 119 125 A new nucleation mechanism of primary Si by like-peritectic coupling of AlP and Al 4 C 3 in near eutectic Al Si alloy Lina Yu, Xiangfa Liu, Haimin Ding, Xiufang Bian Key Laboratory of Liquid Structure and Heredity of Materials, Ministry of Education, Shandong University, 73 Jingshi Road, Jinan 250061, PR China Received 8 March 2006; received in revised form 1 April 2006; accepted 5 April 2006 Available online 3 July 2006 Abstract The effect of TiC particles on the phosphorous modification efficiency in near eutectic Al Si alloy and a new nucleating mechanism was studied in this paper. The results show that the phosphorous modification efficiency can be significantly improved and the sizes of the primary Si can be obviously refined after the addition of trace TiC particles. EPMA shows there is a coupling of AlP and Al 4 C 3 particles in the center of primary Si, and a new nucleation mechanism of primary Si by like-peritectic coupling of AlP and Al 4 C 3 was presented. The formation mechanism of like-peritectic coupling of the particles was also discussed. 2006 Published by Elsevier B.V. Keywords: Al Si alloy; Primary silicon; Modification; Nucleation mechanism 1. Introduction Considerable effort has been devoted to the development of lightweight engineering materials during the past decades [1 3]. Recently, near-eutectic and hypereutectic Al Si alloys have been widely used in automotive applications, especially in the piston industry, because of their excellent combination of properties, including good abrasion and corrosion resistance, low coefficient of thermal expansion and high strength-to-weight ratio [4,5]. As the desirable combination of characteristics of Al Si near-eutectic and hypereutectic alloys depends on the primary Si grain size to a large extent, the modification of primary Si is being studied more widely with increasing usage. Various methods have been used for the modification of primary Si particles, such as rapid cooling [6], low temperature casting [7] and various alloying additions [8]. Microstructure control using minor element addition has been the most popular method due to its simplicity. Phosphorous has been the most widely used minor element for the modification of primary Si in near-eutectic and hypereutectic Al Si alloys. Al Ti C, as a grain refiner of commercial purity aluminum and binary hypoeu- Corresponding author. Tel.: +86 531 88395414; fax: +86 531 88395414. E-mail address: xfliu@sdu.edu.cn (X.F. Liu). tectic Al Si alloys, can be used to refine -Al dendrites [9]. While some researchers [10,11] have found that the addition of TiC particles significantly improves the phosphorous modification effect, the effect mechanism has not been understood until now. This article discusses the effect mechanism of TiC particles on P modification efficiency in near eutectic Al Si alloys by using EPMA and crystal lattice coherence analysis. This also provides new practical and theoretical information for the neareutectic and hypereutectic Al Si production industry. 2. Experimental procedure The eutectic Al Si alloy used in the experiments was produced by a 20 kw medium frequency induction furnace using commercial purity aluminum (99.85%) and super-purity crystalline Si (99.999%). The TiC particles were added in the form of Al 8Ti 2C master alloy. The chemical compositions of the experimental alloy are shown in Table 1 (all compositions quoted in this work are in wt.% unless otherwise stated). The eutectic Al Si alloy was re-melted in a clay-bonded graphite crucible, heated in an electric resistance-heating furnace at 800 C and held at this temperature for 30 min. Part of the melt was poured into an iron chill mould with dimensions of 70 mm 35 mm 20 mm and Sample-1 was obtained without any additions or further treatment. Sample-2 was obtained 40 min after the addition of 1%Al 3.5P master alloy to the melt. Then, following the addition of 0.2%Al 8Ti 2C master alloy, Sample-3 was obtained after 10 min. More Al 8Ti 2C master alloy was added until there was 2%Al 8Ti 2C master alloy in the melt. After 10 min, the melt was cast and Sample-4 was obtained. 0925-8388/$ see front matter 2006 Published by Elsevier B.V. doi:10.1016/j.jallcom.2006.04.011

