Microstructure and grain refining efficiency of Al-5Ti-1B master alloys prepared by halide salt route

Microstructure and grain refining efficiency of Al-5Ti-1B master alloys prepared by halide salt route

Journal of Materials Processing Technology 246 (2017) 205–210 Contents lists available at ScienceDirect Journal of Materials Processing Technology j...

3MB Sizes 1 Downloads 19 Views

Journal of Materials Processing Technology 246 (2017) 205–210

Contents lists available at ScienceDirect

Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec

Microstructure and grain refining efficiency of Al-5Ti-1B master alloys prepared by halide salt route Lili Zhang a,b,1 , Hongxiang Jiang a,1 , Jiuzhou Zhao a,b,∗ , Jie He a a b

Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China University of Chinese Academy of Sciences, Shenyang 110016, China

a r t i c l e

i n f o

Article history: Received 2 December 2016 Received in revised form 28 March 2017 Accepted 28 March 2017 Keywords: Grain refinement Microstructure Master alloy

a b s t r a c t Al-5Ti-1B master alloys were prepared by reacting K2 TiF6 and KBF4 salts with Al melt. During the reaction period between the fluoride salts and Al melt, the non-homogenous distribution of solutes B and Ti in the Al melt close to the salts/Al melt interface causes the formation of the unstable AlB2 phase. The AlB2 phase transforms to TiB2 particles by the diffusion of solutes Ti and B in the Al melt during the afterward holding temperature period, which causes an increase both in the number density of TiB2 particles and in the grain refining efficiency of the master alloys. After the transformation of AlB2 phase to TiB2 phase, TiB2 particles coarsen by Ostwald ripening, which leads to a decrease both in the number density of TiB2 particles and in the grain refining efficiency of the master alloys. © 2017 Published by Elsevier B.V.

1. Introduction Adding grain refiners to alloy melts has become a common industrial practice to achieve a fine, equiaxed grain structure of Al alloys. Al-Ti-B master alloys, in particular Al-5Ti-1B (all compositions quoted in this work are in wt.% unless otherwise specified), have been widely used as grain refiners over the past half century. Much work has been done to investigate the solidification behaviours of the Al alloys inoculated with Al-Ti-B master alloys since the 1970s. Models, such as Maxwell-Hellawell model (Maxwell and Hellawell, 1975), Free Growth model (Greer et al., 2000) and Interdependence Theory (StJohn et al., 2011) were proposed to describe the microstructure formation under the effect of inoculants. These researches clearly demonstrate that the grain refining efficiency is closely related to the concentration of solute Ti as well as the number density and size distribution of TiB2 particles in the melt. Currently, Al-5Ti-1B master alloys are mostly prepared by reacting K2 TiF6 and KBF4 salts with Al melt. Numerous techniques have been used to control the microstructures of Al-5Ti-1B master alloys. Li et al. (2006) found that rapid solidification processing could narrow the distribution of TiB2 particles and thus lead to an improved grain refining efficiency of the master alloy. Han et al. (2005) found

∗ Corresponding author at: Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China. E-mail address: [email protected] (J. Zhao). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.jmatprotec.2017.03.029 0924-0136/© 2017 Published by Elsevier B.V.

that the application of high-intensity ultrasound during the reaction of fluoride salts with Al melt could significantly reduce the size as well as its spread of TiB2 particles and enhance the grain refining efficiency of the master alloy. Zhao et al. (2015) found that the application of electromagnetic stirring led to a homogenous dispersion of the second phases in the Al-5Ti-1B master alloy. Although much work has been done to control the microstructure of Al-5Ti-1B master alloys, there are few systematic studies on the microstructure evolution during the preparation especially during the holding temperature period after the reaction between the fluoride salts and Al melt. The present work was thus carried out to investigate the microstructure evolution during the holding temperature period after the reaction between the fluoride salts and Al melt as well as its effect on the grain refining efficiency of the Al-5Ti-1B master alloys.

