EBM: Electron Beam Melting; CCM/SM: Cold Crucible Melting or Skull Melting; VAR: Vacuum Arc Remelting; PAM: Plasma Arc Melting; ISM: Induction Skull Melting; EDS: Energy-Dispersive Spectroscop

Due to its advantages of light weight, good toughness, excellent high temperature performance, biocompatibility and corrosion resistance, titanium alloys are widely used in aerospace and automotive industries, biomedical components and surgical instrument manufacturing, chemical and petrochemical engineering, marine applications and other fields [1-5]. Although the potential is great, the production of titanium and titanium alloys is usually low due to high costs since the very high processing temperature (pure Ti, melting point (m.p.)=1668 °C; Ti6Al4V alloys, m.p.=1640 °C; TiAl(γ)alloys, 42~50 at.% Al, m.p.=1485~1575 °C; TiNi alloys, m.p.=1240~1310 °C). Unfortunately, at such high temperatures, molten titanium becomes extremely active, and it is chemically corrosive to almost anything it comes into contact with [6]. How to obtain titanium and titanium alloys with low cost, high purity, uniform microstructure and excellent performance is a real problem to be solved at present.

To obtain titanium alloy products, casting is indispensable, of course, powder metallurgy can also be selected, but the powder metallurgy routes has some disadvantages such as ingot chemical and microstructural heterogeneity, components geometry are limited, porosity leads to serious oxygen contamination, and high cost [7]. Thus, casting seems to be the first way to obtain high quality titanium alloy products at low cost. However, due to the active chemical nature of molten titanium, it is easily contaminated, such as oxygen and nitrogen in the air, and the crucible material used for melting. Therefore, the preparation process of titanium alloy must be carried out in a vacuum or inert protective atmosphere, which will undoubtedly increase the cost of production, and more importantly, how to select a suitable crucible refractory. Although with the rapid development of metallurgical technology, some new concepts of melting titanium alloys have been developed, such as levitation melting (LM) [8] and Cold Crucible Melting or Skull Melting (CCM/SM) [9], the core idea of which is to avoid direct contact between the molten metal and the crucible. LM technique uses electromagnetic force or gas pressure to suspend the molten metal, while CCM/SM technique forms a solid skull with the same composition as the parent alloy on the surface of the water-cooled copper crucible. According to different heating principles, CCM/SM technique has spawned various techniques such as Vacuum Arc Remelting (VAR) [10], Induction Skull Melting (ISM) [11], Electron Beam Melting (EBM) [12], and Plasma Arc Melting (PAM) [13]. However, LM technique is obviously not suitable for large-scale production. VAR technique needs to press the pre-electrode before melting, which not only increases the cost but also requires high cleanliness of raw materials. In addition, the inclusion and segregation of VAR castings are serious, which requires multiple remelting and long subsequent heat treatment. ISM technique, due to a part of the heat has taken away by the water-cooled copper crucible, which results in low energy efficiency and is not easy to form superheat molten metal, as well as heterogeneous composition and microstructure of castings. EBM and PAM technique also have the problem of low superheat, only superheat near the heat source, resulting in poor fluidity of the molten metal. What’s more, EBM technique will lead to the volatilization of element with high evaporation pressures, such as Al, Sn, Cr, etc., so that the fluctuation of ingot composition is difficult to control. In general, contactless melting technique has its advantages, but its disadvantages are also obvious. Perhaps in the future, various heat sources can be integrated, combined with the advantages of all parties, through numerical modeling and analysis, to further understand the distribution of temperature field in the melting process, so as to optimize the entire process.

Since the preparation method without direct contact with the crucible has low energy efficiency, high cost and heterogeneous composition and microstructure of the ingot (generally requires subsequent processing and modification, which further increases the cost), so it goes back to the previous point of view-choose the appropriate crucible refractories are particularly importantThe development of inert and durable refractories for melting titanium alloys is definitely valuable work. Refractories that need to be used and in close contact with the melt are mainly Vacuum Induction Melting (VIM) [14,15]. Compared with other melting methods, because the thermal conductivity of refractory materials is generally lower, VIM technique allows the metal liquid to superheat, and the composition and microstructure of the casting are more homogenization due to electromagnetic stirring. More valuable, VIM technique is inexpensive and suitable for large-scale industrial production. The most critical part of VIM is how to choose refractories that are inert, durable and inexpensive, so as to apply to industrial production.

