Interaction mechanism between BaZrO₃/Y₂O₃ refractory and Ti₂Ni alloy melts

Abstract

The titanium alloys with highly chemical activity require stable crucible refractories that can withstand the erosion of alloy melts. The phase composition and microstructure are crucial factors that affect the stability of the refractory crucible. The effect of Y₂O₃ on the composition and microstructure of BaZrO₃ crucible was systematically investigated, and the improved mechanism of the stability of BaZrO₃/Y₂O₃ crucible was clarified in comparison with the BaZrO₃ crucible. The results showed that the erosion layer thickness of the BaZrO₃/Y₂O₃ crucible was only 63 μm, which was far less than that in the BaZrO₃ crucible (485 μm), and the erosion layer in the BaZrO₃/Y₂O₃ crucible also exhibited a higher density than that in the BaZrO₃ crucible. During the sintering, Y₂O₃ could improve the densification of the BaZrO₃ crucible due to the solid solution effect between Y₂O₃ and ZrO₂, which also caused the evaporation of BaO, resulting in the generation of a Y₂O₃ (ZrO₂) film on the surface of the crucible. Furthermore, the Y₂O₃ (ZrO₂) had higher thermodynamic stability than Y₂O₃, confirming that the BaZrO₃/Y₂O₃ crucible with high density exhibited a superior erosion resistance to titanium alloys. This dual-phase structure provides a strategy to design a long-life and stable refractory for melting titanium alloys.

1 Introduction

Titanium alloys attract the attention due to their high specific strength, corrosion resistance [1, 2] and biocompatibility [3]. However, their extensive use for civilian applications is still hindered by the high costs associated with complicated metallurgical processing [4]. Fortunately, the vacuum induction melting method can help to reduce the production costs of titanium alloys [5]. Additionally, the key challenge is to select a suitable refractory to contain the molten titanium alloys [6].

Common oxides, such as Al₂O₃ [7, 8], ZrO₂ [9], CaO [10] and Y₂O₃ [11, 12], have demonstrated higher corrosion resistance in contact with titanium alloys than carbide, nitride and boride ceramics [13]. However, none of these materials proves to be totally inert against titanium alloy melts. Recently, a perovskite structure material, CaZrO₃ [14, 15] or BaZrO₃, shows good potential as an alternative refractory [16].

Functional thin CaZrO₃ coatings obtained by centrifuging and spraying according to the replica technique improve the properties of titanium alloy casting molds [17]. Schafföner et al. [18] demonstrated that CaZrO₃ exhibits good corrosion resistance in contact with titanium alloy melts. In addition, self-prepared shell molds based on CaZrO₃ ceramic are better than commercial shell systems [19].

A study by Gao et al. [20] revealed that the shape-memory TiNi alloy prepared using the BaZrO₃ crucible could achieve a higher shape-memory rate and fatigue life than that prepared using the graphite crucible. In addition, the TiAl alloy prepared using the BaZrO₃ crucible exhibited a higher purity than that using the MgO and Al₂O₃ crucibles on the laboratory scale [21]. Furthermore, a pilot-scale preparation of TiAl alloy was also achieved in the BaZrO₃ crucible [22].

In fact, the difficulties of the metallurgy of titanium alloys with a refractory crucible lie in the contamination of alloys and the erosion of the refractory crucible. These phenomena will get worse with the increase in titanium content in the alloys due to the increased chemical activity. The researchers found that the BaZrO₃ refractory exhibited insufficient stability during melting the alloys with high titanium content, confirming serious oxygen contamination in the alloy and damage to the crucible [23]. It limits the further application of the BaZrO₃ refractory. Thus, to improve the stability of BaZrO₃ refractories, it is necessary to broaden its extent of melting titanium alloys.

Several composite molds, including Y₂O₃–Al₂O₃–MgO–CaO [24], BN–ZrO₂ [25] and ZrO₂–M_xO_y (M = La, Ce, and Nb) [26], have been proposed. In our previous study, we proposed an effective strategy for designing a dual-phase material using Y₂O₃ and BaZrO₃ [27]. This material was able to maintain the stability of both BaZrO₃ and Y₂O₃, and the dissolution of Y₂O₃ in BaZrO₃ leads to the formation of a solid-solution phase BaZr_1−xY_xO_3−δ, which improved the stability of the refractory. Kuang et al. [28] demonstrated that the penetration of alloy melts into the refractory is related to the wettability, and a theoretical calculation by Zheng et al. [29] indicated that the non-wettability performance of BaZrO₃ refractory was better than that of Y₂O₃ refractory when in contact with TiAl alloy melts.

