A Novel Directional Solidification of TiAl-Based Alloys by Electromagnetic Cold Crucible Zone Melting Technology with Y2O3 Moulds
2018 Volume 59 Issue 5 Pages 816-821
Details
2018 Volume 59 Issue 5 Pages 816-821
In order to decrease mould contamination, a novel directional solidification process was developed for TiAl-based alloys, where an Y2O3 mould was incorporated into electromagnetic cold crucible (EMCC) zone melting technology. To determine the characteristics of this process, the macro/microstructures and mechanical properties of directionally solidified (DS) Ti–45Al–2Cr–2Nb ingots prepared by two kinds of directional solidification techniques, namely traditional graphite heating (control group) and EMCC heating, were extensively investigated using electromagnetic field and temperature field. Compared with the control group, this new technique can induce bigger electromagnetic force in the tangential direction, generate a more rapid heating and higher temperature gradient, and decrease the interaction between the mould and melt; however the heat transfer is altered to inclining outward owe to the lateral heat transfer. The DS sample prepared by this method can achieve finer columnar crystals growing toward the axis, α2/γ lamellae, and lower levels of contamination with regard to Y2O3 particles and oxygen. These are beneficial to improve room temperature fracture toughness and tensile properties.
Lightweight TiAl-based alloys are materials that could potentially be applied in high thrust-to-weight turbine engines in the form of either compressor blades or turbine blades. However, components manufactured from TiAl-based alloys have exhibit two key disadvantages, namely low room-temperature (RT) elongation and poor processing ability, which limit their industrial application.1–4) It is noted that the Bridgman directional solidification process, with a ceramic mould, is an effective approach to not only improve the RT elongation and strength, but also to enable greater geometrical complexity.3) To date, two types of heating mode have been typically used in this type of process, namely refractory resistance heating and graphite heating.5,6) These are indirect heating mode, which means that the heating process occurs over a long period of time. In addition, during all stages of the heating process, the ceramic mould inside the heating chamber is held at a relatively high temperature because of the external heat-transfer behavior. Interaction between the TiAl melt and the mould is promoted by an extended period of contact at the melting temperature.7) Consequently, significant interaction between the melt and mould will occur, inducing poor mechanical properties. Considering the aforementioned information, it is necessary to develop a novel directional-solidification process, which can improve the heating efficiency, shorten the smelting time, and limit the interaction between the mould and the melt.
An internal induction heating mode, which is a direct heating mode, was developed; this process can raise the temperature rapidly and maintain it efficiently. Zhang8) used direct inductive heating to prepare a directionally solidified (DS) AZ31 magnesium alloy within a ceramic mould, which generated a high temperature gradient and reduced the period of interaction between the mould and melt. However, as demonstrated by Li,9) in the case of DS TiAl-based alloys prepared by a direct induction heating mode, the columnar grains were relatively short and discontinuous. As shown in a previous report,10) DS TiAl-based alloys were successfully prepared via electromagnetic cold crucible (EMCC) directional solidification. To incorporate an EMCC, it is effective to reduce and homogenize the electromagnetic forces, and enlarge the melt area. In this case, under the effect of EMCC, the growth interface of DS TiAl alloy became relative stable, and then continuous columnar crystals were obtained.10) Based on the above results, during this study, a novel directional solidification process was developed for TiAl-based alloys, where an Y2O3 mould was incorporated into the EMCC zone melting technology. The macro/microstructure and mechanical properties of the material were intensively investigated using electromagnetism and temperature to determine the characteristics of this novel process. In addition, samples were prepared by traditional graphite heating using identical heating rates, temperatures, and growth rates, and served as the control group.
A master ingot with a nominal composition of Ti–45Al–2Cr–2Nb (at%) was prepared using a cold crucible induction furnace under an argon atmosphere. The experimental installations, with two types of heating modes, are shown in Fig. 1. Figure 1(a) shows the experimental installation with graphite heating; the principle used is similar to that of the installations reported by Lapin et al.11) The graphite body was heated via induction, and subsequently, the entire mould and TiAl ingot within the graphite were heated via radiation. Meanwhile, the base of the TiAl ingot was maintained at a relatively low temperature because of the cooling of the Ga-In alloy. Consequently, during the solidification process, a vertical temperature gradient existed to promote the directional growth of the TiAl ingot. Figure 1(b) shows the installation with EMCC, in which an inductive coil connected to a high-frequency power supply unit was wrapped around a bottomless copper crucible with a cross-section of φ33 mm, which was divided into 12 water-cooled segments. The portion near the base of the TiAl ingot was directly superheated via induction. In addition, the presence of the external cold crucible can reduce the temperature of the ceramic. The temperature was measured and controlled using Φ0.15 mm PtRh30–PtRh6 thermocouples.10)
Experimental installations with two types of heating modes; (a) graphite heating, and (b) EMCC heating.