120 L. Yu et al. / Journal of Alloys and Compounds 429 (2007) 119 125 Table 1 Chemical compositions of produced eutectic Al Si alloy and Al Ti C master alloy (wt.%) Samples Si Cu Mg Ni Ti C P Sr Al Al 12.6Si 12.61 0.002 0.004 0.002 0.218 0.0046 <0.0001 0.0002 Bal. Al 8Ti 2C <0.15 <0.01 7.98 1.99 Bal. Another four samples were obtained in the same experiment process without the addition of Al 8Ti 2C master alloy for comparison. 1%Al 3.5P master alloy was added to the eutectic Al Si alloy and samples were taken after 40, 50, and 60 min, respectively. All the samples were poured into the same type of iron chill mould, preheated to 150 C before casting. The microstructure analysis was carried out on as-cast samples to investigate the morphologies and transformation of the silicon phase. Metallographic specimens were cut directly from 10 mm above the bottom of the ingots and mechanically ground and polished using standard routines. The structure and qualitative analysis were conducted by using a high scope video microscope (HSVM) and JXA-8840 electron probe micro-analyzer (EPMA). 3. Results and discussion Fig. 1 shows the microstructures of Al 12.6Si alloy before and after the addition of Al 3.5P and Al 8Ti 2C master alloys. There are a few primary Si grains in Al 12.6Si alloy, and their sizes are large block-like and unequal, as shown in Fig. 1(a). After the addition of 1%Al 3.5P master alloy, there is an obvious phosphorous modification effect and the size of the primary Si decreased significantly. The holding time used in this experiment has no influence on the phosphorous modification effect of Al P modified Al 12.6Si alloy, which is consistent with the previous result [12]. The phosphorous modification effect is retained after the addition of 0.2%Al 8Ti 2C master alloy in the melt. The average size of the primary Si decreased from 40 m to approximately 20 m, indicating that the Al 8Ti 2C master alloy improved the phosphorous modification effect. In addition, when 10 times the Al 8Ti 2C master alloy was added to the melt, the positive phosphorous modification effect was retained as before without any poisoning phenomenon. According to the literature [13 15], AlP and Si are both diamond cubic with very similar lattice parameters. Primary Si nucleates heterogeneously on the solid AlP particles with a cubecube orientation relationship and solidifies [13], which promotes the precipitation and refinement of primary silicon. In terms of this theory, Al P master alloy can modify primary silicon effectively, as shown in Fig. 1(b). Fig. 2 presents the EPMA of a primary silicon nucleus in Sample-4 (Al 12.6Si with the addition of both 1%Al 3.5P and 2%Al 8Ti 2C master alloys), and there is a light particle inside the dark nucleus of the primary silicon. The X-ray images show that the dark nucleus contains Al and P elements, indicating that it was AlP compound. While the light particle contains Al and C, without Ti, which indicates that the particle is not TiC but Al 4 C 3 compound. Furthermore, the Al 4 C 3 particle lies in the core of the AlP compound. Fig. 3 shows another nucleus of the primary silicon in Sample-4, and the compositions of the magnified nucleus are illustrated in Fig. 4. It is noted that the nucleus contains Al, P and C elements, without Ti. Furthermore, the P and C are not overlapping each other, while they both overlap the Al, indicating that they are AlP and Al 4 C 3 compounds, respectively. This is similar to the result of Fig. 2, since the Al 4 C 3 compound Fig. 1. Microstructures of Al 12.6Si alloy before and after the addition of Al 3.5P and Al 8Ti 2C master alloys: (a) Al 12.6Si; (b) Al 12.6Si + 1%Al 3.5P; (c) Al 12.6Si + 1%Al 3.5P + 0.2%Al 8Ti 2C; (d) Al 12.6Si + 1%Al 3.5P + 2%Al 8Ti 2C.

L. Yu et al. / Journal of Alloys and Compounds 429 (2007) 119 125 121 Fig. 2. EPMA of a primary silicon nucleus: (a) SEI of the primary Si; (b f) the X-ray images for respective elements, Al, Si, P, C and Ti. Fig. 3. Another nucleus of the primary silicon in the Al 12.6Si added 1%Al 3.5P and 2%Al 8Ti 2C master alloys. is not surrounded with AlP compound but adjacent to AlP compound. In order to further confirm whether that the P and C elements overlap each other or not, the composition along the line A B across the primary silicon in Fig. 5(a) is illustrated in Fig. 5(b). In the nucleus of the primary silicon, the lower Si composition corresponds to an Al peak. The P also exhibits a peak offset to the left of the Al peak, and C peak occurs to the right of the Al peak. It can be concluded that there are two types of compound, AlP and Al 4 C 3, in the nucleus of the primary silicon. In a pure Al system, a maximum in reactivity was observed in the liquid-state, between 700 and 750 C, where TiC decomposes to form Al 4 C 3 and TiAl 3 [16,17]. TiC was found to be stable above 900 C [16]. In this work, the melt was heated to Fig. 4. EPMA of another primary silicon nucleus: (a) SEI of the primary Si; (b f) the X-ray images for respective elements, Al, Si, P, C and Ti.