Although the previous study also revealed that the BaZrO₃/Y₂O₃ refractory exhibited good performance in the process of melting the TiAl alloys [27]. However, the effect of Y₂O₃ on the phase composition and microstructure of BaZrO₃ crucibles has not been thoroughly studied, and the activity of Ti in the TiAl alloy melt is insufficient to obtain a significant difference in the interaction between the refractory and TiAl alloys. In this study, the phase composition and microstructure of the surface and the cross-section of BaZrO₃/Y₂O₃ crucible were carefully investigated, and the interaction mechanism between the BaZrO₃/Y₂O₃ crucible and a high titanium content Ti₂Ni alloy was further clarified in comparison with the BaZrO₃ crucible. The basic research focused on the dual-phase refractory may help meet future demands for melting titanium alloys.

2 Experimental

2.1 Preparation of crucibles

In this study, two different types of crucibles made of fused BaZrO₃ powder (size of  13 μm, Shandong Haoyao New Material Co., Ltd., Shandong, China) and commercial Y₂O₃ powder (purity of 99.9%, size of 5 μm, Sinopharm Chemical Regent Co., Ltd., Shanghai, China) were designed and produced. To prepare the BaZrO₃/Y₂O₃ dual-phase crucible, according to previous research results [27], the ingredients were weighed carefully according to the molar ratio of BaZrO₃ toY₂O₃ of 2:1. In order to mix the ingredients well, the raw materials were first mixed in alcohol, ball-milled on a ball mill for 8 h, and finally dried at 120 °C for 12 h. As shown in Fig. 1, the dried raw material was isostatically cold pressed for 3 min at a pressure of 150 MPa in a U-shaped crucible mold and pressed into a green body. Finally, the BaZrO₃/Y₂O₃ dual-phase crucibles were sintered at 1750 °C for 6 h in a Si–Mo rod sintering furnace. In addition, the preparation conditions of the BaZrO₃ crucible are the same as those of the BaZrO₃/Y₂O₃ dual-phase crucible. The BaZrO₃ crucible after sintering was 5.5 cm in width and 7 cm in height, and the Y₂O₃ crucible after sintering was 4.5 cm in width and 6 cm in height.

Fig. 1

Schematic diagram of preparation route of crucibles and melting experiment

2.2 Erosion resistance test

Ti₂Ni alloys were fabricated in a water-cooled copper induction furnace using a 2:1 molar ratio of titanium sponge (99.9%) to nickel plate (99.9%). The alloys were melted in a medium frequency vacuum induction furnace with the crucibles. Before the melting, the furnace chamber was evacuated down to 10⁻³ Pa and then backfilled with pure argon up to 0.06 MPa for three times. The alloys were melted in the crucible at 1600 °C for 3 min and then poured into a graphite crucible to obtain the alloy ingots. This melting process was repeated for three times, and each crucible is used three times with each time using a different master alloy prepared from the same batch. The detailed flowchart including the preparation of refractory crucibles and the master alloys is also shown in Fig. 1. The phase composition was analyzed by X-ray diffraction (XRD) (D8 Advance, Bruker) with nickel-filtered Cu Kα radiation. The microstructure analysis was examined using a scanning electron microscope (JSM-6700F) combined with the energy-dispersive spectrometry (EDS) equipment. Inductively coupled plasma atomic emission spectrometry (ICP) and LECO TC600 O/N analyzer were utilized to examine the alloy impurity element caused by the crucible refractory. The erosion layer was measured with the Ruler software, and an average was obtained after multiple measurements.

3 Results and discussion

3.1 Phase composition and microstructure of crucibles

Figure 2 shows the scanning electron microscopy (SEM) images of the surface and inner structure of the BaZrO₃ and BaZrO₃/Y₂O₃ crucibles after sintering at 1750 °C for 6 h. Combining the EDS results in Table 1 and Fig. 2a, the surface of the BaZrO₃ crucible was composed of BaZrO₃ and ZrO₂. Our previous study revealed that ZrO₂ could be formed as a second phase during the electric arc furnace process of BaZrO₃ raw material due to the evaporation of BaO [27]. The backscattered electron (BSE) image shows that the amount of ZrO₂ in the surface was little. After the grinding and polishing, the interior of the BaZrO₃ crucible consisted of ZrO₂ (grain C) and BaZrO₃ (grain D), as shown in Fig. 2b, and this was consistent with the phase composition of the crucible surface. In addition, the number of pores in the surface was less than that in the interior. A study by Ahmed et al. [30] revealed that the growth of grains on the surface had more advantages due to easier elimination of pores and lower stress.