2. Experiments Al-5Ti-1B master alloys were prepared as follows: commercialpurity Al ingot (99.7%) was first melted and heated to 1023 K in a graphite crucible using a medium-frequency induction furnace. The mixture of K2 TiF6 and KBF4 salts, weighed to produce an Al-5Ti-1B master alloy nominally, was then added to the Al melt. The melt was finally cast into a mould after holding at 1023 K for different time. A total of 10 alloys were prepared with the holding temperature time (simply HTT in the following) of 25 min, 35 min, 45 min, 55 min, 60 min, 65 min, 90 min, 110 min, 130 min and 150 min, respectively. The compositions of Al-Ti-B master alloys were analysed using

206

L. Zhang et al. / Journal of Materials Processing Technology 246 (2017) 205–210

Table 1 Recovery rates of Ti and B in the alloys with different HTT. HTT (min)

Ti recovery (%)

B recovery (%)

25 35 45 55 65

86.6 97.9 99.8 99.4 99.1

87.9 95.2 99.9 99.2 99.4

inductively coupled plasma-atomic emission spectrometry (ICPAES). The microstructures were characterized by field-emission scanning electron microscopy (FESEM, Inspect F50) equipped with an energy dispersive X-ray spectroscopy (EDS). The phases were identified by X-ray diffraction (XRD, Moldel D/Max 2500PC Rigaku). TiB2 particles in the master alloys were extracted by dissolving the Al matrix and aluminide with diluted HCl solution for the analysis of size distributions. The size distribution of TiB2 particles was determined by quantitative metallographic analysis using SISC IAS V8.0 software, which is an image analysis software capable of taking information from photographs and calculating the number density, average size, area fraction, etc. of the phases or defects, just as Image-Pro Plus software does. The experimental procedure for examining the grain refining potency of the master alloys was as follows: commercial-purity Al ingot was first melted and heated to 983 K in a graphite crucible using an electric resistance furnace. 0.4% Al-5Ti-1B master alloy was then added to the melt. After holding at 983 K for 15 min, the melt was poured into a cast-iron mould pre-heated to 423 K to form a frustum sample with a height of 12 cm and top and

bottom diameters of 5 cm and 1 cm, respectively. A reference sample without inoculation was prepared under the same conditions. The temperature at the centre of the sample was measured using tungsten-rhenium thermocouple. The samples were sectioned approximately 2 cm from the base and metallographic specimens were prepared. These specimens were examined using a Zeiss optical microscope with polarized light. SISC IAS V8.0 software was used to determine the average grain size of commercial-purity Al.

3. Experimental results Table 1 shows the recovery rates of Ti and B at different HTT. It demonstrates that the recovery rates are higher than 99% once the HTT reaches 45 min, indicating that the reaction between the fluoride salts and Al melt almost finishes. Three distinguished phases, including white blocky phase, particulate phase and a little of flake-like phase, can be observed in the master alloys with the HTT of 25, 45 and 60 min (Fig. 1). The XRD patterns (Fig. 2) indicate that the three phases are TiAl3 , TiB2 and AlB2 phases, respectively. The volume fraction of TiAl3 particles decreases while the volume fraction of TiB2 particles increases with the increase of HTT when the HTT is less than 90 min and both of them remains almost constant after the HTT of 90 min, as shown in Fig. 3. The amount of AlB2 phase shows an increasing tendency with HTT during the reaction period between the fluoride salts and Al melt and a decreasing tendency with HTT after the reaction. The AlB2 phase disappears when the HTT is longer than 90 min (Fig. 1). These results indicate that AlB2 phase can transform to TiB2 phase and the transformation almost finishes when the HTT is 90 min.

Fig. 1. FESEM images of Al-5Ti-1B alloys with the HTT of (a) 25, (b) 45, (c) 60, (d) 90, (e) 110, (f) 130 and (g) 150 min.

L. Zhang et al. / Journal of Materials Processing Technology 246 (2017) 205–210

207

where [Ti] and [B] represent the solutes of Ti and B in the Al melt, respectively. The standard Gibbs free energy of formation of TiAl3 , TiB2 and AlB2 phases can be calculated by (Geoffery, 1984): ␪ GTiAl = −76.5 + 0.0227T (KJ/mol)

(4)

␪ GTiB = −307.1 + 0.0585T (KJ/mol)

(5)

␪ GAlB = −201.2 + 0.0905T (KJ/mol)

(6)

3

2

2

Eqs. (4)–(6) indicate that, in the temperature range 950–1050 K, the standard Gibbs free energy of formation of TiAl3 , TiB2 and AlB2 phases are all negative, and the standard Gibbs free energy of formation of TiB2 phase is the most negative. In other words, TiB2 phase is thermodynamically more stable than AlB2 and TiAl3 phases. The equilibrium constants for the Reactions (1)–(3) can be ␪ calculated by Eqs. (4)–(6). The solubility product of TiAl3 (KTiAl ),

Fig. 2. XRD patterns of Al-5Ti-1B alloys with different HTT.