The DS ingots were cut into the isometric two halves longitudinally. The macrostructure was recorded by Digital Single Lens Reflex. The microstructures were investigated by the scanning electron microscope (SEM) in backscattered electron (BSE) mode and transmission electron microscopy (TEM). The oxygen content was measured by the inert gas infrared-thermal conductivity technique (IGI). The volume fraction of Y2O3 particles was determined by computerized image analysis.
All properties including fracture toughness and tensile properties were measured by the electronic universal test machine with a loading speed of 0.5 mm/min at RT. Fracture toughness was tested by using single edge notched bend (SENB) specimens (16 mm × 4 mm × 2 mm). The fatigue precrack was introduced by the wire-electrode cutting in consideration of the hard brittleness of TiAl-based alloys. Flat dog-bone-like tensile specimens were with the tensile axis parallel to the growth direction, and had a 7 mm gauge length with a 2 mm × 2 mm cross-section. At least five specimens cutting from each sample were tested.
2.1 Characteristics of the EMCC heating mode 2.1.1 Electromagnetic characteristicsThe flow generated by the electromagnetic effects would strongly influence the solute distribution, grain growth, and lamellar structure.12) Therefore, investigations on the flow are essential with regard to the DS processing of TiAl. However, owing to space limitations and the chemical reactivity of the TiAl melt, it is difficult, or even impossible to measure the flow of the TiAl melt. Electromagnetic analysis can be effectively used to study the flow field. The magnetic flux density of the DS system was measured via the small coil method, which is based on Faraday’s law.13) Consequently, B could be calculated using the following formula:
| \begin{equation} B = \frac{E}{4.44fNS} \end{equation} | (1) |
Electromagnetic characteristics of two heating modes, (a) measured and calculated results for B at various points, where a1–a3 and b1–b3 represent graphite heating and EMCC heating, respectively; (b) scheme of vector B decomposition in alternating magnetic field; (c) electromagnetism on surface of the melt.
As clearly shown in Fig. 2(b), vector B of each unit body can be divided into two directions, tangential Bθ and axial Br.
| \begin{equation} \overrightarrow{B_{\text{r}}} = | \vec{B} |\cos\theta \end{equation} | (2) |
| \begin{equation} \overrightarrow{B_{\theta}} = | \vec{B} |\sin\theta \end{equation} | (3) |
As noted by Xu,14) axial Br generates a melt flow that is driven by an electromagnetic force and rotates around point O, while Bθ generates an electromagnetic pressure on the melt, shown in Fig. 2(c). And the electromagnetic force strength can be calculated in the form:
| \begin{equation} \skew5\vec{F} = \overrightarrow{j_{\text{z}}} \times \skew3\vec{B} \end{equation} | (4) |
| \begin{equation} \skew3\vec{J} = \sigma_{\text{e}}[\skew3\vec{E} + \vec{\mu} \times \skew3\vec{B}] \end{equation} | (5) |
Combining Maxell law in induction heating, expressed the electromagnetic force driving melt to rotate around O point,14) and electromagnetic pressure on the melt:
| \begin{equation} \overrightarrow{F_{\theta}} = - \frac{1}{8}\overrightarrow{B^{2}}\sigma_{\text{e}}\mu_{0}\left(\omega - \frac{V_{\theta}}{r}\right)^{2}\vec{r} \end{equation} | (6) |
| \begin{equation} \overrightarrow{F_{\text{r}}} = - \frac{1}{8}\overrightarrow{B^{2}}\sigma_{\text{e}}{}^{2}\mu_{0}\left(\omega - \frac{V_{\theta}}{r} \right)^{2}\overrightarrow{r^{3}} \end{equation} | (7) |
It is recognized that the heat transport generated by a temperature gradient is an important factor, which can affect the growing interface and columnar growth, especially in the direction of columnar growth. Figures 3(a) and (b) exhibit schematic diagrams of the heat transfer during the directional solidification process under two types of heating modes, in which qx, qy, and q represent the lateral, longitudinal, and actual heat dissipation, respectively. And, the three parties meet the following relationship:
| \begin{equation} q = q_{\text{x}} + q_{\text{y}} \end{equation} | (8) |
(a) and (b) schematic diagram of heat transfer during the directional solidification process; (c) and (d) Temperature distribution along the longitudinal section of the mold during directional solidification at a melt temperature of 1923 K, in which (a) and (c) are these of graphite heating, (b) and (d) are these of EMCC heating.