Fig. 2

SEM image of surface of BaZrO₃ crucible and EDS element mapping image for O, Zr, and Ba elements (a), interior of BaZrO₃ crucible after grinding and polishing (b), SEM image of surface of BaZrO₃/Y₂O₃ crucible and EDS element mapping image for O, Y, Zr, and Ba elements (c) and interior of BaZrO₃/Y₂O₃ crucible after grinding and polishing (d)

Table 1 EDS results of grains A–H in Fig. 2

The surface of the BaZrO₃/Y₂O₃ crucible in Fig. 2c showed that it consisted of some grains of different shapes, marked as grain E and grain F, respectively. EDS results in Table 1 revealed that these two grains are Ba_1−xZr_1−yY_yO_3−δ and Y₂O₃(ZrO₂). To confirm the distribution of Ba element in the surface of the BaZrO₃/Y₂O₃ crucible, an additional element mapping analysis is shown in the inset in Fig. 2c, which indicated that only 3 at.% Ba was detected, confirming the occurrence of evaporation of BaO on the surface during the sintering. As shown in Fig. 2c, the main phase of the surface of the BaZrO₃/Y₂O₃ crucible is Y₂O₃(ZrO₂), and only a small number of Ba_1−xZr_1−yY_yO_3−δ phase distributed in it. Additionally, no pores were observed on the surface, indicating that the densification of the surface in the BaZrO₃/Y₂O₃ crucible was better than that in the BaZrO₃ crucible. Figure 2d shows the interior of the BaZrO₃/Y₂O₃ crucible, presenting that it also consisted of two kinds of grains, marked as grain G and grain H. EDS analysis in Table 1 exhibits that they were Y₂O₃(ZrO₂) and BaZr_1−xY_xO_3−δ, resulting from the solution of ZrO₂ in Y₂O₃ and the solution of Y₂O₃ in BaZrO₃, respectively. Although some pores appeared in the interior of the BaZrO₃/Y₂O₃ crucible, the number of pores was significantly lower by comparing with the interior in the BaZrO₃ crucible. It confirmed that the Y₂O₃ addition was beneficial to the densification of the BaZrO₃ crucible.

Figure 3 illustrates the XRD patterns of the surface of the BaZrO₃ and BaZrO₃/Y₂O₃ crucibles before and after grinding and polishing. In Fig. 3a, it can be seen that the surface of the BaZrO₃ crucible without grinding and polishing mainly consisted of BaZrO₃. After grinding and polishing, the amount of second phase ZrO₂ was significantly increased, confirmed by the analysis in Fig. 2a, b. During the sintering, the low surface stress was conducive to the growth of the main phase BaZrO₃ grains on the surface, which also influenced the sintering of second phase ZrO₂ [31]. Figure 3b shows that the BaZrO₃/Y₂O₃ crucible consisted of a mixture of BaZrO₃ and Y₂O₃. As shown in Fig. 3c, the magnified XRD pattern of the crucible over the 2$\theta$ range of 27°–37° exhibited a significant peak shifting of BaZrO₃ toward lower angles. It was because the Y³⁺ ions (0.90 Å) could replace the Zr⁴⁺ ions (0.72 Å) in the crystal lattice of BaZrO₃, leading to the formation of BaZr_1−xY_xO_3−δ solid solution [32]. According to the Bragg’s law 2dsin$\theta$=nλ, the increase in d value would cause a decrease in $\theta$ value. Therefore, it can be concluded that the Y₂O₃ peak shifts to higher angles due to the substitution of Zr⁴⁺ ions by Y³⁺ ions. In addition, it can be seen from the phase diagram in Fig. 4 that ZrO₂ could dissolve in Y₂O₃ to form α-Y₂O₃. Especially for the BaZrO₃/Y₂O₃ crucible, the surface of the crucible before grinding and polishing had a small amount of main phase BaZrO₃, and it was different from the interior of the crucible after grinding. Thus, it can be concluded that the solid solution effect between BaZrO₃ and Y₂O₃ was beneficial to the evaporation of BaO [33], which was consistent with the phenomenon in Fig. 2c, d. The theoretical reactions could be used to describe the solid solution between BaZrO₃, ZrO₂ and Y₂O₃ below.