122 L. Yu et al. / Journal of Alloys and Compounds 429 (2007) 119 125 Fig. 5. Line distribution of chemical composition along the line across the primary silicon: (a) line A B across the primary silicon; (b) chemical composition distribution along line A B. a temperature of 800 C below the stable temperature, so similarities in the thermal stability of TiC in the melt were observed and some Al 4 C 3 compounds were formed through the following reaction: TiC + Al Al 4 C 3 + TiAl 3 (1) When Al P master alloy is added into the melt, many of AlP particles may not be absorbed by the Al melt because of the lower solubility of phosphorus and the density difference. The undissolved AlP do not act as the nuclei of the primary Si crystals during solidification. It is well understood that there are only TiC particles and the Al matrix in the Al 8Ti 2C master

L. Yu et al. / Journal of Alloys and Compounds 429 (2007) 119 125 123 alloy. Thus, the improvement in the phosphorous modification effect is only due to the TiC particles. When TiC particles are added to the melt, the Al 4 C 3 formed can absorb the AlP, coupling together as a like-peritectic reaction to form an AlP-rich transition layer around the Al 4 C 3 particles, which can act as heterogeneous nucleating sites for the primary Si grains during solidification. With the dissolved AlP precipitating to Al 4 C 3 compound, more AlP will dissolve into the Al melt. Therefore, most undissolved P can provide a modification effect after the addition of TiC particles (Al 4 C 3 is unstable in air), thus more AlP particles act as the nuclei of the primary silicon, increasing the quantity of the primary Si and decreasing the size, as shown in Fig. 1(c) and (d). It can be concluded that the addition of TiC particles can significantly increase the absorptivity of P in the melt, greatly improving the modification effect of the Al P master alloy. Heterogeneous nucleation can be regarded as a geometrically modified case of homogeneous nucleation by which the activation barrier is decreased by the presence of a foreign substrate [18]. A purely geometrical calculation shows that when the solid/liquid interface of the substance is partly replaced by an area of low energy solid/solid interface between the crystal and a foreign solid, nucleation can be greatly facilitated. Normally, the potential of the foreign substrate in facilitating the nucleation process is estimated from the following relation [19]: (2 + cos θ)(1 cos θ) f (θ) = (2) 4 where θ is the wetting angle between the growing crystal and the foreign substrate within the melt. Under conditions of good solid/solid wetting, i.e. small θ, the foreign substrate can have a dramatic effect on the nucleation process. Good solid/solid wetting is expected when there are planes of low lattice disregistry between the foreign substrate and the growing crystal. Therefore, the need for near-perfect epitaxy of AlP on at least one of the crystal faces of the Al 4 C 3 substrate has been suggested as being essential for effective nucleation. It is well known that AlP is a diamond cubic with lattice parameter: a = 5.42 [13], as shown in Fig. 6. The Al atoms, shown in black, form a normal face-centered cubic, while the P atoms, shown in white, are distributed in the tetrahedral interstices of the crystal cell, with each P atom surrounded by four proximate Al atoms. Al 4 C 3 is hexagonal with the lattice parameters: a = 3.335, c = 24.96, with the Al atoms located on the corner angles of a hexagon and in the centre of the back surface. Table 2 Possible coherent interface of Al 4 C 3 and AlP crystals Number Al 4 C 3 AlP δ (%) d (Å) (hkl) d (Å) (hkl) 1 1.8888 1 0 10 1.9162 2 2 0 1.43 < 5 2 1.6675 1 1 0 1.6341 3 1 1 2.04 < 5 3 1.6644 0 0 15 1.6341 3 1 1 1.85 < 5 4 1.3872 0 0 18 1.3550 4 0 0 2.38 < 5 5 1.3385 0 2 7 1.3550 4 0 0 1.22 < 5 6 1.3011 1 1 12 1.3550 4 0 0 3.98 < 5 7 1.2011 1 1 12 1.2434 3 3 1 4.64 < 5 Note: d, interplanar distance; (hkl), indices of crystallographic plane; δ, the degree of disregistry for interplanar distance. When there is a good coherent relationship existing on the interface of two types of phases, one phase can act as very fine heterogeneous nucleating site for the other phase [20]. The interatomic distance of the crystal face of the two phases should be close to each other, besides the atomic arrangement of the crystal faces should be similar. When the interface of two phases has a good coherent relationship, the interplanar distances should be also similar to each other. Based on the interplanar distances of Al 4 C 3 and AlP crystals, several possible coherent interfaces between them can be obtained, as shown in Table 2, and the degree of disregistry δ is below 5%. Generally, the degree of disregistry between the substrate phase and the crystalline phase is measured from the following Turnbull Vonnegut equation: δ = a s a c a c 100% (3) where a s and a c is the interatomic/interplanar distance of substrate plane and crystalline