During the directional solidification process with graphite heating, there is no lateral heat dissipation because of the external heater. However, in case of the process with EMCC heating, lateral heat dissipation is inevitable owing to the internal heater and the cooling of the external EMCC. The direction of the actual heat dissipation will alter, inclining outwards.
The interaction between the melt and mould is an essential and important part of this type of process, which depends greatly on the interfacial temperature (T) and reaction time (tr). The interfacial temperature may be related to the thermal-field distribution within the Y2O3 mould, which is owed to its porous structure. As indicated by Cui et al.,15) the reaction time is determined by the heating and solidification (depending on the temperature gradient) processes. To clearly characterize the difference between the interactions that occur using the graphite and EMCC heating modes, investigations on the temperature distributions inside the Y2O3 moulds, as well as on the heating process and temperature gradient, were performed.
The results of calculations regarding the differences in the thermal-field distributions within the Y2O3 ceramic moulds, based on ANSYS analyses, are shown in Figs. 3(c) and (d). Figure 3(c) relates to the indirect graphite heating mode. To ensure the appropriate melt temperature, the graphite heater should be maintained at a temperature 150 K above that of the mould. In this case, the heating is gradually conducted inward to the inner surface of the mould, and the entire section of the mould is at a high temperature, with an estimated range of 1923–2100 K. In particular, the temperature of the interface between the metal and mould is approximately 1923 K, which is similar to the melting temperature of the TiAl ingot. As shown in Fig. 3(d), the temperature of the entire mould is relatively low; the highest temperature is determined at the metal/mould interface; however, this temperature is not greater than the melting temperature of the TiAl ingot. The temperature of the outer surface of the mould would decrease to approximately 600 K because of the cooling effect of the EMCC. Therefore, the heat is gradually transferred outwards to the surface through the cross section of the mould.
Figure 4 shows the heating curves of the TiAl ingots obtained during the melting processes with the two types of heating modes; the red curve indicates the graphite heating mode, and the black curve indicates the EMCC heating mode. The analysis and statistics results regarding the temperature gradients and melt-residence period during the melting process are shown in Table 1. At identical melt temperatures, the EMCC heating mode is advantageous because it offers more rapid heating and a greater temperature gradient. Irrespective of the melt residence time during the melting process or the solidification process period, the period of local interfacial compatibility, tr, decreased with regard to the melt and mould. In addition, the zone-melting characteristics will further reduce the interaction period. Hence, compared with the graphite heating mode, the use of the EMCC heating mode will greatly weaken the interaction. The above conclusions can be verified by the observational results of the metal/mould interfacial morphologies, and the occurrence of contamination, including Y2O3 particles and oxygen, following the directional solidification experiments under identical melt temperatures (1923 K) and growth rates (0.8 mm/min), as shown in Figs. 5(a)–(c). A comparison of the metal/mould interfacial morphologies is shown in Figs. 5(a) and (b); the ceramic layer of the sample prepared by graphite heating is much thicker than that of the sample prepared by EMCC heating. Figure 5(c) shows the results obtained for the DS samples prepared via the two types of heating modes under identical melt temperatures and growth rates; the results show increases in the volume fraction of the Y2O3 particles and the oxygen content. In the case of the DS samples prepared by the two heating modes, there is an increase in both the volume fraction of the Y2O3 particles and the oxygen content; however, the values of the results obtained for the samples prepared via graphite heating are approximately double those of the samples prepared by EMCC heating.
Time–temperature curves, in which the red curve indicates the graphite heating mode, and the black curve indicates the EMCC heating mode.