Fig. 3

XRD patterns of crucibles before and after grinding and polishing. a BaZrO₃ crucible; b BaZrO₃/Y₂O₃ crucible; c magnified image over 2θ range of 27°–37° in b

Fig. 4

YO_1.5–ZrO₂ binary diagram. T—Temperature; x(ZrO₂)—content of ZrO₂; Css—cubic structure; Tss—tetragonal structure

$${\mathrm{BaZrO}}_{3}+{\mathrm{Y}}_{2}{\mathrm{O}}_{3}\xrightarrow{\text{sintering}, {1750{\;}^\circ} \mathrm{C}}{\mathrm{BaZr}}_{1-x}{\mathrm{Y}}_{x}{\mathrm{O}}_{3-\delta }$$

(1)

$${\mathrm{Y}}_{2}{\mathrm{O}}_{3}+{\mathrm{ZrO}}_{2}\xrightarrow{\text{sintering}, 1750\;^\circ \mathrm{C}}{\mathrm{Y}}_{2}{\mathrm{O}}_{3}({\mathrm{ZrO}}_{2})$$

(2)

$${\mathrm{BaZrO}}_{3}+{\mathrm{Y}}_{2}{\mathrm{O}}_{3}\xrightarrow{\text{sintering}, 1750\;^\circ \mathrm{C}}{\mathrm{Ba}}_{1-x}{\mathrm{Zr}}_{1-y}{\mathrm{Y}}_{y}{\mathrm{O}}_{3-\delta }+x\mathrm{BaO}$$

(3)

Figure 5 depicts the cross-section of the BaZrO₃ and BaZrO₃/Y₂O₃ crucibles before the melting. From Fig. 5a, the crucible had a two-phase structure, consisting of light gray (spot 1) and black gray (spot 2) substances. EDS results in Table 2 show that they were BaZrO₃ (spot 1) and ZrO₂ (spot 2), respectively, confirming the analysis in the crucible surface. The element mapping in Fig. 5a shows that the two phases exhibited a good homogeneity. Additionally, some pores were significantly observed in the crucible. Figure 5b shows the cross-section of the BaZrO₃/Y₂O₃ crucible, and the crucible also consisted of black (spot 3) and light gray (spot 4) substances. Being different from the BaZrO₃ crucible in Fig. 5a, a large amount of the black gray substance was enriched along the inside wall of the BaZrO₃/Y₂O₃ crucible. The analysis of the element mapping in Fig. 5b exhibited that a large amount of Y and Zr elements were distributed in the black gray substance, and the deficiency of Ba element in this area confirming the occurrence of its evaporation was discussed in Figs. 2c and 3b. In comparing with the BaZrO₃ crucible in Fig. 5a, the Y- and Zr-rich layer had a small number of pores, indicating that it had a higher density. Additionally, this layer also had a higher densification than the crucible matrix. Thus, the Y₂O₃ addition also helped to improve the density of the BaZrO₃ crucible due to the effect of solid solution sintering.

Fig. 5

SEM images and elemental mapping of cross-section of BaZrO₃ crucible (a) and BaZrO₃/Y₂O₃ crucible (b) before melting

Table 2 EDS results of spots 1–4 in Fig. 5

3.2 Interaction between BaZrO₃ crucibles and Ti₂Ni alloys

Figure 6 illustrates the surface microstructure of the BaZrO₃ and BaZrO₃/Y₂O₃ crucibles before and after melting Ti₂Ni alloys. Figure 6a displays that the surface of the BaZrO₃ crucible before melting exhibited an even surface along with the appearance of some pores. From Fig. 6b, the surface of the BaZrO₃ crucible after melting was eroded by the alloy melts, leading to the appearance of the inside grains in the crucible, and a large number of pores were also observed. This was due to the destruction of the grain boundaries on the crucible surface by the molten alloy, which then penetrated inside through the pores of the crucible, further damaging the internal structure. From Fig. 6c, the surface of the BaZrO₃/Y₂O₃ crucible before the melting had a complete and dense structure. From Fig. 6d, a similar erosion phenomenon was also observed in the BaZrO₃/Y₂O₃ crucible after the melting. However, a residual flat surface, which consisted of subsphaeroidal grains, appeared in the BaZrO₃/Y₂O₃ crucible. Apparently, the surface of the BaZrO₃/Y₂O₃ crucible was not totally damaged by the molten alloys, indicating that the Y₂O₃ addition can improve the erosion resistance of the BaZrO₃ crucible in the process of melting titanium alloy.