In the past two decades, in order to find suitable refractory materials for melting titanium and titanium alloys, various high-temperature ceramic materials such as oxides, carbides, nitrides, silicides, sulfides, and borides have been evaluated, but the results have been not satisfying [16]. For example, BN and AlN ceramic crucible melting titanium alloy will cause interface reaction [17,18], while using graphite crucible will cause carbon contamination of the titanium alloy [19]. Among the most used ceramic oxides, generally only Al2O3, ZrO2, CaO, and Y2O3 meet the relevant thermodynamic considerations (Figure 1). But Al2O3 and ZrO2 will form an inevitable reaction layer with the titanium alloy melt at the interface [20,21]. Due to thermodynamically stable and inexpensive, CaO is a very promising refractory for melting titanium alloys, but CaO crucibles have poor water resistance and high oxygen content in melted titanium alloys. Studies have shown that the contamination of Y2O3 crucibles for titanium alloys is far less than other refractories, but expensive and poor thermal shock resistance are fatal flaws. It can be seen that a single refractory material will always have one or other defects, and it is difficult to meet all needs. Faced with this situation, how should we choose? In order to solve this problem, two ideas of coating [21,22] and doping [23-26] came into being. For example, the corrosion resistance of Y2O3 crucible is good, but the cost is high and the thermal shock resistance is poor, while the performance of Al2O3 crucible is just the opposite. Therefore, Y2O3 can be coated on the inner wall of the Al2O3 crucible. The formed Y2O3/ Al2O3 crucible perfectly combines the advantages of the two materials and avoids the disadvantages of both sides. There are also many examples of doping, such as CaO doped ZrO2, BaO doped ZrO2 (CaCO3 or BaCO3 and ZrO2 are solid-phase sintered at a molar ratio of 1:1 to form CaZrO3 or BaZrO3) [26-29]. CaZrO3 and BaZrO3 are perovskite refractories and meet the thermodynamic conditions for melting titanium and titanium alloys (Figure 1). Since BaZrO3 has higher thermodynamic stability than CaZrO3, and literature shows that BaZrO3 crucible has less contamination to titanium alloy than Y2O3 crucible (Figure 2), so we will focus on BaZrO3 refractories.

According to the previously mentioned ideas, a single refractory material is difficult to meet all needs [30]. Even if the performance of the BaZrO3 crucible is already very good, we can also dope with other refractories in order to pursue a more perfect effect. For example, BaZrO3 crucible doped with CaO [24] or Y2O3 [25,31], in order to determine the best doping ratio, can be guided by the ternary phase diagram. (Figure 3) shows the isothermal section of BaO-ZrO2-YO1.5 and BaO-CaO-ZrO2 at 1750 °C and 1820 °C, respectively. In order to avoid mismatch of physicochemical properties of different phases, the composition of materials should be controlled in the single-phase region or two-phase region with similar properties [32]. Therefore, refractory materials located in the BZ and BCZ phase reigns are the most likely candidates for melting titanium alloys. (Figure 4) shows the SEM images of surface microstructure of BaZrO3 crucible with different doping. (Table 1) lists the energy-dispersive spectroscopy (EDS) results that were taken in regions indicated in (Figure 4). It can be seen from the (Figure 4(a)) that the densification of the BaZrO3 crucible is excellent. Although the densification will be slightly reduced after doping with CaO or Y2O3, the exciting thing is that the grain size is significantly reduced (Figures 4(b) & (C)), mainly because CaO or Y2O3 will become the nucleus point (Table 1), the formation of new phases also has the effect of hindering the movement of the grain boundary, which will undoubtedly increase the strength of the material. (Figures 4(d) & 4(e)) is the doping of different molar amounts of CaZrO3, which can form a microstructure inlaid with small and large grains. This special microstructure will improve the bundling and densification of crucible, which is conducive to improving the thermal shock resistance and Corrosion resistance of the crucible, and the crucible will be more durable. Of course, excessive addition of CaZrO3 will also have a counterproductive effect (Figure 4(d) & Table 1).