3

␪ ␪ TiB2 (KTiB ) and AlB2 (KAlB ) in the Al melt can thus be obtained. They 2

2

limit respectively the lowest concentration (x) or concentration product (K) for the precipitation of TiAl3 , TiB2 and AlB2 intermetallic compounds, as shown in Fig. 8. The fact that the master alloy, which is cast just after the reaction between the fluoride salts and Al melt, contains both TiAl3 and AlB2 phases besides TiB2 phase (see Fig. 1(a)) indicates that the molar fraction of solute Ti (xTi ) is above 0.19 at% at some locations in the melt (see the red region in Fig. 8), while at some other locations the molar fraction of solute B (xB ) is above 0.169 at% (see the blue region in Fig. 8). In other words, the concentration of solutes Ti and B is non-homogenous in the Al melt close to the interface between the fluoride salts and Al melt.

Fig. 3. Volume fractions of TiAl3 and TiB2 particles in the Al-5Ti-1B alloys with different HTT.

Calculations based on the volume fraction of TiB2 particles indicate that the volume fraction of AlB2 in the master alloy with the HTT of 45 min, in which the content of AlB2 is the highest, is less than 0.2 vol%. Figs. 4 and 5 show the size distribution, number density (NTiB2 ) and average diameter of TiB2 particles in the Al-5Ti-1B master alloys with different HTT. They demonstrate that the number density of TiB2 particles reaches its maximum and the average diameter of TiB2 particles achieves its minimum when the HTT is 90 min. Figs. 6 and 7 show the microstructures and average grain size of commercial-purity Al. They demonstrate that a significant grain refinement of ␣-Al can be achieved by adding 0.4% Al-5Ti-1B master alloy, and the Al-5Ti-1B master alloy with the HTT of 90 min exhibits the best grain refining performance. 4. Discussion 4.1. Phases in the master alloy just after the reaction between the salts and Al melt When the salts of K2 TiF6 and KBF4 are introduced into the Al melt, Ti and B atoms are reduced from the fluoride salts and enrich at the interface between the salts and Al melt. The enrichment of Ti and B may cause the formation of TiAl3 , TiB2 and AlB2 intermetallic compounds (Fan et al., 2005): [Ti] + 3Al(l) → TiAl3 (s)

(1)

[Ti] + 2[B] → TiB2 (s)

(2)

2[B] + Al(l) → AlB2 (s)

(3)

4.2. Microstructure evolution during the holding temperature period after the reaction between the salts and Al melt 4.2.1. Transformation of AlB2 phase to TiB2 phase Although AlB2 phase can form during the reaction period between the fluoride salts and Al melt, it is not thermodynamically stable in the presence of solute Ti. It tends to transform to TiB2 phase. The transformation process is controlled by the diffusion of solutes Ti and B in the melt, as shown schematically in Fig. 9. The B content in the Al melt at the interface of Al(l)/AlB2 (s) is determined by the solubility product of AlB2 . The boundary conditions for the concentration field on the surface of an AlB2 flake are thus as follows:

⎧  ⎨ x (z = ␦, t) ≈ K ␪ B AlB2 ⎩ xTi (z = ␦, t) = K ␪ /x2 (z = ␦, t) TiB2

(7)

B

The molar ratio of Ti/B in the Al-5Ti-1B alloy is 1.15. It is much greater than 0.5 (the molar ratio of Ti/B in the TiB2 phase). There are thus a lot of TiAl3 particles in the master alloy. It may be reasonable to assume that the solute Ti content in the melt far from an AlB2 flake is determined by the solubility product of TiAl3 . That is to say that the boundary conditions for the concentration field in the Al melt far from AlB2 are as follows:

⎧ ␪ ⎨ xTi (z → ∞, t) ≈ KTiAl 3  ⎩ xB (z → ∞, t) = K ␪ /xTi (z → ∞, t) TiB2

(8)

The non-homogenous distribution of xTi and xB causes a diffusion transfer of solutes Ti and B in the melt. Because xB (z = ␦, t) is much higher than xB (z → ∞ , t) and xTi (z = ı, t) is much lower than xTi (z → ∞ , t), the solute B diffuses from the surface of AlB2 flake to the Al melt while the solute Ti diffuses in the opposite direction. The diffusion transfer of solutes Ti and B leads to the growth of the TiB2 particles in the melt. It may also cause the nucleation of TiB2

208

L. Zhang et al. / Journal of Materials Processing Technology 246 (2017) 205–210

Fig. 4. Size distributions of TiB2 particles in the Al-5Ti-1B alloys with different HTT. (a) 45, (b) 60, (c) 90, (d) 110, (e) 130 and (f) 150 min.

4.3. Grain refining efficiency of Al-5Ti-1B master alloys

Fig. 5. Number density (solid line) and average diameter (dashed line) of TiB2 particles in the Al-5Ti-1B alloys with different HTT.

␪ particles in the region where the supersaturation (KTiB2 − KTiB ) is 2

higher enough (shadowed area in Fig. 9). This process continues until all the AlB2 flakes transform to TiB2 phase.

4.2.2. Coarsening of TiB2 particles After the transformation of AlB2 phase to TiB2 phase, TiB2 particles coarsen under the effect of the interfacial energy between Al melt and TiB2 particles. The experimental results demonstrate that ¯ of TiB2 particles the time dependence of the average diameter (d) follows a power law d¯ 3 ∝ t, as shown in Fig. 10. This is in agreement with the Ostwald ripening theory (Balluffi et al., 2005).

Recently, Wang et al. found that AlB2 has a grain refining ability on the commercial-purity Al (Wang et al., 2011). The average grain size of the commercial-purity Al inoculated with 0.011% AlB2 by using Al-3B master alloy is larger than 150 ␮m (Wang et al., 2011). The grain refinement experiments of the present work were carried out under the similar solidification conditions as Wang’s. Calculations indicate that the content of AlB2 in the Al melt due to the addition of Al-5Ti-1B master alloy is less than 0.001%, much lower than 0.011. One can, thus, conclude from the fact that the average grain size of commercial-purity Al inoculated by the Al-5Ti-1B master alloy with the HTT of 45 min is much smaller than 150 ␮m (about 80 ␮m) that the refinement is caused by TiB2 , not by AlB2 . The grain refining efficiency of Al-Ti-B master alloys is closely related to the size distribution of TiB2 particles. Quested and Greer (2004) proposed that the critical diameter dc for a TiB2 particle to act as the effective nucleation substrate is inversely proportional to the undercooling Tfg of the melt: dc = 4/SV T

(9)

where ␴ = 0.158 J/m2 and SV = 1.112 × 106 J/(m3 K) (Qian et al., 2010) are the solid-liquid interfacial energy and the entropy of fusion per unit volume of Al, respectively. In the present experiment, the commercial-purity Al with the addition of 0.4% Al-5Ti-1B master alloys achieved an undercooling of about 2.7 K. It can be concluded from Eq. (9) that the critical diameter is about 0.21 ␮m. This result demonstrates that all the TiB2 particles in the master alloys (see Fig. 4) can act as the nucleating substrates under the present experimental conditions. In other

L. Zhang et al. / Journal of Materials Processing Technology 246 (2017) 205–210

209

Fig. 6. Microstructures of commercial-purity Al. (a) non-inoculated and inoculated with 0.4% Al-5Ti-1B master alloy with the HTT of (b) 45, (c) 60, (d) 90, (e) 110, (f) 130 and (g) 150 min.

