Free Access
Issue
Metall. Res. Technol.
Volume 117, Number 2, 2020
Article Number 206
Number of page(s) 7
DOI https://doi.org/10.1051/metal/2020015
Published online 17 April 2020

© EDP Sciences, 2020

1 Introduction

Inclusion issue is one of the most significant things to assure the quality of superalloys. They are usually compounds of metals (aluminum, magnesium, titanium, calcium, cerium, etc.) with non-metallic elements (oxygen, sulfur, and nitrogen) [14]. Numerous studies have demonstrated that inclusions degraded the mechanical properties (e.g., tensile strength, and fracture, specifically fatigue properties at high temperatures) of superalloys [57]. For example, Texier et al. [5] found that surface or near-surface inclusions generated fatigue cracks because of the high anisotropic elastic properties of nickel-based superalloys and the large differences in the elastic properties between inclusions and the surrounding metallic matrix of IN718 superalloy. Therefore, inclusion characterization is a prerequisite for removing inclusions. Inclusions in superalloys are mainly oxides and nitrides. In most cases, oxides are surrounded by nitrides and carbides [24]. This leads to difficulty in characterizing the size and composition of oxides accurately. However, there are few kinds of literature regarding the separation of oxides from nitrides and nitrides.

This paper focuses on the separation of oxides from nitrides and carbides based on two factors. The first one is levitation melting technique. In the past two decades, levitation melting is widely used to refine various alloys because of sufficient stirring and turbulence in molten alloy [816]. Moreover, Toh et al. [17,18] mentioned that inclusions are accumulated to the surface of the melt during levitation melting. The other factor is the different behavior of oxides, nitrides and carbides between melting and solidification. It is reported that oxides remain in the alloy melt, while nitrides and nitrides dissolved during melting and precipitated during solidification and the subsequent cooling process [1921].

2 Experimental

The billet material used in this study was from the industrial FGH96 superalloy billet, of which the chemical composition was Ni-16.2Cr-13.1Co-3.9W-3.9Mo-0.8Nb-3.7Ti-2.2Al-0.03C (wt. %). The contents of O and N in the billet were 21 and 28 ppm, respectively. A 5-kg-scale vacuum electromagnetic levitation melting furnace (ZGXF-0.002/Q0004) was employed for the experiment. A total of 4.3 kg billet was used after grinding and polishing to remove the dust and oxidation layer. Before levitation melting, the chamber was evacuated to the vacuum pressure of 9.0×10−3 Pa and backfilled with pure argon up to 0.05 MPa to reduce re-oxidation of the melt by air atmosphere, which would have altered the inclusions. The furnace mainly consists of a water-cooled copper crucible with a diameter of 90 mm, which is different from traditional vacuum induction melting furnace. Therefore, opportunities for contamination at refining temperatures are minimized.

Scanning Electron Microscope with Energy Dispersive X-ray Spectroscopy (SEM/EDS) as well as optical microscopy (OM) and X-ray diffraction (XRD) are used to characterize the morphology, composition, and size of inclusions. Also, automated SEM (ZEISS, EVO18) with an EDS (Oxford Instrument, INCA) method was explored to quantitatively characterize the number, size, composition and especially spatial distribution of inclusions in a large area (4 mm × 4 mm for each sample). The surface oxide film of the samples was eliminated through polishing. Fine contaminants, such as dust remaining on the sample surface, were eliminated through ultrasonic cleaning. In the present work, the minimum particle size was set at 1.0 µm, which indicates that only inclusions larger than 1.0 µm can be obtained using automated SEM-EDS.

However, directly observing the variation of inclusions during levitation melting is challenging. An ultrahigh temperature confocal laser scanning microscopy (CLSM) was therefore used in the present study to simulate this process. Reports have demonstrated that CLSM in effective to observe the real time behavior of individual inclusions moving on the surface of the molten melt, including their nucleation, collision, and agglomeration [2224]. Specific details of the CLSM technique have been widely published and are detailed elsewhere [19]. The specimen was machined into a disc (φ7.8 mm × 2.5 mm), carefully polished into a plane, and placed in an alumina crucible (ϕ8.0 mm × 3.0 mm) after the inclusions were characterized by SEM-EDS and OM. Argon gas was passed through a device to capture water and oxygen and then introduced into the chamber as a protective atmosphere to prevent re-oxidation of the specimen surface. Additionally, a tantalum foil was wrapped around the crucible to act as an oxygen getter considering the high contents of active elements (Al and Ti). The program-controlled heating rate of CLSM was set as 2 K/s from room temperature to 1473 K and then 25 K/min to 1673 K to observe the melting process clearly. The melt was kept at 1673 K for 60 s to obtain homogeneous temperature and composition. Immediately after that, the melt was cooled at 25 K/min to observe the solidification process to 1473 K. The deviation of temperature due to the thermocouple placement in the furnace chamber was approximately ±2 K.