Fig. 6

SEM images of surface of crucibles before and after melting Ti₂Ni alloys. a BaZrO₃ crucible before melting; b BaZrO₃ crucible after melting; c BaZrO₃/Y₂O₃ crucible before melting; d BaZrO₃/Y₂O₃ crucible after melting

Figure 7a displays the magnified images of area I in Fig. 6d, and some subsphaeroidal grains existed on the surface of the crucible. From the surface analysis in Fig. 7b, a uniform distribution of Y and Zr elements in the subsphaeroidal grains was observed, indicating that they were Y₂O₃(ZrO₂) grains. In addition, BaZrO₃ grains were mainly distributed inside the crucible from the distribution form of Zr and Ba elements. This suggests that the residual high-stability Y₂O₃ refractory formed a thin film in the crucible surface, which could effectively prevent the erosion by the alloy melts.

Fig. 7

Magnified image of area I in Fig. 6d (a), combination of all elements (b), and EDS element mapping images for Y, Zr, Ba, and O elements (c–f)

The XRD patterns of the BaZrO₃ and BaZrO₃/Y₂O₃ crucible surface before and after melting are shown in Fig. 8. From Fig. 8a, it can be seen that no other new phases were observed on the inner surface of the BaZrO₃ crucible after the melting. Study by Chen et al. [16] indicated that the physical dissolution of the crucible refractory in titanium melts was the main interaction mechanism between the crucible and the alloy melt. However, the peaks of ZrO₂ phase at 35.088°, 50.428° and 59.999° disappeared on the BaZrO₃ crucible surface after melting. In addition, the intensity of BaZrO₃ peaks at 30.238°, 43.264° and 53.687° also exhibited a decreasing trend than that of the peaks before the melting. The detailed peak values are listed in Table 3. It can be concluded that the molten alloys dissolved the second phase of ZrO₂ on the crucible surface. Furthermore, the dissolution of BaZrO₃ refractory caused the loose and uneven surface structure, leading to the decreased intensity of BaZrO₃ peaks. The XRD result shows that no interaction product was generated in the BaZrO₃/Y₂O₃ crucible surface after the melting. From Table 3, it can be seen that the intensity of BaZr_1−xY_xO_3−δ peaks at 30.357°, 43.403° and 53.889° before the melting exhibited a decreasing trend than that of the peaks before the melting. However, the intensity of Y₂O₃ peaks at 29.599°, 34.269° and 49.157° exhibited an increasing trend after the melting. It appears that the molten alloy dissolved a significant amount of the BaZrO₃ refractory in the surface, leaving behind residual Y₂O₃ refractory, as observed in Fig. 6. Hence, the residual Y₂O₃ refractory exhibited a larger XRD intensity than BaZrO₃ refractory. According to the comprehensive analysis in Figs. 6 and 7, it can be concluded that the BaZrO₃ refractory exhibited a priority erosion phenomenon than the Y₂O₃ refractory in the surface of the BaZrO₃/Y₂O₃ crucible. The study by Oyama et al. [34] revealed that the thermodynamic stability of Y₂O₃ was higher than that of BaZrO₃. It should be noted that this interesting phenomenon might be decided by the competition erosion behavior of refractory in molten titanium alloy. It meant that the refractory with lower stability would exhibit a priority erosion phenomenon. And the relative erosion degree of the refractory in the molten titanium alloys should be declined in the following sequence: ZrO₂ > BaZrO₃ > Y₂O₃.