plane without deformation, respectively. Because of the inherent limitation of the Turnbull Vonnegut equation, and in order to broaden the equation so that it would be applicable to crystallographic combinations of two phases with planes of differing atomic arrangements, it was found necessary to modify the equation in terms of the angular difference between the crystallographic directions within the planes [21]. The modified equation is presented in two forms, the first of which is the generalized form: δ = δ 1 + δ 2 + δ 3 3 100% (4) where δ 1, δ 2 and δ 3 are the disregistries calculated along the three lowest-index directions within a 90 quadrant of the planes of the nucleated solid and the substrate. Table 3 Parameters for the disregistry equation [5] Case d [uvw]s d [uvw]n θ ( ) d [uvw]s cos θ Fig. 6. Crystal structure of AlP. (1 1 0)Al 4 C 3 (3 1 1)AlP 3.335 3.83 0 3.335 11.55 11.50 6.5 11.48 6.67 6.63 13 6.50

124 L. Yu et al. / Journal of Alloys and Compounds 429 (2007) 119 125 Fig. 7. The crystallographic relationship at the interface between the (1 1 0) of Al 4 C 3 and the (3 1 1) of AlP. The second, more specific form of the modified equation, is as follows: δ (hkl)s (hkl)n = 3 i=1 ( (d [uvw] i s cos θ) d [uvw] i n )/d [uvw] i n 3 100% where (hkl) s is the a low-index plane of the substrate, [uvw] s the a low-index direction in (hkl) s,(hkl) n the a low-index plane in the nucleated solid, [uvw] n the a low-index direction in (hkl) n, d [uvw]n the the interatomic spacing along [uvw] n, d [uvw]s the interatomic spacing along [uvw] s, and θ is the angle between the [uvw] s and [uvw] n. From Table 2, the Number 2 entry can be taken as an example to examine the atomic arrangement of the crystal plane. Fig. 7 illustrates the typical planar atomic arrangement in the (1 1 0) crystal face of Al 4 C 3 and in the (3 1 1) crystal face of AlP, in which the broken-line circles represent the Al atoms in the (1 1 0) crystal face of Al 4 C 3 while the shaded circles represent the P atoms in the (3 1 1) crystal face of AlP. The corresponding parameters for Eq. (5) are listed in Table 3. Therefore, (5) The nucleation process of primary silicon with and without the like-peritectic coupling of nucleating sites can be described as follows: in the case of the addition of AlP particles only, many of AlP particles may not be absorbed by the Al melt because of the lower solubility of phosphorus and the density difference. The undissolved AlP which cannot act as the nuclei of the primary Si crystals during solidification become the dregs and float to the melt surface. With a decrease in temperature, the solubility of the AlP in the melt decreases, so the dissolved AlP segregates or grow up. When the temperature is low enough, primary silicon nucleates on the AlP nucleating sites. Because the absorptivity of the AlP particles is low, few primary silicon crystals are precipitated and their sizes are coarser. In comparison, with the addition of both AlP and TiC particles, the nucleation process of the primary silicon with the like-peritectic coupling of nucleating sites can be schematically illustrated in Fig. 8. There will be quantities of Al 4 C 3 particles (white particles in Fig. 8(a)) formed in the melt together with a small amount of AlP particles (black particles). With a decrease in temperature, AlP can easily segregate on the Al 4 C 3 particles and have a like-peritectic coupling with them to form a coupling compound of Al 4 C 3 AlP, which is an AlP-rich transition layer formed around the Al 4 C 3 particles. The coupling function can be shown as follows: Al 4 C 3 (s) + AlP(l) Al 4 C 3 AlP(s) (6) The coupling compounds of Al 4 C 3 AlP can act as heterogeneous nucleating sites of the primary Si during solidification. With the dissolved AlP precipitating to the Al 4 C 3 compound, more AlP will dissolve into the Al melt. Therefore, most undissolved P can provide a modification effect. Furthermore, because the sizes of Al 4 C 3 particles are submicron and the number of them is large, the coupling compounds are abundant and they can distribute evenly in melt, as shown in Fig. 8(b). The large number of evenly distributed nuclei is available for the primary silicon, so the quantity of the primary Si increases significantly, and accordingly the size decreases, as presented in Fig. 8(c). This is a proposed mechanism for the improvement of AlP δ (1 1 0)Al 4C 3 [( 3.335 3.83 )/3.83] + [( 11.48 11.50 )/11.50] + [( 6.50 6.63 )/6.63] (3 1 1)AlP = 100% 5% 3 This indicates that Al 4 C 3 compound is a good nucleation substrate for AlP compound, and AlP can easily nucleate. modification efficiency in the Al Si alloy after the addition of a small amount of TiC particles. Fig. 8. Schematic illustration of primary silicon nucleation process with the like-peritectic coupling of nucleating sites: (a) quantities of Al 4 C 3 particles formed in the melt; (b) like-peritectic coupling of Al 4 C 3 and AlP particles; (c) nucleation of primary silicon.