3 Results and discussion

3.1 Inclusions in the billet

Before levitation melting, inclusions in the billet can be classified to four types: oxide-nitride two-layer inclusions (Fig. 1a), oxide-nitride-carbide three-layer inclusions (Fig. 1b), nitride-carbide two-layer inclusions (Fig. 1c), and pure nitrides (Fig. 1d). Nitrides are TiN and carbides are (Nb,Ti,Mo)C. Figure 1e–f shows the spatial distribution of inclusions by manual and automated SEM, respectively. In manual SEM image, only several inclusions can be seen at an observed field. In automated SEM image, however, more inclusions can be shown in a picture, which can reflect the spatial distribution of inclusions more accurately. The result shows that the spatial distribution of inclusion in the billet is even. No inclusion cluster is observed. EDS result of core oxides shows that oxides are Al-Mg-Ti oxides. It should be noted that the EDS results can be interrupted by the surrounding phase when the size of oxides is smaller than 1 µm. In the present study, the surrounding phase is TiN. Therefore, it is uncertain whether the core oxide contains Ti element, which is also an active element to generate oxide inclusions during melting.

thumbnail Fig. 1

(a–d) Morphology, and (e–f) spatial distribution of inclusions in the billet before levitation melting.

3.2 Inclusions after levitation melting

After levitation melting, inclusions in samples cut from different positions of the levitation melted ingot are observed. Figure 2a is the schematic of sampling. A total of six samples are analyzed and referred to as S1∼S6. Figure 2b shows that inclusion clusters are observed in each sample. The size of inclusion clusters is larger at the top (100∼200 µm) than that at the bottom (30∼50 µm). This implies that inclusions gather and float to the top during levitation melting. Elemental mapping of inclusion clusters indicates that inclusions in the cluster are oxides (Al2O3-MgO) and nitrides (TiN), as shown in Figure 3. Some oxides are still surround by TiN. In addition, OM image is effective for observation of inclusion cluster (Fig. 2c). The dark particles represent oxides, while the yellow ones represent nitrides. The size of oxides is clearly smaller than that of nitrides. It is obvious that oxides and nitrides have been separated, although there are also some mixed inclusions. The typical morphology of pure oxides is shown in Figure 4a. Small oxide cluster consisting of several oxides are observed. The size distribution of separated oxides is shown in Figure 4b. About 82.3% oxide inclusions are under the size of 5 µm and the average size is 3.2 µm. In this study, randomly more than 100 Al-Mg oxides are analyzed. The MgO content in the Al2O3-MgO oxides is shown in Figure 4c. The result shows that the composition of Al-Mg oxides is in the scope of MgO · Al2O3 spinel. It should be noted that MgO · Al2O3 spinel is solid solution, rather than stoichiometric compound (MgAl2O4).

Automated SEM-EDS method further confirms the aggregation distribution of inclusions, as shown in Figure 5. It is widely acknowledged that clusters are more easily to float up to the surface of molten alloys because of the buoyancy than single inclusions. However, in the present study, there are still inclusions in the ingot and approximately 82% of them existed in the form of clusters. This can be explained by the solidification condition. The alloy melt solidified in a water-cooled copper crucible after levitation melting, which means that the solidification process was at a high speed. Therefore, there was not sufficient time for inclusion clusters especially those at the bottom to float to the surface thoroughly.

In the sample S6, the number fraction of pure oxides, pure nitrides and oxide-nitride inclusions are 56%, 34.5% and 9.5%, respectively. This means that the separation efficiency of oxides from nitrides is approximately 85.5%. Besides the clusters, the number of single oxides in other area of the samples is very few, while the number of single nitrides is much more. The distribution of nitrides is even at the whole sample. This can be explained by the precipitation of nitrides during solidification.