(a) and (b) metal/mold interfacial microstructures following the directional solidification experiments with graphite heating and EMCC heating, respectively, under a temperature of 1923 K and growth rate of 0.8 mm/min; (c) volume fraction of Y2O3 particles, and increase in oxygen content.
Figure 6 shows the structures of the DS samples prepared via the two types of heating modes under identical smelting temperatures (1923 K) and growth rates (0.8 mm/min). Figures 6(a)–(d) show the typical macrostructures of the DS samples, in which the white dotted line and the red line represent the growth directions of the interface and columnar crystals, respectively. The structure of the longitudinal section parallel to the growth direction can be divided into two regions, namely a transient region composed of coarse isometric grains and a DS region composed of slender columnar crystals, as shown in Figs. 6(b) and (d). As shown in Figs. 6(e) and (f), the typical microstructure of the DS region consists of α2/γ lamellae, B2 phase and blocky γ phase,15) and Y2O3 particles, which are generated by the reaction between the Y2O3 mould and the melt, as described in previous reports.16,17) As described in Refer,18,19) the samples prepared by two kinds process under same solidification parameters have essentially same alignment of lamellae.
Structures of DS samples prepared by the two types of heating modes under a temperature of 1923 K and growth rate of 0.8 mm/min; (a) and (c) cross-section; (b) and (d) longitudinal section; (e) and (f) typical microstructure of the DS region; (g) and (h) TEM images showing the α2/γ lamellae of the DS samples prepared by the graphite heating and EMCC heating modes, respectively.
However, some differences do exist in the case of the structures of the DS samples prepared by the two types of heating modes. The DS sample prepared via the EMCC heating mode offers a shorter transient region, and finer columnar crystals (Figs. 6(a)–(d)) and lamellae (Figs. 6(g) and (h)). In detail, the mean width of columnar crystals of sample prepared via the EMCC heating mode is about 1.3 mm which is much less than 2.2 mm of samples prepared by graphite heating. And, the mean lamellar spacing is about 0.34 µm which is much less than 0.55 µm of samples prepared by graphite heating. Previous works demonstrated substantial flow in the tangential direction; a large temperature gradient could result in the refinement of the columnar crystals and α2/γ lamellae,10,20) and promote the elimination of non-preferential grains as well as accelerate the transition from equiaxed grains to columnar crystals (which is equivalent to the length of the transient region).20) This can explain the results regarding the finer structure and shorter transient region in the case of the DS samples prepared by the EMCC heating mode. Moreover, the interface becomes concave and the columnar crystals grow toward the axis (Fig. 6(d)). As shown by Yang,12) columnar crystals grow in the direction opposite to that of the heat dissipation. As described in section 2.1.2, under the influence of lateral heat dissipation, the direction of actual heat dissipation inclines outward. Hence, columnar crystals will grow toward the axis. Similar phenomena can be observed in the case of samples prepared by the cold crucible directionally solidified (CCDS) process.10,12)
2.3 Mechanical propertiesFigures 7(a) and (b) show the mechanical properties, including the RT fracture toughness and tensile properties, of the DS samples prepared by the two types of heating mode. It is clear that the RT fracture toughness and tensile properties of the DS samples prepared by the EMCC heating mode are superior to those of the DS samples prepared by the graphite heating mode.
Variation in the mechanical properties of the DS samples prepared by the two types of heating modes under a temperature of 1923 K and growth rate of 0.8 mm/min; (a) fracture toughness, and (b) tensile properties.
The RT fracture toughness and tensile properties of DS TiAl-based alloys depend on several microstructural parameters, namely the orientation of the lamellae to the loading axis, inter-lamellar spacing, precipitation of the B2 phase, and contamination, such as that involving Y2O3 particles and increased oxygen contents.1,17–21) As described in previous chapters, compared with the DS sample prepared by the graphite heating mode, the sample prepared by the EMCC heating mode possesses finer columnar crystals and lamellae, and exhibits lower levels of contamination, which are all beneficial with regard to improving the RT fracture toughness and tensile properties. These would explain the superior mechanical properties of the samples prepared by this novel process.
In order to improve the heating efficiency and weaken the interaction between mould and TiAl melt, a directional solidification process with EMCC heating mode was developed. The following conclusions are reached:
We gratefully acknowledge the support of this work by National Nature Science Foundation of China (Grant No. 51471062 and Grant No. 51671072).