L. Yu et al. / Journal of Alloys and Compounds 429 (2007) 119 125 125 4. Conclusions The effect of TiC particles on the phosphorous modification efficiency in near eutectic Al Si alloy was studied by using HSVM and EPMA, and the mechanism was also discussed. The phosphorous modification effect can be significantly improved and the size of the primary Si can be refined from 40 m to approximately 20 m after the addition of only 0.2%Al 8Ti 2C into the Al 12.6Si alloy. EPMA shows there is an AlP-rich transition layer around or near to Al 4 C 3 compound, which can act as the heterogeneous nuclei for the primary silicon grains. Crystal lattice correspondence analysis shows that Al 4 C 3 has a good lattice matching coherence relationship with AlP, and the disregistry is just 5%, so AlP crystals can easily nucleate on Al 4 C 3 surface and have a like-peritectic coupling with them to form quantities of coupling compounds which act as heterogeneous nuclei for the primary silicon grain, which is a possible mechanism for the improvement of phosphorous modification efficiency after the addition of trace TiC particles. Acknowledgements This work was supported by a grant from National Science Fund for Distinguished Young Scholars (No. 50625101), Key Project of Science and Technology Research of Ministry of Education of China (No. 106103) and Shandong Natural Science Foundation (No. Z2004F03). References [1] K. Matsuura, M. Kudoh, H. Kinoshita, H. Takahashi, Mater. Chem. Phys. 81 (2003) 393. [2] B. Yang, F. Wang, J. Zhang, Acta Mater. 51 (2003) 4977. [3] X. Liu, Y. Wu, X. Bian, J. Alloys Compd. 391 (2005) 90. [4] M.M. Haque, A. Sharif, J. Mater. Process. Technol. 118 (1 3) (2001) 69. [5] M.M. El-Alat, J. Eng. Appl. Sci. 50 (3) (2003) 603. [6] J. Zhou, J. Duszczyk, B.M. Korevaar, J. Mater. Sci. 26 (1991) 3041. [7] E.L. Rool, AFS (Am. Foundrymen s Soc.) Trans. 80 (1972) 421. [8] G.K. Sigworth, AFS (Am. Foundrymen s Soc.) Trans. 95 (1987) 303. [9] Y. Kenichi, T. Hiroyasu, S. Tatsuo, K. Akihiko, Mater. Sci. Forum 391 (2000) 331 337. [10] X. Liu, J. Qiao, Z. Wang, L. Yu, Y. Han, X. Bian, Rare Metal. Mat. Eng. 33 (2004) 924. [11] Y. Han, X. Liu, H. Wang, Z. Wang, X. Bian, J. Zhang, Trans. Nonferrous Met. Soc. China 13 (2003) 92. [12] X. Song, X. Bian, X. Qin, X. Liu, J. Zhang, B. Wang, L. Zhu, J. Univ. Sci. Tech. Beijing 11 (2004) 81. [13] C.R. Ho, B. Cantor, Acta Metall. Mater. 43 (1995) 3231. [14] H. Lescuyer, M. Allibert, G. Laslaz, J. Alloys Compd. 279 (1998) 237. [15] P.H. Shingu, J.I. Takamura, Metall. Trans. 1 (1970) 2339. [16] A.R. Kennedy, D.P. Weston, M.I. Jones, Mater. Sci. Eng. A 3 (2002) 16 32. [17] V.H. Lopez, A. Scoles, A.R. Kennedy, Mater. Sci. Eng. A 356 (2003) 316. [18] P.S. Mohanty, J.E. Gruzleski, Acta Metall. Mater. 43 (1995) 2001. [19] W. Kurz, D.J. Fisher, Fundamentals of Solidification, vol. 27, Trans. Tech. Publications, Switzerland, 1986. [20] J.M. Rigsbee, H.I. Aaronson, Acta Metall. 27 (1979) 351. [21] L.B. Bruce, Metall. Trans. 1 (1970) 1987.