Based on the above experimental results, oxides and nitrides are separated effectively by levitation melting. The reason might be that oxides and nitrides have different dissolution and precipitation behavior during melting and solidification. TiN inclusions began to dissolve before the base alloy was completely liquefied [19] during heating to solidus temperature and melting. This results in the exposure of core oxide inclusions, which will gather due to strong collision by electromagnetic force during levitation melting. Then the oxide clusters grow bigger and float to the top surface. Figure 6 shows the morphology and area-scanning of the top surface in the ingot. MgO and Al2O3 are tested. Figure 6c displays the oxide particles at the top surface. Figure 6d confirms that the oxides are MgO-Al2O3 spinel. This is in agreement with the work by Toh et al. [17] that levitation melting enables about 80% of oxide inclusions to go out to the surface. Finally, nitrides precipitate again during solidification [20,21]. Some mixed inclusions are still observed because pre-existing oxides are always the preferential sites for the nucleation of nitrides.

thumbnail Fig. 2

Inclusions in the samples from different positions of the levitation melted ingot.

thumbnail Fig. 3

Elemental mapping of a typical inclusion cluster.

thumbnail Fig. 4

(a) Morphology, (b) size distribution, and (c) MgO content of separated oxides in the levitation melted alloy.

thumbnail Fig. 5

Spatial distribution of inclusions in the levitation melted alloy by automated SEM-EDS.

thumbnail Fig. 6

Top surface in the levitation melted ingot: (a) overall image, (b) elemental mapping of a randomly selected area, (c) high resolution of oxide particles, (d) XRD analysis.

3.3 In situ observation of inclusion behavior

The dissolution of oxides and nitrides inclusions in the specimen upon heating is displayed in Figure 7a–f. At 1274.3 K and 1374.5 K, the shape of a nitride is clearly observed and remains unchanged. At 1547.3 K, however, the size of nitride shrinks slightly. At 1674.4 K, which is higher than the liquidus line (1623 K) of the superalloy, nitrides disappear (Fig. 7f). This confirms the TiN dissolution during melting. Since the heating rate is larger during CLSM, the time for complete dissolving of TiN is insufficient. It also can be seen that oxide inclusions still remain at 1673 K, suggesting that oxides are stable at melting temperature.

Subsequently, in situ CLSM of oxides and nitrides with decreasing temperature during solidification was observed, as shown in Fig. 7g–j. Single and cluster oxides are observed at 1674.4 K, which is about 50 K higher than the liquidus line of the superalloy, suggesting that oxides are stable during melting and will not dissolve into the alloy melt. At 1560.7 K, which is 17 K higher than the solidus temperature, a TiN particle with regular shape precipitates near the primary dendrite, as shown in Figure 7i. This confirms that TiN nitrides precipitate during solidification.

Moreover, the agglomeration behavior of oxides at melting temperatures is observed by CLSM, as shown in Figure 8. A small cluster consisting of two oxide particles rotates and moves to a bigger cluster at temperatures of 1573.7 K to 1568.8 K. At 9 s, the small cluster collided and merged with the big cluster. This is in good agreement with the work by Shibata [22,23] revealing that inclusion particles form larger clusters both due to the fluid flow that exists at the surface as well as the attraction forces that exist between inclusions. Furthermore, the attraction is a strong attractive force which nature is capillary effect and can result in the rapid collision among the single particles and small aggregates as soon as they emerge on the surface. Factors having an impact on the agglomeration of inclusions are mainly the size and composition of inclusions. Surfactant element such as sulfur content of the alloy melt has little effect on the attraction when sulfur content is lower than 0.021 wt.%.

thumbnail Fig. 7

In situ observation of inclusions (a–f) during melting, and (g–j) during solidification by CLSM.

thumbnail Fig. 8

In situ CLSM observation of oxides agglomeration at temperatures of 1573.7 K to 1568.8 K.

4 Conclusions

The present study offers an approach separating oxides and nitrides in the superalloy by levitation melting. According to the different dissolution and precipitation behavior of oxides and nitrides, levitation melting boosts the agglomeration and floatation of oxides, most of which floats to the top surface. The remaining inclusions in the inner of the ingot exist mainly in the form of clusters. In situ confocal laser scanning microscopy clarifies the separation mechanism of oxides from nitrides and carbides. Nitride inclusions are dissolved during melting and precipitate again during solidification. However, oxide inclusions are stable during melting. This approach is also useful to separate core inclusions (oxides) from other shell inclusions (nitrides, sulfides) in a variety of alloys such as special steels.