Figure 5 shows the erosion of the metal-crucible interface. Analysis the composition of the three points A, B and C by EDS and the morphology of the metal-crucible interface in (Figure 5(a) &Table 2). It can be seen that the TiNi alloy corrodes the CaO crucible more seriously. In contrast, the thickness of the interface reaction layer of CaZrO3 and TiNi alloy is about 30μm (Figure 5(b)). Even better, the thickness of the interface reaction layer of BaZrO3 and TiNi alloy is only about 8μm (Figure 5(c)). Best of all, there is almost no interface reaction layer between the Y2O3-doped BaZrO3 crucible and TiNi alloy (Figure 5(d)). It can be seen that compared with CaO and CaZrO3 crucibles, BaZrO3 crucibles are more effective for preparing TiNi alloys, and the performance of Y2O3- doped BaZrO3 crucible has been further improved. This may be because Y+3 occupies the lattice position of Zr+4 to form two new phases, BaZr1-xYx O3 and Ba2 YZrO6 [25], and the new phases are widely distributed at the grain boundaries (Erosion generally starts at the grain boundaries), thus resisting the corrosion of TiNi melt. Considering the high cost of the BaZrO3 crucible, it is possible to adopt the idea of coating. For example, the costeffectiveness of the BaZrO3 /Al2 O3 crucible is very attractive [33,34].

Finally, to obtain high-quality titanium and titanium alloys, it is necessary to explore more novel melting processes, break the barriers of materials, physics, chemistry, mathematics and other disciplines, and integrate the knowledge of multiple disciplines. In addition, the traditional preparation process is still worthy of in-depth study, such as obtaining better refractories. To evaluate the practicality of a crucible, we must consider its corrosion resistance, durability, cost, high temperature stability, thermal shock resistance, hydration resistance and wettability to the melt. BaZrO3 is a kind of refractory material with great potential, but it still needs a long way to go to the standard of industrial mass production, such as choosing one or more better doping materials and the corresponding ratio, explore more detailed erosion mechanisms, establish corresponding thermodynamic models, and optimize related multiphase diagrams.

The authors wish to thank the Open Project and Independent Research Project of State Key Laboratory of Advanced Special Steel and Shanghai Key Laboratory of Advanced Ferro metallurgy, Shanghai University, (Contract: SKLASS2019-11; SKLASS2019-Z019) and the National Nature Science Foundation of China (Contract: U1760109) for its support. Also thanks to the anonymous referee of this paper for their constructive suggestions.