Fig. 8

XRD patterns of surface of crucibles before and after melting Ti₂Ni alloys

Table 3 Intensity of characteristic peaks of different phases in surface of crucibles

To further investigate the performance of the BaZrO₃ and BaZrO₃/Y₂O₃ crucibles for melting Ti₂Ni alloys, the microstructure images of the cross-section of the crucibles before and after melting are shown in Fig. 9. From Fig. 9a, the BaZrO₃ crucible before melting is observed to have a complete structure with a light-yellow color. After melting, a contact layer with a gray color, with 3190 μm in thickness, was observed in the BaZrO₃ crucible, as shown in Fig. 9b. The magnified SEM image of area II in Fig. 9b is shown in Fig. 9c. It can be seen that the crucible had a 485 μm thick erosion layer, confirmed by the loose and lamellar structure. During the melting, the alloy melt would erode the surface of the crucible and further permeate along the porosity in the crucible, resulting in the formation of the erosion layer. In addition, due to the capillary force, the alloy melt would further permeate into the crucible, resulting in a thicker contact layer. Figure 9d shows that the cross-section of the BaZrO₃/Y₂O₃ crucible before the melting also presented a complete structure. After the melting, it can be seen from Fig. 9e that a contact layer with a thickness of 3750 μm appeared in the crucible. The enlarged view of area III in Fig. 9f shows that the thickness of the erosion layer in the crucible was only 63 μm with no appearance of cracks. Hence, it indicates that the BaZrO₃/Y₂O₃ crucible exhibited a better erosion resistance to Ti₂Ni alloy melts than the BaZrO₃ crucible.

Fig. 9

Images of cross-section of crucibles. a BaZrO₃ crucible before melting; b BaZrO₃ crucible after melting; c magnified picture of area II in b; d BaZrO₃/Y₂O₃ crucible before melting; e BaZrO₃/Y₂O₃ crucible after melting; f magnified picture of area III in e

Figure 10 depicts the SEM images of the cross-section of the erosion layer in the BaZrO₃ and BaZrO₃/Y₂O₃ crucibles. In Fig. 10a, it can be observed that the structure of the erosion layer in the BaZrO₃ crucible was damaged by the alloy melts, resulting in dispersed grains. In addition, from Fig. 10a, there was no obvious ZrO₂ phase in the erosion layer, indicating that the molten alloys dissolved a large amount of phase ZrO₂, and it was in agreement with the analysis in Fig. 8. Figure 10b shows the microstructure of the erosion layer in the BaZrO₃/Y₂O₃ crucible. Although the Y and Zr-rich layer in the crucible was damaged by the alloy melts, this erosion layer still exhibited a relatively dense structure in comparison with that in the BaZrO₃ crucible. Furthermore, the BaZrO₃/Y₂O₃ crucible showed excellent erosion resistance to the alloy melts due to the uniform distribution of the Y, Zr, and Ba elements in the erosion layer.

Fig. 10

SEM images and element mapping of cross-section of BaZrO₃ crucible (a) and BaZrO₃/Y₂O₃ crucible (b) after melting

Figure 11 displays the measured concentration of O, Zr and Y elements in the alloys obtained from the BaZrO₃ and BaZrO₃/Y₂O₃ crucibles. From Fig. 11a, it can be seen that the oxygen concentration in the alloys melted in the BaZrO₃ crucible decreased from 0.50 to 0.39 wt.%. In addition, the alloys melted in the BaZrO₃/Y₂O₃ crucible also exhibited a decreased oxygen concentration (from 0.33 to 0.11 wt.%). Furthermore, as demonstrated in Fig. 11b, the Zr concentration in the alloys decreased from 0.640 to 0.635 wt.% after melting in the BaZrO₃ crucible. However, the Zr concentration in the alloys melted in the BaZrO₃/Y₂O₃ crucible also had a significant downward trend (from 0.350 to 0.075 wt.%). Compared to the BaZrO₃ crucible, the BaZrO₃/Y₂O₃ crucible showed less O and Zr contaminations in the alloys, indicating its superior erosion resistance to the Ti₂Ni melts. The analysis in Fig. 6 supports this conclusion by suggesting that the remaining Y₂O₃ refractory in the crucible surface had excellent erosion resistance to the Ti₂Ni melts, which prevented the further dissolution of the crucible refractory. Figure 11b shows that the Y concentration in the alloys obtained from the BaZrO₃/Y₂O₃ crucible was 0.078, 0.130 and 0.130 wt.%, respectively. It was because there would be some Y contamination in the alloy melts due to the presence of Y₂O₃. Apparently, the low refractory contamination of the BaZrO₃/Y₂O₃ crucible could reflect its good erosion resistance to the alloys.