Acknowledgements

This work is financially supported by National Natural Science Foundation of China (51574029, 51574030, 51974029) and National Major Science and Technology Projects of China (grant No. 2017-I-0001-0001).

References

  1. M.H. Manjili, M. Halali, Removal of non-metallic inclusions from nickel base superalloys by electromagnetic levitation melting in a slag, Metall. Mater. Trans. B 49, 61–68 (2018) [CrossRef] [Google Scholar]
  2. J.D. Busch, J.J. Debarbadillo, M.J.M. Krane, Flux entrapment and titanium nitride defects in electroslag remelting of INCOLOY alloys 800 and 825, Metall. Mater. Trans. A 44, 5295–5303 (2013) [CrossRef] [Google Scholar]
  3. X.C. Chen, C.B. Shi, H.J. Guo, F. Wang, H. Ren, D. Feng, Investigation of oxide inclusions and primary carbonitrides in Inconel 718 superalloy refined through electroslag remelting process, Metall. Mater. Trans. B 43, 1596–1607 (2012) [CrossRef] [Google Scholar]
  4. H.E.O. Kellner, A.V. Karasev, O. Sundqvist, A. Memarpour, P.G. Jönsson, Estimation of non-metallic inclusions in industrial Ni based alloys 825, Steel Res. Int. 88(4), 1–8 (2017) [Google Scholar]
  5. D. Texier, J. Cormier, P. Villechaise, J.C. Stinville, C.J. Torbet, S. Pierret, T.M. Pollock, Crack initiation sensitivity of wrought direct aged alloy 718 in the very high cycle fatigue regime, Mater. Sci. Eng. A. 678, 122–136 (2016) [CrossRef] [Google Scholar]
  6. G.L. Miao, X.G. Yang, D.Q. Shi, Competing fatigue failure behaviors of Ni-based superalloy FGH96 at elevated temperature, Mater. Sci. Eng. A. 668, 66–72 (2016) [CrossRef] [Google Scholar]
  7. J. Jiang, J. Yang, T.T. Zhang, F.P.E. Dunne, T.B. Britton, On the mechanistic basis of fatigue crack nucleation in Ni superalloy containing inclusions using high resolution electron backscatter diffraction, Acta Mater. 97, 367–369 (2015) [Google Scholar]
  8. S.J. Sun, Y.Z. Tian, H.R. Lin, X.G. Dong, Y.H. Wang, Z.J. Zhang, Z.F. Zhang, Enhanced strength and ductility of bulk CoCrFeMnNi high entropy alloy having fully recrystallized ultrafine-grained structure, Mater. Des. 133, 122–127 (2017) [Google Scholar]
  9. S.Q. Xia, Y. Zhang, Deformation mechanisms of Al0.1CoCrFeNi high entropy alloy at ambient and cryogenic temperatures, Mater. Sci. Eng. A. 733, 408–413 (2018) [CrossRef] [Google Scholar]
  10. S.Q. Xia, M.C. Gao, T.F. Yang, P.K. Liaw, Y. Zhang, Phase stability and microstructures of high entropy alloys ion irradiated to high doses, J. Nucl. Mater. 480, 100–108 (2016) [Google Scholar]
  11. S.Q. Xia, X. Yang, T.F. Yang, S. Liu, Y. Zhang, Irradiation resistance in AlxCoCrFeNi high entropy alloys, JOM 67, 2340–2344 (2015) [CrossRef] [Google Scholar]
  12. Y.Z. Li, F. Hu, L. Luo, J.Y. Xu, Z.W. Zhao, Y.H. Zhang, D.L. Zhao, Hydrogen storage of casting MgTiNi alloys, Catal. Today 318, 103–106 (2018) [Google Scholar]
  13. C.D. Rabadia, Y.J. Liu, G.H. Cao, Y.H. Li, C.W. Zhang, T.B. Sercombe, H. Sun, L.C. Zhang, High-strength β stabilized Ti-Nb-Fe-Cr alloys with large plasticity, Mater. Sci. Eng. A 732, 368–377 (2018) [CrossRef] [Google Scholar]
  14. M.Y. Wu, X.R. Yang, R.X. Zou, F.J. Qian, S.Y. Hu, W.Y. Wang, G.L. Zhong, X.F. Miao, F. Xu, A time-, energy-, and cost-efficient way of preparing (MnFe)2(P, Si)-type magnetocaloric materials, Mater. Lett. 236, 579–582 (2009) [Google Scholar]
  15. Z.D. Yao, X.Z. Xiao, Z.Q. Liang, H.Q. Kou, W.H. Luo, C.G. Chen, L.J. Jiang, L.X. Chen, Improvement on the kinetic and thermodynamic characteristics of Zr1-xNbxCo (x=0-0.2) alloys for hydrogen isotope storage and delivery, J. Alloys Compd. 784, 1062–1070 (2019) [Google Scholar]
  16. M. Besse, P. Castany, T. Gloriant, Mechanisms of deformation in gum metal TNTZ-O and TNTZ titanium alloys. A comparative study on the oxygen influence, Acta Mater. 59, 5982–5988 (2011) [Google Scholar]
  17. T. Toh, H. Yamamura, H. Kondo, M. Wakoh, S.I. Shimasaki, S. Taniguchi, Kinetics evaluation of inclusions removal during levitation melting of steel in cold crucible, ISIJ Int. 47, 1625–1632 (2007) [CrossRef] [Google Scholar]
  18. T. Toh, H. Yamamura, H. Kondo, M. Wakoh, E. Takeuchi, Inclusions behavior analysis during levitation melting of steel in cold crucible for application to cleanliness assessment, ISIJ Int. 45, 984–990 (2005) [CrossRef] [Google Scholar]
  19. L. Yang, G.G. Cheng, Characteristics of Al2O3, MnS, and TiN inclusions in the remelting process of bearing steel, Int. J. Miner. Metall. Mater. 24, 869–875 (2017) [CrossRef] [Google Scholar]
  20. Y. Liu, L.F. Zhang, H.J. Duan, Y. Zhang, Y. Luo, A.N. Conejo, Extraction, thermodynamic analysis and precipitation mechanism of MnS-TiN, Metall. Mater. Trans. A 47, 3015–3025 (2016) [CrossRef] [Google Scholar]
  21. Y. Luo, W. Yang, Q. Ren, Z.Y. Hu, M. Li, L.F. Zhang, Evolution of non-metallic inclusions and precipitates in oriented silicon steel, Metall. Mater. Trans. B 49, 926–932 (2018) [CrossRef] [Google Scholar]
  22. J. Appelberg, K. Nakajima, H. Shibata, A. Tilliander, P. J̈onsson, In situ studies of misch-metal particle behavior on a molten stainless steel surface, Mater. Sci. Eng. A 495, 330–334 (2008) [CrossRef] [Google Scholar]
  23. H.B. Yin, H. Shibata, T. Emi, M. Suzuki, In-situ observation of collision, agglomeration and cluster formation of alumina inclusion particles on steel melts, ISIJ Int. 37, 936–945 (1997) [CrossRef] [Google Scholar]
  24. H. Shibata, H.B. Yin, S. Yoshinaga, T. Emi, M. Suzuki, In-situ observation of engulfment and pushing of nonmetallic inclusions in steel melt by advancing melt-solid interface, ISIJ Int. 38, 149–156 (1998) [CrossRef] [Google Scholar]