1. Peters M, Kumpfert J, Ward CH, Leyens C (2003) Titanium alloys for aerospace applications. Advanced Engineering Materials 5(6): 419-427.
2. Oryshchenko A, Gorynin I, Leonov V, Kudryavtsev A, Mikhailov V, et al. (2015) Marine titanium alloys: Present and future. Inorganic Materials: Applied Research 6: 571-579.
3. Fujii H, Takahashi K, Yamashita Y (2003) Application of titanium and its alloys for automobile parts. Nippon Steel Technical Report 88: 70-75.
4. Trevisan F, Calignano F, Aversa A, Marchese G, Lombardi M, et al. (2018) Additive manufacturing of titanium alloys in the biomedical field: Processes, properties and applications. J Appl Biomater Funct Mater 16(2): 57-67.
5. Peters M, Leyens C (2003) Non-aerospace applications of titanium and titanium alloys. pp. 393-422.
6. Weber B, Thompson W, Bielstein H, Schwartz M (2006) Ceramic crucible for melting titanium. Journal of the American Ceramic Society 40(11): 363-373.
7. Fang ZGZ, Paramore JD, Sun P, Chandran KSR, Zhang Y, et al. (2018) Powder metallurgy of titanium-past, present, and future. International Materials Reviews 63: 407-459.
8. Morita A, Fukui H, Tadano H, Hayashi S, Hasegawa J, et al. (2000) Alloying titanium and tantalum by cold crucible levitation melting (CCLM) furnace. Materials Science and Engineering: A 280(1): 208-213.
9. Guo J, Jia J, Yuan L, Liu G, Su Y, et al. (2000) Evaporation behavior of aluminum during the cold crucible induction skull melting of titanium aluminum alloys. Metallurgical and Materials Transactions B 31: 837-844.
10. Tarenkova N, Vykhodets V, Krashaninin V, Kurennykh T, Fishman A (2011) Formation of gas-saturated defects in titanium alloys during vacuum-arc remelting. Russian Metallurgy 127-132.
11. Chronister D, Scott S, Stickle D, Eylon D, Froes F (1986) Induction skull melting of titanium and other reactive alloys. JOM 38: 51-54.
12. Zhang L, Liu Y, Li S, Hao YL (2017) Additive manufacturing of titanium alloys by electron beam melting: A review. Advanced Engineering Materials 20(5): 1700842.
13. Blackburn MJ, Malley DR (1993) Plasma arc melting of titanium alloys. Materials & Design-MATER DESIGN 14: 19-27.
14. Zhang Z, Xing FY, Zhu M, Zhu KL, Lu XG, et al. (2013) Vacuum induction melting of TiNi alloys using BaZrO₃ crucibles. Materials Science Forum 765: 316-320.
15. Schuyler D, Petrusha J, Hall G, Seagle S (1976) Development of titanium alloy casting technology. 3: 332-336.
16. Frueh C, Poirier DR, Maguire M, Harding RA (1996) Attempts to develop a ceramic mould for titanium casting-A review. International Journal of Cast Metals Research 9: 233-239.
17. Faran E, Gotman I, Gutmanas E (2000) Experimental study of the reaction zone at boron nitride ceramic-Ti metal interface. Materials Science and Engineering: A 288: 66-74.
18. Kartavykh A, Tcherdyntsev VV, Zollinger J (2009) TiAl-Nb melt interaction with AlN refractory crucibles. Materials Chemistry and Physics 116(1): 300-304.
19. Frenzel J, Zhang Z, Neuking K, Eggeler G (2004) High quality vacuum induction melting of small quantities of NiTi shape memory alloys in graphite crucibles. Journal of Alloys and Compounds 385(1-2): 214-223.
20. Mulak A (2005) Interfacial reaction between titanium and partially stabilized zirconia (zirconium oxide-9wt% yttrium oxide) at 1800 C, ETD Collection for University of Texas, El Paso, USA.
21. Zhang H, Tang X, Zhou C, Zhang H, Zhang S (2013) Comparison of directional solidification of γ-TiAl alloys in conventional Al₂O₃ and novel Y₂O₃-coated Al₂O₃ crucibles. Journal of the European Ceramic Society 33(5): 925-934.
22. Zhang H, Tang X, Zhou L, Gao M, Zhou C, et al. (2012) Interactions between Ni-44Ti-5Al-2Nb-Mo alloy and oxide ceramics during directional solidification process, Journal of Materials Science 47(17): 6451-6458.
23. Li C, Li M, Zhang H, Ali W, Qin Z, et al. (2017) Fabrication of Y₂O₃ doped BaZrO₃ coating on Al₂O₃ applied to solidification of titanium alloy, Surface & Coatings Technology 320: 146-152.
24. Chen G, Kang J, Lan B, Gao P, Lu X, et al. (2018) Evaluation of Ca-doped BaZrO₃ as the crucible refractory for melting TiAl alloys. Ceramics International 44(11): 12627-12633.
25. Chen G, Gao P, Kang J, Li B, Ali W, et al. (2017) Improved stability of BaZrO₃ refractory with Y₂O₃ additive and its interaction with titanium melts. Journal of Alloys and Compounds 726: 403-409.
26. Chen GY, Li BT, Zhang H, Qin ZW, Lu XG, et al. (2016) On the modification of hydration resistance of CaO with ZrO₂ additive. Int J Appl Ceram Technol 13(6): 1173-1181.
27. Li C, Gao YH, Ding WZ, Ren Z, Deng K (2011) Interaction between the ceramic CaZrO₃ and the melt of titanium alloys. Adv Sci Technol 70: 136-140.
28. Li CH, He J, Wei C, Wang HB, Lu XG (2015) Solidification and interface reaction of titanium alloys in the BaZrO₃ shell-mould. Materials Science Forum 828-829: 106-111.
29. He J, Wei C, Wang S, Meng D, Lu X, et al. (2016) BaZrO₃ refractory applied to the directional solidification of TiAl alloys. IOP Conference Series: Materials Science and Engineering 117: 012033.
30. Zollinger J, Lapin J, Daloz D, Hervé C (2007) Influence of oxygen on solidification behaviour of cast TiAl-based alloys. Intermetallics 15(10): 1343-1350.
31. Chen G, Lan B, Xiong F, Gao P, Qin Z, et al. (2018) Effect of CaO additive on the phase constitution and microstructure of fused BaZrO₃ refractory. Mater Res Express 5: 1-8.
32. Lin C, Wang S, Chen G, Wang K, Cheng Z, et al. (2016) Thermodynamic evaluation of the BaO-ZrO₂-YO_1.5 system. Ceramics International 42(12): 13738-13747.
33. Lan B, Chen G, Xiao Y, Feng Q, Lu X, et al. (2020) Phase and microstructural evolution of BaZrO₃-CaZrO₃ refractory and its interaction with titanium alloy melt. Int J Appl Ceram Technol 17(5): 2193-2201.
34. Sun T, Jiang M, Li C, Liu W (2010) Modification of CaO Refractory for melting titanium alloys and its hydration resistance. Advanced Materials Research 177: 502-505.