Fig. 7. Average grain size of commercial-purity Al inoculated with 0.4% Al-5Ti-1B alloys with different HTT.

words, the grain refining efficiency of the master alloys depends on the number density of TiB2 particles, and thus has the same tendency with the HTT as the number density of TiB2 particles does (see Figs. 5 and 7). 5. Conclusions The microstructure evolution and its impact on the grain refining efficiency of the Al-5Ti-1B master alloys during the holding

Fig. 8. Concentration of solutes Ti and B for the formation of TiAl3 , TiB2 and AlB2 phases.

temperature period after the reaction between the fluoride salts and Al melt were investigated. The conclusions are as follows: (1) During the reaction period between the fluoride salts and Al melt, the concentration of solutes B and Ti in the Al melt close to the fluoride salts/Al melt interface is non-homogenous. The non-homogenous distribution of solute B causes the formation of the unstable AlB2 phase. (2) AlB2 phase formed during the reaction period between the fluoride salts and Al melt is not thermodynamically stable in the

210

L. Zhang et al. / Journal of Materials Processing Technology 246 (2017) 205–210

Acknowledgements This work was supported by the National Natural Science Foundation of China [grant numbers 51501207 and 51471173], Natural Science Foundation of Liaoning Province [grant number 201501043] and China’s Manned Space Station Project [mission number TGJZ800-2-RW024]. References

Fig. 9. Schematic illustration of the transformation of AlB2 phase to TiB2 phase. KTiB2 , c is the critical supersaturation for the nucleation of TiB2 particles, 2ı is the thickness of an AlB2 flake.

Fig. 10. Average diameter of TiB2 particles in the Al-5Ti-1B master alloys with different HTT.

presence of solute Ti. It transforms to TiB2 particles by the diffusion of solutes Ti and B in the Al melt during the afterward holding temperature period. This transformation causes an increase both in the number density of TiB2 particles and in the grain refining efficiency of the master alloys. (3) After the transformation of AlB2 phase to TiB2 phase, Ostwald ripening of TiB2 particles occurs. It leads to a decrease both in the number density of TiB2 particles and in the grain refining efficiency of the master alloys.

Balluffi, R.W., Allen, S.M., Carter, W.C., 2005. Kinetics of Materials. Wiley Interscience, Canada, pp. 363–372. Fan, T.X., Yang, G., Zhang, D., 2005. Thermodynamic effect of alloying addition on in-situ reinforced TiB2 /Al composites. Metall. Trans. A 36, 225–233. Geoffery, K., 1984. The grain refining of aluminium and phase relationships in the Al-Ti-B system. Metall. Trans. A 15, 277–282. Greer, A.L., Bunn, A.M., Tronche, A., Evans, P.V., Bristow, D.J., 2000. Modeling of inoculation of metallic melts: application to grain refinement of aluminium by Al-Ti-B. Acta Mater. 48, 2823–2835. Han, Y.F., Li, K., Wang, J., Shu, D., Sun, B.D., 2005. Influence of high-intensity ultrasound on grain refining performance of Al-5Ti-1B master alloy on aluminium. Mater. Sci. Eng. A 405, 306–312. Li, K., Rao, L., Yan, H., Wang, J., Sun, B.D., 2006. Effect of rapid solidification on the microstructure and grain refinement of Al-Ti-B master alloy. Foundry 55, 894–897. Maxwell, I., Hellawell, A., 1975. A simple model for grain refinement during solidification. Acta Metall. 23, 229–237. Qian, M., Cao, P., Easton, M.A., McDonald, S.D., StJohn, D.H., 2010. An analytical model for constitutional supercooling-driven grain formation and grain size prediction. Acta Mater. 58, 3262–3270. Quested, T.E., Greer, A.L., 2004. The effect of the size distribution of inoculant particles on as-cast grain size in aluminium alloys. Acta Mater. 52, 3859–3868. StJohn, D.H., Qian, M., Easton, M.A., Cao, P., 2011. The interdependence theory: the relationship between grain formation and nucleant selection. Acta Mater. 59, 4907–4921. Wang, T.M., Chen, Z.N., Fu, H.W., Xu, J., Fu, Y., Li, T.J., 2011. Grain refining potency of Al-B master alloy on pure aluminium. Scripta Mater. 64, 1121–1124. Zhao, Z.Y., Guan, R.G., Guan, X.H., Zhang, J., Sun, X.P., Liu, H.N., 2015. Effects of electromagnetic stirring shearing, and extrusion on TiB2 and TiAl3 particles in Al-5Ti-1B(wt.%). Alloy Mater. Manuf. Process. 30, 1223–1228.