Cite this article as: Xiaoyong Gao, Lin Zhang, Xuanhui Qu, Yifeng Luan, Xiaowei Chen, Application of levitation melting to the separation of oxide inclusions from nitrides and carbides in Ni-based superalloy, Metall. Res. Technol. 117, 206 (2020)

All Figures

thumbnail Fig. 1

(a–d) Morphology, and (e–f) spatial distribution of inclusions in the billet before levitation melting.

In the text
thumbnail Fig. 2

Inclusions in the samples from different positions of the levitation melted ingot.

In the text
thumbnail Fig. 3

Elemental mapping of a typical inclusion cluster.

In the text
thumbnail Fig. 4

(a) Morphology, (b) size distribution, and (c) MgO content of separated oxides in the levitation melted alloy.

In the text
thumbnail Fig. 5

Spatial distribution of inclusions in the levitation melted alloy by automated SEM-EDS.

In the text
thumbnail Fig. 6

Top surface in the levitation melted ingot: (a) overall image, (b) elemental mapping of a randomly selected area, (c) high resolution of oxide particles, (d) XRD analysis.

In the text
thumbnail Fig. 7

In situ observation of inclusions (a–f) during melting, and (g–j) during solidification by CLSM.

In the text
thumbnail Fig. 8

In situ CLSM observation of oxides agglomeration at temperatures of 1573.7 K to 1568.8 K.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.