Fig. 11

Concentration of O element (a) and Zr and Y elements (b) in alloys melted in BaZrO₃ and BaZrO₃/Y₂O₃ crucibles

3.3 Improved mechanism of stability of BaZrO₃/Y₂O₃ crucible

In our previous study [27], although the dissolution of the BaZrO₃/Y₂O₃ refractory was mentioned in the melting of titanium alloys, the mechanism of its improved stability was not clarified. Generally, the interaction extent between the crucible and the titanium alloy was influenced by the thermodynamic stability of the crucible refractory and the interaction kinetics [35]. Specifically, the lower Gibbs free energy of the refractory and the higher density of the crucible meant better erosion resistance.

Figure 12a shows the dependence of the change in Gibbs free energy ∆G versus the temperature for the formation of Y₂O₃, BaZrO₃, ZrO₂, Ti and TiO₂, which was obtained from the HSC software. Meanwhile, the Gibbs free energy of Y₂O₃ was more negative than that of BaZrO₃, indicating that Y₂O₃ was more stable than BaZrO₃. Figure 12b is about the Y₂O₃–ZrO₂ phase diagram system, which shows the change of Gibbs free energy of α-Y₂O₃ phase with the amount of ZrO₂ solid solution at 1750 °C. With the increase in ZrO₂ solid solution, the energy of the whole system decreases, and the α-Y₂O₃ phase is more stable. Thus, it can be confirmed that the stability of the BaZrO₃/Y₂O₃ crucible refractory was superior to that of the BaZrO₃ crucible refractory. Furthermore, due to the effect of the solid solution sintering, the BaZrO₃/Y₂O₃ crucible exhibited a high density than that of the BaZrO₃ crucible, confirmed by the analysis in Figs. 2 and 5. Thus, the low porosity of the BaZrO₃/Y₂O₃ crucible meant that it had a lower interaction rate. In summary, the BaZrO₃/Y₂O₃ crucible had a thinner erosion layer and a lower contamination extent to the alloy melts in comparison with the BaZrO₃ crucible.

Fig. 12

Ellingham diagram of Y₂O₃, BaZrO₃, ZrO₂, Ti and TiO₂ (a) and change in Gibbs free energy (G) in solid solution region of ZrO₂–YO_1.5 binary phase diagram (b)

Figure 13 presents a schematic of interactions between the Ti₂Ni melt and the crucible during the melting process. The dissolution of the crucible refractory damaged the structure of the crucible, leading to crucible erosion. As shown in Fig. 13a, ZrO₂ grains, which were on the surface of the BaZrO₃ crucible, would be totally dissolved by the alloy melts. From Fig. 13b, BaZrO₃ grains would be firstly dissolved, resulting in the residual Y₂O₃ grains on the surface of BaZrO₃/Y₂O₃ dual-phase refractory crucible. In addition, Y₂O₃ grains with higher stability could effectively resist melt erosion. Thus, the BaZrO₃/Y₂O₃ dual-phase refractory crucible demonstrated superior erosion resistance to Ti₂Ni alloy melts than the BaZrO₃ crucible.

Fig. 13

Schematic diagram of interaction between crucibles and alloy melts. a BaZrO₃ crucible; b BaZrO₃/Y₂O₃ crucible

4 Conclusions

• 1. The BaZrO₃/Y₂O₃ crucible demonstrated excellent resistance to the Ti₂Ni alloy, with only a 63 μm thick erosion layer compared to the 485 μm thick layer in the fused BaZrO₃ crucible.

• 2. After melting the alloys in the BaZrO₃ crucible, the oxygen concentration in the alloys varied from 0.50 to 0.39 wt.%. However, the BaZrO₃/Y₂O₃ crucible exhibited lower refractory contamination to the alloys, and the oxygen concentration varied from 0.33 to 0.11 wt.%.

• 3. The improved stability of BaZrO₃/Y₂O₃ was attributed to the densification of the BaZrO₃ crucible with the Y₂O₃ addition and the generation of phase Y₂O₃(ZrO₂) with higher stability. It could reduce the interaction extent of the BaZrO₃/Y₂O₃ crucible with Ti₂Ni alloy melts.

• 4. In the melting process, BaZrO₃ in the dual-phase refractory crucible would be firstly dissolved into the alloy melts, leading to the formation of a complete Y₂O₃ layer, which acted as a barrier to resist the erosion of the alloy melts.